Docker image built to run code, all evals run using R2RHEAD master

author: ShelbySolomonDarnell 2024-10-17 12:24:26 +0300
committer: ShelbySolomonDarnell 2024-10-17 12:24:26 +0300
commit: 00cba4b9a1e88891f1f96a1199320092c1962343 (patch)
tree: 270fd06daa18b2fc5687ee72d912cad771354bb0 /gnqa/paper2_eval/data/dataset/human/intermediate_files
parent: e0b2b0e55049b89805f73f291df1e28fa05487fe (diff)
download: gn-ai-master.tar.gz
107 files changed, 6955 insertions, 0 deletions
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_1 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_1
new file mode 100644
index 0000000..677e295
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_1
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2018 -  Mechanisms of Vascular Aging.pdf",
+    "2016 - The dog aging project translational geroscience in companion.pdf",
+    "2015 - Cellular and Molecular Biology of Aging Endothelial Cells.pdf",
+    "2012 - Aging, Rejuvenation, and Epigenetic.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf",
+    "2015 - Cellular and Molecular Biology of Aging Endothelial Cells.pdf",
+    "2017 - Epigenetic aging signatures in mice livers.pdf",
+    "2017 - Diverse interventions that extend mouse.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf"
+  ],
+  "extraction_id": [
+    "3e65812c-453e-53aa-83ab-92f2ce15da29",
+    "2c1fcce1-b723-5f9f-8f66-49ed7895f2ac",
+    "86f9502b-7a3a-501f-9053-8af1d37043b4",
+    "d23b6aab-f299-5370-b3b6-0615112681f0",
+    "a47672ed-9f4d-5aa8-8b7e-f10753246a6e",
+    "42c88d1d-4bb6-50f8-9010-379e15650d96",
+    "0e789eef-b085-5fc2-b10a-8572bc28fa1b",
+    "5d4bf4c1-5bb4-5de6-a1bb-0485163a5373",
+    "d634b92e-0802-5ba8-a4c5-9e45462cd7d5",
+    "a47672ed-9f4d-5aa8-8b7e-f10753246a6e"
+  ],
+  "document_id": [
+    "659b84b6-63dd-5bb1-80ee-7478ed3c47e3",
+    "e841c6bd-78b8-56e1-b3dd-e2bcc8a0f590",
+    "815d7f3e-e219-502f-aba0-57a68ae787d3",
+    "bde26feb-f423-51b0-89ec-6f079bfc8b17",
+    "62b635c3-040e-512a-b016-6ef295308a1e",
+    "62b635c3-040e-512a-b016-6ef295308a1e",
+    "815d7f3e-e219-502f-aba0-57a68ae787d3",
+    "b20b11a6-1490-51b8-9218-c441a2e65ba7",
+    "dc7ad71a-a4d7-5901-a016-9a6fb2b91a2f",
+    "62b635c3-040e-512a-b016-6ef295308a1e"
+  ],
+  "id": [
+    "chatcmpl-ADZV87184EnuXO9GIujWS8NC7oWU2",
+    "57be0715-77c8-55e3-8239-56e1fa11a543",
+    "03e62089-fef5-5ed5-bf7f-36ff595fbaea",
+    "fe5b60e5-ded6-5950-bc1c-72cb39e16234",
+    "d7dcefa4-133c-594c-b8a8-38fe945c6b5c",
+    "907d7d31-04db-5f66-b390-7740142af182",
+    "40cbc230-7175-522e-b0ae-3901f2cfac0b",
+    "a9666b11-4567-52dd-90c8-be2238dafdcb",
+    "729598dc-94e6-5f52-ae19-071c959c7dd2",
+    "cbc86652-98e1-5464-a0ce-2272111246df",
+    "f8630239-fd67-5214-a5cd-f965d878f712"
+  ],
+  "contexts": [
+    "168. Yin L, Ye S, Chen Z, Zeng Y . Rapamycin preconditioning attenuates tran- sient focal cerebral ischemia/reperfusion injury in mice. Int J Neurosci. 2012;122:748756. doi: 10.3109/00207454.2012.721827  169. Spilman P, Podlutskaya N, Hart MJ, Debnath J, Gorostiza O, Bredesen  D, Richardson A, Strong R, Galvan V . Inhibition of mTOR by rapamy-cin abolishes cognitive deficits and reduces amyloid-beta levels in a  mouse model of Alzheimers disease. PLoS One. 2010;5:e9979. doi:  10.1371/journal.pone.0009979",
+    "Anisimov VN, Zabezhinski MA, Popovich IG, Piskunova TS, Semenchenko AV, Tyndyk ML, Yurova MN, Rosenfeld SV,Blagosklonny MV (2011b) Rapamycin increases lifespan and inhibits spontaneous tumorigenesis in inbred female mice. Cell Cycle 10:42304236 Augustine JJ, Bodziak KA, Hricik DE (2007) Use of sirolimus in solid organ transplantation. Drugs 67:369391 Bannister CA, Holden SE, Jenkins-Jones S, Morgan CL, Halcox JP,",
+    "ACCEPTED MANUSCRIPTACCEPTED MANUSCRIPT mTOR complex 2 (mTORC2), the less clearly identified  and less sensitive to rapamycin. Most information to  date on the r ole of mTOR has studied the insulin/nutrient signaling via the mTORC1 and significantly less in  known about the role of mTORC2 ( in this review, future references measure either mTORC1 or general mTOR  activity )[251]. Earlier this decade studies showed that decreasing TOR signaling, genetically or with rapamycin,",
+    "Harrison, D.E., Strong, R., Sharp, Z.D., Nelson, J.F., Astle, C.M., Flurkey, K.,Nadon, N.L., Wilkinson, J.E., Frenkel, K., Carter, C.S., et al. (2009). Rapamycin Cell148, January 20, 2012 2012 Elsevier Inc. 55",
+    "96. Lamming DW, Ye L, Katajisto P, Goncalves MD, Saitoh M, Stevens DM, etal. Rapamycin-  induced insulin resistance is mediated by mTORC2 loss and uncoupled from longevity.  Science. 2012;335:163843.  97. Tataranni T, Biondi G, Cariello M, Mangino M, Colucci G, Rutigliano M, etal. Rapamycin-  induced hypophosphatemia and insulin resistance are associated with mTORC2 activation  and klotho expression. Am J Transplant. 2011;11(8):165664.",
+    "ing these aspects in future studies on the effects of resveratrol could help to study in  greater depth the mechanisms of action of this compound [56].  Rapamycin Rapamycin is a macrolide isolated from Streptomyces hygroscopicus, a bacteria  from Pascua Island (Rapa Nui). It has functions as an antibiotic, an immune sup- pressant drug, and it is also proposed as a CRM.After the first studies, it was found  that rapamycin could induce the extension of the replicative life of yeast through the",
+    "[257] Wilkinson JE, Burmeister L, Brooks SV, Chan CC, Friedline S, Harrison DE, et al. Rapamycin slows  aging in mi ce. Aging Cell. 2012;11:675 -82.  [258] Selman C, Tullet JM, Wieser D, Irvine E, Lingard SJ, Choudhury AI, et al. Ribosomal protein S6 kinase 1  signaling regulates mammalian life span. Science. 2009;326:140 -4.  [259] Reihl K, Seals D, Henson G, LaRocca T, Mag erko K, Bosshardt G, et al. Dietary rapamycin selectively  improves arterial function in old mice. FASEB Journal. 2013;27:1194.17.",
+    "29. Wilkinson JE, Burmeister L, Brooks SV, Chan C-C, Friedline S, Harrison DE, et al. Rapamycin slows aging in mice. Aging Cell. 2012;11:675 82. 30. Lamming DW, Ye L, Katajisto P, Goncalves MD, Saitoh M, Stevens DM, et al. Rapamycin-induced insulin resistance is mediated by mTORC2 loss and uncoupled from longevity. Science. 2012;335:1638 43. 31. Zampieri M, Ciccarone F, Calabrese R, Franceschi C, Brkle A, Caiafa P. Reconfiguration of DNA methylation in aging. Mech Ageing Dev. 2015;151:60 70.",
+    "files [55, 62]. Of note, rapamycin in particular appears to induce additional changes u nrelated to age-associated changes. While both CR and rapamycin induced these non-age-related effects, this effect was much more marked for rapamycin. These non age-related epigenetic changes include gains of methylation at genes, enhancers and CpG islands and losses of methylation at genes and enhancers. Conceivably, such non age-related effects of rapamycin in",
+    "23  94. Chakrabarti P, English T, Shi J, Smas CM, Kandror KV .Mammalian target of rapamycin  complex 1 suppresses lipolysis, stimulates lipogenesis, and promotes fat storage. Diabetes.  2010;59:77581.  95. Miller RA, Harrison DE, Astle CM, Fernandez E, Flurkey K, Han M, et al. Rapamycin-  mediated lifespan increase in mice is dose and sex dependent and metabolically distinct from  dietary restriction. Aging Cell. 2014;13:46877."
+  ],
+  "answer": "Here are the studies that include rapamycin:  1. **Yin L, Ye S, Chen Z, Zeng Y (2012)**: \"Rapamycin preconditioning attenuates transient focal cerebral ischemia/reperfusion injury in mice\" [1]. 2. **Spilman P, Podlutskaya N, Hart MJ, Debnath J, Gorostiza O, Bredesen D, Richardson A, Strong R, Galvan V (2010)**: \"Inhibition of mTOR by rapamycin abolishes cognitive deficits and reduces amyloid-beta levels in a mouse model of Alzheimer's disease\" [1]. 3. **Anisimov VN, Zabezhinski MA, Popovich IG, Piskunova TS, Semenchenko AV, Tyndyk ML, Yurova MN, Rosenfeld SV, Blagosklonny MV (2011)**: \"Rapamycin increases lifespan and inhibits spontaneous tumorigenesis in inbred female mice\" [2]. 4. **Lamming DW, Ye L, Katajisto P, Goncalves MD, Saitoh M, Stevens DM, et al. (2012)**: \"Rapamycin-induced insulin resistance is mediated by mTORC2 loss and uncoupled from longevity\" [5]. 5. **Tataranni T, Biondi G, Cariello M, Mangino M, Colucci G, Rutigliano M, et al. (2011)**: \"Rapamycin-induced hypophosphatemia and insulin resistance are associated with mTORC2 activation and klotho expression\" [5]. 6. **Wilkinson JE, Burmeister L, Brooks SV, Chan CC, Friedline S, Harrison DE, et al. (2012)**: \"Rapamycin slows aging in mice\" [7]. 7. **Reihl K, Seals D, Henson G, LaRocca T, Magerko K, Bosshardt G, et al. (2013)**: \"Dietary rapamycin selectively improves arterial function in old mice\" [7]. 8. **Miller RA, Harrison DE, Astle CM, Fernandez E, Flurkey K, Han M, et al. (2014)**: \"Rapamycin-mediated lifespan increase in mice is dose and sex dependent and metabolically distinct from dietary restriction\" [10].  These references provide a comprehensive list of studies involving rapamycin.",
+  "question": "List as many studies as you can that include rapamycin."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_10 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_10
new file mode 100644
index 0000000..ff7c2dc
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_10
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2001 - The genetics of aging.pdf",
+    "2011 - A genome-wide association study confirms APOE as the major gene influencing.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2018 - Sex Differences in Aging Genomic Instability.pdf",
+    "2007 - Genetic correlates of brain aging on MRI and cognitive test measures a genome-wide association and linkage analysis in the Framingham study.pdf",
+    "2021 - Genome-wide association studies identify.pdf",
+    "2021 - Footprints in the Sand Deep Taxonomic Comparisons in Vertebrate Genomics to Unveil the Genetic Programs of Human Longevity.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf",
+    "2016 - Progress on the role of DNA methylation in aging.pdf"
+  ],
+  "extraction_id": [
+    "76bae746-eabf-51ed-a01f-d32ecc89c11b",
+    "210aa417-372c-5bf6-b961-e281a1817458",
+    "7ada6b55-99c2-5e20-bf96-d153f927256c",
+    "0104338d-cc9c-538f-be29-8343a64da37d",
+    "3f7edb2f-b8c8-511a-9fb3-c746b6f7f213",
+    "cd7730b6-22dc-5256-9310-79fc348b3226",
+    "34223e0e-590c-5f26-b120-b7250cd91b99",
+    "fd143ef8-c605-51d8-9969-4e681493b826",
+    "4ea8424f-1cd8-569c-a1df-3f0f54206e70",
+    "bcb3c620-b960-5af6-95ea-13215c31672e"
+  ],
+  "document_id": [
+    "aa9a9193-b6f3-5ef8-aefd-e01ec44abb46",
+    "63b27b06-db2c-5542-9b1a-cb9ebe64d339",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "8cfb5529-7f0c-58fc-b6e4-b3ee800fb72f",
+    "56b25b5a-fc9e-5d61-8502-1c110466ba16",
+    "60c2e869-1fee-53ea-b332-26d9c2abc747",
+    "0dc45abe-ab02-5b07-9916-7093b53323c0",
+    "62b635c3-040e-512a-b016-6ef295308a1e",
+    "62b635c3-040e-512a-b016-6ef295308a1e",
+    "e4cdc02f-4415-5638-aab8-f848b4d64a22"
+  ],
+  "id": [
+    "chatcmpl-ADZWBvqLDNy2CwnWNpuZ7tlbpCfWr",
+    "e09c33ea-4139-5cc2-9cf5-a40045f26a0c",
+    "2d0a20b8-4196-5451-9d99-282f82234464",
+    "3117c019-7311-53ae-8ab1-927ca822c709",
+    "a9434032-4a9d-54f8-a7a6-16110d1b3118",
+    "3ff3cf0b-c00f-5548-b4fb-21a57042c91c",
+    "9fa00091-9661-57bd-91c7-f0bf436805a7",
+    "f6ed1f0c-e4ea-5459-bb63-136decc11eee",
+    "12868be9-1c5a-57ea-a5e3-f521b977cef2",
+    "a0672677-71ad-5603-8427-a0648eec407f",
+    "c1b5a31a-066d-571b-af1f-db746d9d17f6"
+  ],
+  "contexts": [
+    "Recent developments on the genetics of aging can be seen as several streams of effort. In general, humans show a relatively modest ( <50%) heritability of",
+    "effect  genetic  variants  on  human  longevity.  Aging  2,  612620. Yu,  C.E.,  Seltman,  H.,  Peskind,  E.R.,  Galloway,  N.,  Zhou,  P.X.,  Rosenthal,  E.,  Wijsman, E.M.,  Tsuang,  D.W.,  Devlin,  B.,  Schellenberg,  G.D.,  2007.  Comprehensive  analysis of  APOE  and  selected  proximate  markers  for  late-onset  Alzheimers  disease: patterns  of  linkage  disequilibrium  and  disease/marker  association.  Genomics",
+    "It is undisputed that genetic factors influence aging. In a remarkable",
+    "males: what are the molecular and evolutionary causes? Aging Cell. 2007;6:225233. doi:10.1111/j.1474-9726.2007.00279.x  63. Benayoun BA, Pollina EA, Brunet A. Epigenetic regulation of ageing: link- ing environmental inputs to genomic stability. Nat Rev Mol Cell Biol. 2015;16:593610. doi:10.1038/nrm4048  64. Sen P, Shah PP, Nativio R, Berger SL. Epigenetic mechanisms of longevity  and aging. Cell. 2016;166:822839. doi:10.1016/j.cell.2016.07.050",
+    "Genet  1998, 81:92-97. 3. Pedersen NL, Posner SF, Gatz M: Multiple-threshold models for genetic influences on age of  onset for Alzheimer disease: findings in Swedish twins.   Am J Med Genet  2001, 105:724-728. 4. Gudmundsson H, Gudbjartsson DF, Frigge M, Gulcher JR, Stefansson K: Inheritance of human longevity in Iceland.   Eur J Hum Genet 2000, 8:743-749. 5. Flossmann E, Schulz UG, Rothwell PM: Systematic review of methods and results of studie s of the genetic epidemiology",
+    "population dynamics on the genetic architecture of human longevity. Aging (Albany NY). 2018;10(8):1947 63. 68. Bellenguez C, Kucukali F, Jansen I, Andrade V, Morenau-Grau S, Amin N, et al. Large meta-analysis of genome-wide association studies expands knowledge of the genetic etiology of Alzheimer disease and highlights potential translational opportunities. medRxiv. 2020. 69. Kojima T, Shimazui T, Hinotsu S, Joraku A, Oikawa T, Kawai K, et al. Decreased expression of CXXC4 promotes a",
+    "discover core mechanisms of regulation.ANALYSIS OF HUMAN VARIATION IN THE GENETIC CONTROL OF LONGEVITY Heritability studies have convincingly demonstrated that at least some fraction of human lifespan is heritable. In tandem, large-scale genome-wide association studies (GWAS) have identied numerous loci associated with age-related traits (Buniello et al., 2019). While genetic studies have functionally shown an inverse eect of multiple age-related, disease-",
+    "than in healthy elderly patients [71].  Concluding Remarks The study of the human aging process is complex and multifactorial, where genetic  and environmental variables are key players in its development. That is why we sug- gest a series of different biomarkers which include hormonal, inflammatory, and  oxidative stress biomarkers. However, it is possible that other biomarkers such as  DNA damage, telomere length determination, DNA repair mechanisms and p53",
+    "Clinical Genetics and Genomics of Aging",
+    "standing the cause and mechanisms of aging is imperative in assisting to suppress age-related diseases and promote healthylongevity. It is well-known that aging is influenced by a combin- ation of genetic and environmental factors. Previous twin stud- ies have shown that the genetic contribution to general human longevity is about 2030% [ 4,5], whereas environmental factors in human aging and longevity still account for the largest effect. Epigenetic factors influence the regulation of gene expres-"
+  ],
+  "answer": "Genetic factors play a significant role in influencing aging in humans. Here are some key points based on the provided context:  1. **Heritability of Longevity**: Human longevity shows a relatively modest heritability of less than 50% [1]. Twin studies have estimated the genetic contribution to general human longevity to be about 20-30% [10].  2. **Genome-Wide Association Studies (GWAS)**: Large-scale GWAS have identified numerous loci associated with age-related traits, indicating that specific genetic variations can influence aging [7].  3. **Specific Genetic Variants**: The APOE gene and its variants have been comprehensively analyzed for their association with late-onset Alzheimer's disease, which is an age-related condition [2]. This suggests that certain genetic markers are linked to age-related diseases.  4. **Epigenetic Regulation**: Epigenetic mechanisms, which involve changes in gene expression without altering the DNA sequence, also play a crucial role in aging. Environmental inputs can affect genomic stability through epigenetic regulation [4].  5. **Inheritance Studies**: Studies on the inheritance of human longevity, such as those conducted in Iceland, have provided insights into the genetic factors that contribute to a longer lifespan [5].  In summary, aging in humans is influenced by a combination of genetic factors, including specific genetic variants, heritability, and epigenetic regulation [1], [2], [4], [5], [7], [10].",
+  "question": "What genetic factors influence aging in humans?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_11 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_11
new file mode 100644
index 0000000..c32fefd
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_11
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2010 - Do different neurons age differently Direct genome-wide analysis of aging in single identified cholinergic neurons.pdf",
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2012 - Human Ageing Genomic Resources Integrated.pdf",
+    "2018 - Biological Processes Modulating Longevity across Primates.pdf",
+    "2011 - Genetics and genomics of human ageing.pdf",
+    "2010 - Genetics and genomics of human ageing.pdf",
+    "2008 - Estrogen, not intrinsic aging, is the major regulator of delayed.pdf",
+    "2013 - Gene expression changes with age in skin.pdf"
+  ],
+  "extraction_id": [
+    "52c67b46-63f2-54ae-a78e-e9d54a55f6e4",
+    "d59d7882-333d-5576-86ab-3cfa6354b946",
+    "81c68113-aa96-5af3-b4fc-5898fa20e379",
+    "d59d7882-333d-5576-86ab-3cfa6354b946",
+    "25e9d8a3-54ac-5412-8efb-3b56d93f363f",
+    "c07d6709-8dbe-5437-b7df-0849b92c0ea0",
+    "07a34581-749c-5556-bdea-806b2c9c7915",
+    "59227f74-f1c7-58ad-a886-aa9e3799a132",
+    "eeffae01-ce08-54a8-955f-6f0c9d07eedc",
+    "dfb687b2-f1ff-5e22-8a67-4a1db9ebeb3c"
+  ],
+  "document_id": [
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "153b070f-0291-5ed4-ad33-edea5e3fa8f7",
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "5f554cc7-c94d-5fbd-9567-528499663ed6",
+    "930103c1-e98e-524c-aa68-233a45dc6726",
+    "08eee102-d627-5f1b-84c7-603c38981adf",
+    "633f3149-e966-53ef-aa7d-b759398ed541",
+    "04a3d8f1-64c1-5e25-ab0a-3eb749c06c92",
+    "5c121bbb-57b8-51cc-8461-effa1bfd87b9"
+  ],
+  "id": [
+    "chatcmpl-ADZWHUX5oZWH5Bj3eh2vkudPOLcus",
+    "8fd5ab85-67ed-55e6-bbfa-09436c4fdbfb",
+    "413f8f54-b5cc-5089-9f5c-d9e3b8bcf594",
+    "6f04401a-b938-5a60-8b69-d37f9086748c",
+    "786d2756-4c4d-5ac0-8d3d-63f914d51664",
+    "0ae63c75-df5f-59b0-9561-30d5115f0f74",
+    "f2fbfb29-0a51-5f94-8b67-d47ab4de68bd",
+    "fd6cfc2c-76b1-5620-a68c-fb37db9b6f78",
+    "df45a752-e866-54bb-ab49-daff9a702eef",
+    "66f72bdc-d38b-5c7a-afdd-4c7549ce2131",
+    "d53018ae-0881-5ef4-9c49-48623e8aa342"
+  ],
+  "contexts": [
+    "www.ncbi.nlm.nih.gov/homologene)  of genes strongly asso-ciated with aging in model organisms. Also included are genesin which mutations result in segmental progeroid syndromes,such as the Werners syndrome gene, as well as genes criticalin pathways previously related to aging, such as the insulin/insulin-like signalling pathway (de Magalhes   et     al  ., 2005a). The",
+    "overexpressed with age seem to be a response to aging,in that they have been previously found to have protec-tive functions (de Magalha es et al., 2009b). As such,these genes may help organisms manage aging andcould be targets for manipulation. Likewise, gene ex-pression analysis of CR has been conducted to identifyassociated genes (Lee et al., 1999, 2000). A number ofmolecular signatures have emerged from such studiesthat could be useful to identify candidate processes andpathways that affect aging,",
+    "OTHER AGING RELATED GENES",
+    "In addition to aging- and CR-related genes, another source of candidate genes and pathways for drug designare human longevity-associated genes (Barzilai andShuldiner, 2001; Browner et al., 2004; Kenyon, 2010).Dozens of genes have now been associated with humanlongevity (de Magalha es et al., 2009a), although only ahandful of genes have been shown to have consistenteffects across populations. Many longevity-associated genes are related to spe-",
+    "potentially associated with human ageing. For eachgene, a description compiled from the studies that linkthe gene to ageing is provided. It should be noted thatour focus is on genes that might affect the ageingprocess, rather than individual age-related pathologies; genes affecting multiple, even if not all, age-related",
+    "Pleiotropies and Aging-Related Genesets To study genes that have been previously related to aging, a list of curated human genes associated with aging in different model systems was obtained from the GenAge data set ( de Magalh ~aes et al. 2005 ). We used gene ontology (GO) anno-",
+    "aging in human muscle reveals a common aging signa-ture. PLoS Genet. 2, e115. ( doi:10.1371/journal.pgen. 0020115 ) 64 Lener, T ., Moll, P . R., Rinnerthaler, M., Bauer, J., Aberger, F. & Richter, K. 2006 Expression proling ofaging in the human skin. Exp. Gerontol. 41, 387397. (doi:10.1016/j.exger.2006.01.012 ) 65 Kim, S. K. 2008 Genome-wide views of aging gene net- works . Molecular Biology of Aging Monograph 9. Cold Spring Harbor, CT: Cold Spring Harbor LaboratoryPress.",
+    "aging in human muscle reveals a common aging signa-ture. PLoS Genet. 2, e115. ( doi:10.1371/journal.pgen. 0020115 ) 64 Lener, T ., Moll, P . R., Rinnerthaler, M., Bauer, J., Aberger, F. & Richter, K. 2006 Expression proling ofaging in the human skin. Exp. Gerontol. 41, 387397. (doi:10.1016/j.exger.2006.01.012 ) 65 Kim, S. K. 2008 Genome-wide views of aging gene net- works . Molecular Biology of Aging Monograph 9. Cold Spring Harbor, CT: Cold Spring Harbor LaboratoryPress.",
+    "tive-gerontogenes and genes with established aging-relatedfunctions were identified by interrogation of the GenAgeonline database [12], from aging-associated Gene Ontology( G O )  g r o u p s  a n d  f r o m  h a n d  a n n o t a t i o n  ( s e e  M a t e r i a l s  a n dmethods/Results for a detailed description of the analysis). We show that the fundamenta l changes in genes and proc-",
+    "on model organisms [3] or have been confined to specificaging-associated disorders such as progeria syndromes [4]. A study of postmortem human brain tissue from 30 individuals aged 26 to 106 years [5] showed that approxi- mately 4% of approximately 11,000 genes analyzed show a significant age-related expression change (1.5-fold or more) in individuals aged >40 years. These genes were reported to play central roles in synaptic plasticity, vesi- cular transport, and mitoch ondrial function. Another"
+  ],
+  "answer": "Several genes have been associated with aging. Here are some key points from the provided context:  1. **Genes in Model Organisms**: A list of genes strongly associated with aging in model organisms includes those involved in segmental progeroid syndromes, such as the Werner syndrome gene, and genes critical in pathways previously related to aging, such as the insulin/insulin-like signaling pathway [1].  2. **Gene Expression Analysis**: Studies have identified genes overexpressed with age that have protective functions, suggesting they help manage aging and could be targets for manipulation. Gene expression analysis of caloric restriction (CR) has also been conducted to identify associated genes [2].  3. **Human Longevity-Associated Genes**: Dozens of genes have been associated with human longevity, although only a handful have shown consistent effects across populations [4].  4. **GenAge Data Set**: A curated list of human genes associated with aging in different model systems is available from the GenAge data set [6].  5. **GenAge Online Database**: Genes with established aging-related functions were identified by interrogation of the GenAge online database, aging-associated Gene Ontology groups, and hand annotation [9].  These references collectively highlight the involvement of various genes and pathways in the aging process.",
+  "question": "what genes are associated with aging?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_12 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_12
new file mode 100644
index 0000000..73a4c9b
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_12
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2010 - Do different neurons age differently Direct genome-wide analysis of aging in single identified cholinergic neurons.pdf",
+    "2011 - A genome-wide association study of aging.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2018 - Biological Processes Modulating Longevity across Primates.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2010 - Genetics and genomics of human ageing.pdf",
+    "2011 - Genetics and genomics of human ageing.pdf"
+  ],
+  "extraction_id": [
+    "d59d7882-333d-5576-86ab-3cfa6354b946",
+    "52c67b46-63f2-54ae-a78e-e9d54a55f6e4",
+    "81c68113-aa96-5af3-b4fc-5898fa20e379",
+    "a5be18f8-c263-5635-87d7-57c5addd65e5",
+    "52c67b46-63f2-54ae-a78e-e9d54a55f6e4",
+    "c07d6709-8dbe-5437-b7df-0849b92c0ea0",
+    "52c67b46-63f2-54ae-a78e-e9d54a55f6e4",
+    "52c67b46-63f2-54ae-a78e-e9d54a55f6e4",
+    "59227f74-f1c7-58ad-a886-aa9e3799a132",
+    "07a34581-749c-5556-bdea-806b2c9c7915"
+  ],
+  "document_id": [
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "153b070f-0291-5ed4-ad33-edea5e3fa8f7",
+    "8e9c1150-1047-54a2-bf85-1cc5000a6811",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "930103c1-e98e-524c-aa68-233a45dc6726",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "633f3149-e966-53ef-aa7d-b759398ed541",
+    "08eee102-d627-5f1b-84c7-603c38981adf"
+  ],
+  "id": [
+    "chatcmpl-ADZWOuZDmIcGuvC8wjb6oX7vSBFDg",
+    "786d2756-4c4d-5ac0-8d3d-63f914d51664",
+    "a21de3e8-ed2c-5c06-a351-ccb8f92f4e21",
+    "6f04401a-b938-5a60-8b69-d37f9086748c",
+    "06e319e1-b054-5f33-9b40-ee892f507736",
+    "9defe0af-80a1-56da-90df-551fd55baa13",
+    "f2fbfb29-0a51-5f94-8b67-d47ab4de68bd",
+    "8fd5ab85-67ed-55e6-bbfa-09436c4fdbfb",
+    "650300e1-898c-56e2-9358-0bb6625b0073",
+    "df45a752-e866-54bb-ab49-daff9a702eef",
+    "fd6cfc2c-76b1-5620-a68c-fb37db9b6f78"
+  ],
+  "contexts": [
+    "In addition to aging- and CR-related genes, another source of candidate genes and pathways for drug designare human longevity-associated genes (Barzilai andShuldiner, 2001; Browner et al., 2004; Kenyon, 2010).Dozens of genes have now been associated with humanlongevity (de Magalha es et al., 2009a), although only ahandful of genes have been shown to have consistenteffects across populations. Many longevity-associated genes are related to spe-",
+    "GenAge features a data set of genes that may regulate agingin humans or that at least appear to be considerably associated with the human aging phenotype. This data set includes orthologues  derived from established databases, mainly In-Paranoid (OBrien   et     al  ., 2005) but also HomoloGene (http://",
+    "OTHER AGING RELATED GENES",
+    "processes in human longevity and aging. Ten of the 22 suggestive associations identied in our analyses are in ornear genes that are highly expressed in the brain (HECW2[Rotin and Kumar, 2009], HIP1 [Blanpied et al., 2003], BIN2, GRIA1), were previously related to the regulation of neuronal excitability and plasticity (KCNQ4 [Van Eyken et al., 2006], LMO4 [Joshi et al., 2009; Leuba et al., 2004],",
+    "genes analyzed for their possible association with human lon-gevity (http://genomics.senescence.info/genes/longevity.html).All longevity association studies in humans we could find by thetime of the latest update were added to this list. These includestudies reporting negative results, which we see as essentialsince many genes display population-specific associations withlongevity. Fig. 1 From the main page of the Human Ageing",
+    "Pleiotropies and Aging-Related Genesets To study genes that have been previously related to aging, a list of curated human genes associated with aging in different model systems was obtained from the GenAge data set ( de Magalh ~aes et al. 2005 ). We used gene ontology (GO) anno-",
+    "www.ncbi.nlm.nih.gov/homologene)  of genes strongly asso-ciated with aging in model organisms. Also included are genesin which mutations result in segmental progeroid syndromes,such as the Werners syndrome gene, as well as genes criticalin pathways previously related to aging, such as the insulin/insulin-like signalling pathway (de Magalhes   et     al  ., 2005a). The",
+    "shown that genes associated with aging and/or longevity inmodel organisms are evolutionary conserved in terms of havingmore homologues than predicted by chance (Budovsky   et     al  .,2007, 2008) and exhibiting slower molecular evolution rates (de Magalhes & Church, 2007). Therefore, it is now clear that atleast some genes identified in model organisms may be relevantto human aging. To allow researchers to focus specifically on human aging,",
+    "aging in human muscle reveals a common aging signa-ture. PLoS Genet. 2, e115. ( doi:10.1371/journal.pgen. 0020115 ) 64 Lener, T ., Moll, P . R., Rinnerthaler, M., Bauer, J., Aberger, F. & Richter, K. 2006 Expression proling ofaging in the human skin. Exp. Gerontol. 41, 387397. (doi:10.1016/j.exger.2006.01.012 ) 65 Kim, S. K. 2008 Genome-wide views of aging gene net- works . Molecular Biology of Aging Monograph 9. Cold Spring Harbor, CT: Cold Spring Harbor LaboratoryPress.",
+    "aging in human muscle reveals a common aging signa-ture. PLoS Genet. 2, e115. ( doi:10.1371/journal.pgen. 0020115 ) 64 Lener, T ., Moll, P . R., Rinnerthaler, M., Bauer, J., Aberger, F. & Richter, K. 2006 Expression proling ofaging in the human skin. Exp. Gerontol. 41, 387397. (doi:10.1016/j.exger.2006.01.012 ) 65 Kim, S. K. 2008 Genome-wide views of aging gene net- works . Molecular Biology of Aging Monograph 9. Cold Spring Harbor, CT: Cold Spring Harbor LaboratoryPress."
+  ],
+  "answer": "Several genes have been associated with aging in humans according to the provided context:  1. **GenAge Data Set**: This data set includes genes that may regulate aging in humans or are considerably associated with the human aging phenotype [2]. 2. **HECW2, HIP1, BIN2, GRIA1, KCNQ4, LMO4**: These genes are highly expressed in the brain and have been previously related to the regulation of neuronal excitability and plasticity [4]. 3. **Werners Syndrome Gene**: Mutations in this gene result in segmental progeroid syndromes, which are critical in pathways previously related to aging, such as the insulin/insulin-like signaling pathway [7].  These references indicate that there are multiple genes associated with aging in humans, with some being highly expressed in specific tissues like the brain and others being involved in critical aging-related pathways.",
+  "question": "Which genes are associated with aging in human"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_13 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_13
new file mode 100644
index 0000000..27527e2
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_13
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2012 - Genome-Scale Studies of Aging Challenges and Opportunities.pdf",
+    "2020 - A multidimensional systems biology.pdf",
+    "2020 - Gene network a completely updated tool for systems genetics analyses.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2018 - Human Ageing Genomic Resources new and updated.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2012 - Human Ageing Genomic Resources Integrated.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf"
+  ],
+  "extraction_id": [
+    "aecbe8a8-aeed-5cfa-b0f3-be29f19d849d",
+    "53c57cc4-4d43-505a-974c-442d06e144df",
+    "fe4ec57e-6ae7-59c4-b8fa-da73fe77ce96",
+    "858f630f-9443-5f13-ac40-8e16eadd9ba1",
+    "6b10898e-0906-5fff-9c70-b3be2d562fda",
+    "03c88365-c56c-56f2-a15f-e183398d3dfe",
+    "7ada6b55-99c2-5e20-bf96-d153f927256c",
+    "aecbe8a8-aeed-5cfa-b0f3-be29f19d849d",
+    "25e9d8a3-54ac-5412-8efb-3b56d93f363f",
+    "aecbe8a8-aeed-5cfa-b0f3-be29f19d849d"
+  ],
+  "document_id": [
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "b77aace0-fa36-5fd4-8e2a-c8932198acd1",
+    "d040bfe3-e409-5b5c-b8f8-f3dd4fc060e3",
+    "128224f1-3545-52c3-93cb-77c3cf4ec70a",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "82726cea-f77c-5a92-9f2e-ecccc369953a",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "5f554cc7-c94d-5fbd-9567-528499663ed6",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529"
+  ],
+  "id": [
+    "chatcmpl-ADZWTp42DWHZeK1fZT0MSpkOitZfP",
+    "496d27de-6dd0-5f6a-bedb-64d4c252981d",
+    "df726361-271a-5dbb-b6d1-03dab5a63006",
+    "9716c2c9-6f43-57f2-bad4-6d96c82d5c16",
+    "c63cfaee-749e-547b-9c0a-086266f10670",
+    "42464f0d-d8ce-5f73-9c7c-0cdec45e7f4f",
+    "3153cd1e-de1c-52fb-aede-4065019d8c6b",
+    "676b5bff-01e8-58cf-93e5-ac14d8e82760",
+    "4c4f5670-cb9a-59b5-b9cc-ba5bce662035",
+    "cf8bf1ec-4919-59b2-a60d-183fc5a04bb0",
+    "1d7f120f-20c4-5d6c-983f-41534fb30503"
+  ],
+  "contexts": [
+    "the different pathways linked with aging and even study genenetworks. In such works, GenAge is an adequate resource asit provides a framework for the functional genomics of aging.For example, Xue   et     al  . (2007) used GenAge to construct a modular network of aging and obtain insights into aging, including thefact that genes connecting different modules are more likely toaffect longevity and/or aging, an hypothesis the authors validatedexperimentally in worms (Xue   et     al",
+    "[111], and for generation of networks based on known gene  interactions such as GeneMania [112] and Cytoscape [113], as well as for identifying cross-species orthology relation-ships [114], network-based thinking has been increasingly applied to the study of aging and lifespan [115-118]. Re-cently, the novel computational method of network identifi- cation by regression (NIR) [119] has been used to identify",
+    "networks can be built using protein interaction and gene co-expression data. A previous paper used protein- protein interactions to build genetic networks identifying potential longevity genes along with links between genes and aging-related diseases [ 30]. Here, we present the network of proteins and genes co-expressed with the CellAge senescence genes. Assaying the networks, we find links between senescence and immune system func- tions and find genes highly connected to CellAge genes",
+    "GeneNetwork have reinvigorated it, including the addition  of data from  10 species, multi -omics  analysis, updated code, and new tools. The new GeneNetwork is now an exciting resource for  predictive medicine and systems genetics, which is constantly being maintained and improved.    Here, we give a brief overview of the process  for carrying out some of the most common  functions on GeneNetwork, as a gateway to deeper analyses , demonstrating how a small",
+    "of GenAge involved finding novel genes that may be linked toaging by way of an analysis of proteinprotein interactions. Theprinciple being that proteins not previously thought to berelated to aging which interact with a large number of proteinsdirectly linked to aging might too be involved in aging and arethus promising candidates for future studies (de Magalhes &Toussaint, 2004; Budovsky   et     al  ., 2007). Similar works are made",
+    "2009, with over 400 genes added in the current update (Ta-ble1), includingmiRNAs for thefirst time. GenAge has proven a valuable resource for ageing re- search, as evidence by many publications. A systems levelanalysis of the GenAge human genes database identified a robust group of ageing-specific network characteristics, re- vealingageinggenesasnetworkhubs( 11).Moreover,inan analysis of genes in the ageing human brain, 54 genes with sustained, consistent expression and 23 genes with DNA",
+    "a curated database of genes potentiallyassociated with human aging, and a list of genes testedfor their association with human longevity. A myriad ofbiological data and information is included for hundredsof genes, making GenAge a reference for research thatreflects our current understanding of the genetic basis ofaging. GenAge can also serve as a platform for thesystems biology of aging, and tools for the visualizationof proteinprotein interactions are also included. AnAgeis a database of aging in",
+    "et     al  ., 2007). In a sense, GenAge offers an overall view of what is presently known about thegenetics of aging in model organisms and in humans that canbe used for numerous studies, including in contemporary functionalgenomics and systems biology methods.Table 2 Criteria used to select entries for inclusion in the GenAge human data set Main reason for selectionNumber  of genes Evidence directly linking the gene product to aging in humans 3",
+    "senescence.info/genes/) is to host high-quality curatedgene-centric information relevant to human ageing.Although initially GenAge was designed to include onlyhuman genes potentially associated with ageing, thedatabase has signicantly grown since, and several newgene sets have been added to it. For example, GenAge includes, since 2008, a list of genes from model organisms based on genetic manipulation experiments (2). Currently, the database is divided into three main",
+    "information source on the genetics of aging. In particular forthe human data set, a wealth of biological data is provided foreach entry, including relevant information in the context ofbiogerontology (see example below) to an extent that is notavailable in larger, more generic databases like Entrez Gene.Therefore, to learn about the involvement of a given gene inaging, a quick search in GenAge is the best place to start.GenAge features a sophisticated search engine and its user-friendly interface is easy"
+  ],
+  "answer": "GeneNetwork is a resource that has been significantly updated and enhanced to include data from 10 species, multi-omics analysis, updated code, and new tools. It serves as an exciting resource for predictive medicine and systems genetics, constantly being maintained and improved [4].  In relation to aging research, GeneNetwork is used to study genetic networks and pathways linked with aging. For example, researchers use GeneNetwork to construct modular networks of aging, which can provide insights into how different genes interact and affect longevity and aging processes [1]. This network-based approach allows for the identification of potential longevity genes and the links between genes and aging-related diseases [3]. Thus, GeneNetwork plays a crucial role in the functional genomics of aging by enabling the analysis and visualization of complex genetic interactions and their implications for aging and longevity.",
+  "question": "What is GeneNetwork and how does it relate to aging research?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_2 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_2
new file mode 100644
index 0000000..2e25fb7
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_2
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2002 - Pharmacology, Genomics, and the Evolutionary Biology.pdf",
+    "2001 - A genome-wide scan for linkage to human.pdf",
+    "2010 - Genetics and genomics of human ageing.pdf",
+    "2011 - Genetics and genomics of human ageing.pdf",
+    "2023 - A transcriptome-based single-cell biological age model.pdf",
+    "2011 - A genome-wide association study of aging.pdf",
+    "2021 - Footprints in the Sand Deep Taxonomic Comparisons in Vertebrate Genomics to Unveil the Genetic Programs of Human Longevity.pdf",
+    "2002 - Pharmacology, Genomics, and the Evolutionary Biology.pdf",
+    "2011 - Genomics of human longevity.pdf",
+    "2011 - A genome-wide association study of aging.pdf"
+  ],
+  "extraction_id": [
+    "aa03a9d5-4e30-5fb0-bee1-6dd8e6a549b3",
+    "17246c43-2e44-579b-867d-3dc7150ceedd",
+    "04babc6e-5138-5804-a150-70254859800d",
+    "27e291f1-e6bf-5e76-9245-522de74ea63b",
+    "0fd46f00-d3e1-54f4-9395-6c3e8294ed51",
+    "253a4339-29d4-58c2-8a01-5137d94873b6",
+    "34223e0e-590c-5f26-b120-b7250cd91b99",
+    "e501662f-ffca-563b-97a7-b682a5d7f6ba",
+    "7b101eb3-7990-5345-b510-c0be15f063a7",
+    "beab62d0-2e6f-5d77-b0a2-7375a9ed1364"
+  ],
+  "document_id": [
+    "1bc636a3-6ce0-5fea-b549-0dae90a78f1b",
+    "1431984a-82d9-51d4-a23c-5f76a02ab554",
+    "633f3149-e966-53ef-aa7d-b759398ed541",
+    "08eee102-d627-5f1b-84c7-603c38981adf",
+    "9be234b7-f37d-5cd5-8895-bfe676441b2f",
+    "8e9c1150-1047-54a2-bf85-1cc5000a6811",
+    "0dc45abe-ab02-5b07-9916-7093b53323c0",
+    "1bc636a3-6ce0-5fea-b549-0dae90a78f1b",
+    "2e038219-fdaa-506f-9cd3-51379054130e",
+    "8e9c1150-1047-54a2-bf85-1cc5000a6811"
+  ],
+  "id": [
+    "chatcmpl-ADZVIyiCYn4oPG1At1d3sSBHcOoYZ",
+    "bf384c33-974b-57c3-867d-3515e1d45c49",
+    "e0cce1c5-8709-5218-99b6-48a6ba242931",
+    "62e2bf90-fdb9-5499-a063-cee6c92feb40",
+    "25a0cb1d-0207-5197-9b6a-389b16c1f17e",
+    "9f9fef49-0bda-5948-93bd-0f8f43bbefdf",
+    "f1f870c1-b1ed-5eeb-8831-3484d35414b2",
+    "f6ed1f0c-e4ea-5459-bb63-136decc11eee",
+    "86393802-9171-57d8-806d-6d2ccfb3f0b2",
+    "45e52016-b6ef-5efb-a9e8-5a88341f3300",
+    "146229ff-4d17-5319-88e4-6040f30cf0c1"
+  ],
+  "contexts": [
+    "that is differentiated at hundreds of loci. Many ofthe loci that control aging in Drosophila will not have the same effect on human aging. On the other hand,we expect that other loci will work in a parallelmanner in humans. We have no way of knowing a priori which group any particular locus will belong in. Thus, the individual mutants that increase Drosophila lifespan may or may not come from loci",
+    "effect fundamental mechanisms of aging (14, 16). The drawbacksof such studies include the improbability of picking the right geneto study the myriad of known and unknown genes affecting theprocess of interest (17). The linkage study described heremarkedly improves the efficiency of such association studies bydefining a region likely to contain polymorphism(s) with signif-icant influence on life span. Additional association studies with these families and repli-",
+    "understanding of molecular mechanisms underlyingthe human ageing process. Like other complexhuman traits, nding common variants that accountfor the entire genetic component of human lifespan variability has proved difcult. If rare variants rather than common variants explain most of the genetic vari-ation in ageing among humans, new genotypingtechniques and new analysis methods must be devel-oped to nd genes and pathways involved in ageing.Next-generation sequencing technologies are faster",
+    "understanding of molecular mechanisms underlyingthe human ageing process. Like other complexhuman traits, nding common variants that accountfor the entire genetic component of human lifespan variability has proved difcult. If rare variants rather than common variants explain most of the genetic vari-ation in ageing among humans, new genotypingtechniques and new analysis methods must be devel-oped to nd genes and pathways involved in ageing.Next-generation sequencing technologies are faster",
+    "Map contains 1119 and 1459 curated human and mouse aginggenes, respectively, covering almost all scales of aging, rangingfrom molecular damage to genetic predisposition. Cross-speciescomparison revealed a modest overlap between known humanand mouse aging genes, suggesting both conservation of core sen- escence pathways and fundamental differences in aging between mice and humans (Fig. 2E). Aging-associated genes can alternatively be identified in a",
+    "Several explanations are possible for the lack of genome- wide signicant ndings. First, mortality is arguably 1 ofthe most complex phenotypes, and several trajectories to-ward extreme old age have been identied (Evert et al.,2003). Multiple genes could mediate the aging process butwould have their effects through numerous different patho-physiological processes and diseases that act as intermediate",
+    "discover core mechanisms of regulation.ANALYSIS OF HUMAN VARIATION IN THE GENETIC CONTROL OF LONGEVITY Heritability studies have convincingly demonstrated that at least some fraction of human lifespan is heritable. In tandem, large-scale genome-wide association studies (GWAS) have identied numerous loci associated with age-related traits (Buniello et al., 2019). While genetic studies have functionally shown an inverse eect of multiple age-related, disease-",
+    "[12]More than 1000 loci exhibit age-dependent changes in geneexpression (1264 genes). This is a substantialproblem, because not all of these loci will be causally involved in aging, and there are so many to sort out. An additional application of gene chip technologyis to compare ies with and without a lifespanmodulating physiological treatment. Pletcher et al.",
+    "such alleles. The frequency of genetic variants wastypically compared between highly aged cases andyoung controls, revealing loci at which genetic variantsmay contribute to a higher or lower probability ofsurvival into old age. So far, this approach hasmainly been applied to study single candidate genessuch as the mammalian orthologues of loci in IIS sig-nalling pathways that emerged from lifespan extensionstudies in animal models. An interesting observationthat needs to be taken into human studies is the",
+    "Kenyon, 2010; Vellai et al., 2003 ). However, in humans, common variants within genes involved in these pathways have not been consistently associated with lifespan ( Chris-tensen et al., 2006; Kenyon, 2010; Kuningas et al., 2008; Vijg and Suh, 2005 ). The lack of success in the identication of genes related to aging in humans may be due to the complexity of the phenotype. One approach to investigate aging and longevity is to compare frequencies of genetic variants between no-"
+  ],
+  "answer": "Mapping gene loci that control aging in humans is difficult due to several reasons:  1. **Complexity of the Phenotype**: Aging is a highly complex phenotype influenced by numerous genes and their interactions with environmental factors. This complexity makes it challenging to identify specific genetic loci that control aging [10].  2. **Multiple Pathophysiological Processes**: Multiple genes could mediate the aging process through various pathophysiological processes and diseases, adding layers of complexity to the genetic mapping [6].  3. **Rare vs. Common Variants**: Finding common genetic variants that account for the entire genetic component of human lifespan variability has proven difficult. If rare variants explain most of the genetic variation in aging, new genotyping techniques and analysis methods are required to identify these genes and pathways [3], [4].  4. **High Number of Loci**: There are more than 1000 loci that exhibit age-dependent changes in gene expression. Not all of these loci are causally involved in aging, making it a substantial challenge to sort out which ones are relevant [8].  5. **Species Differences**: Many loci that control aging in model organisms like Drosophila may not have the same effect on human aging. This lack of direct correlation between species complicates the identification of relevant human aging genes [1].  6. **Heritability and Genetic Variation**: While some fraction of human lifespan is heritable, the genetic variation contributing to aging is complex and involves numerous loci associated with age-related traits, further complicating the mapping process [7].  These factors collectively contribute to the difficulty in mapping gene loci that control aging in humans.",
+  "question": "Why is it so diffuclut to map gene loci that control aging in humans?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_3 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_3
new file mode 100644
index 0000000..1d57222
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_3
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2020 - Protecting the Aging Genome.pdf",
+    "2013 - Pathways, Networks and Systems Medicine Conferences.pdf",
+    "2012 - Pleiotropic Cellular Functions of PARP1 in Longevity.pdf",
+    "2008 - Biotools for Determining the Genetics of Susceptibility to Infectious Diseases.pdf",
+    "2008 - (Infectious Disease) Karl A. Western (auth.), Vassil St. Georgiev PhD, Karl A. Western MD, John J. McGowan PhD (eds.) - National Institute of Allergy and Infectious Diseases, NIH_ Frontiers in Researc (3).pdf",
+    "1999 - The NOD mouse model of type 1 diabetes.pdf",
+    "2012 - Genome-Wide Analysis of Yeast Aging.pdf",
+    "2005 -Liang- GENETIC REGULATION OF HEMATOPOIETIC STEM CELL NUMBERS IN MICE.pdf",
+    "2005 - GENETIC REGULATION OF HEMATOPOIETIC STEM CELL NUMBERS IN MICE.pdf",
+    "2006 - Molecular pathogenesis of thyroid cancer the significance.pdf"
+  ],
+  "extraction_id": [
+    "58c6c8e0-734b-539d-8e50-fd3cb02f650e",
+    "ee9fd19c-ae3c-5da6-9fcd-264bafc68b55",
+    "254dda83-4350-5b57-b6e4-638addaf7ce3",
+    "30fc6495-2cc4-5c3a-9d49-555478243db1",
+    "630a9f0d-d04f-581b-a9a0-3d2de4fff6fe",
+    "4c08454a-1c63-52e3-b382-0a33cd46b523",
+    "5436985c-1a11-544d-b935-fe68ee75a956",
+    "20ef68d8-5bd7-5ed1-937f-4be14c6ce1b0",
+    "70332eb3-8348-53b6-abd6-724515f363db",
+    "48a746ad-a07a-5c53-89dc-3c2768900f0d"
+  ],
+  "document_id": [
+    "bb774030-2570-5596-b2ab-b8f57ff81086",
+    "b50a9732-7d01-5d4d-8f33-a9d43dbc7df3",
+    "e67324c0-474b-5280-8cbc-3778c6c0e5f0",
+    "fcbbb3ce-6524-50e3-9f8d-c191dc551231",
+    "4db8c752-c8e2-5f6d-a091-dc4f1d0c48bc",
+    "cc139813-6771-5434-b948-381291c86509",
+    "cf24db9a-e013-5780-8b0f-369c56143f29",
+    "6f3d464d-8df0-560e-b579-942810e1107c",
+    "815a8b4a-6902-5bb9-87e0-563e1ba7a38a",
+    "d05e3aba-f8c1-5c5c-afff-679fa14b9a16"
+  ],
+  "id": [
+    "chatcmpl-ADZVPjiJlVpPgOZJvVtmrRMtG8Eg7",
+    "dfd32439-5b44-5e43-a85b-6dd58810b9ed",
+    "c02a78d4-b932-5d71-b183-8b1965fef470",
+    "fdfc9b00-1bce-5f6b-b20f-c516c7b5448a",
+    "2258748b-d21f-577b-a1f8-0ba4f61b6e30",
+    "de267316-5a20-5a54-b22f-30c8e0bf426e",
+    "31910427-870d-5c8c-846f-d355211c632a",
+    "ee2ce54b-850a-5d36-8781-f8b23585f97d",
+    "c48f36fa-a9c7-5b9c-a7d1-26850026f3a3",
+    "3a15c325-3b6e-54ff-a58a-0e61631ed073",
+    "fe4906aa-37b1-5514-962c-1e8dc5b2fb13"
+  ],
+  "contexts": [
+    "Cell Death A form of programmed cell death, apoptosis is necessary for normal cell turnover and is essential to a plethora of other biological processes. Apoptosis can be executed via Bcl-2 activation of caspases, via signals from the death receptor on the plasma membrane, or via induction by granzyme Bsecreted from cytotoxic T cells (Tc cells) [ 35]. Endonucleases and proteases are activated by active caspases, eventually leading to the death of the cell. With age, however, apoptotic activity changes.",
+    "(during development and for maintenance of homeostasis) in multi -cellular  organism is apoptosis, which is character ized by a  sequence of well -defined  events resulting in cell destruction. Dysregulation of apoptosis is responsible for  many physiological health problems and diseases; therefore, it is necessary to  understand  the responsible signaling pathways and complex interplay of  cellularprocesses. Results:   A combined mathematical model of apoptosis",
+    "is, apoptosis and necrosis. Apoptosis is considered as thedefault pathway, where cell death occurs in a controlledmanner resulting in the elimination of cells by macrophageswithout secondary damage of the surrounding cells. In con-trast, necrosis is considered an uncontrolled process whichleads to disruption of cells promoting tissue inammation[187]. Several transition states between the two pathways",
+    "tion of cells undergoing apoptosis.   Immunol Today   14:  131  136.       82.     Platt   N,     Silva   RP,   da    Gordon   S    (1998)   Recognizing death: the  phagocytosis of apoptotic cells.   Trends Cell Biol   8:  365  372.       83.     Giles   KM,     Hart   SP,     Haslett   C,     Rossi   AG,     Dransfield   I    (2000)    An appetite for apoptotic cells? Controversies and challenges.    Br J Haematol   109:  1  12.",
+    "tion of cells undergoing apoptosis.   Immunol Today   14:  131  136.       82.     Platt   N,     Silva   RP,   da    Gordon   S    (1998)   Recognizing death: the  phagocytosis of apoptotic cells.   Trends Cell Biol   8:  365  372.       83.     Giles   KM,     Hart   SP,     Haslett   C,     Rossi   AG,     Dransfield   I    (2000)    An appetite for apoptotic cells? Controversies and challenges.    Br J Haematol   109:  1  12.",
+    "the induc-tion of apoptosis.",
+    "to cancer , b ut probably not rele v ant to the i ntrinsic aging process i n yeast. Apoptosis Cell suicide, or apoptosis, i s a well-studied biological phenomenon in multicellular or g anisms t hat allo ws specic cells to be remo v e d during t he de v e lopment of com- ple x tissues, o r potentially dangerous damaged cells to be destro yed for t he benetof the w hole o r g anism. T he lack of an apparent e v olutionary benet for s uch a p ro-",
+    "15Apoptosis is caused by the activation of the caspase cascade, which isinitiated by two signaling routes (stress-induced death and death-domainreceptor-induced death) (Domen 2001). This process can be prevented by anti-apoptotic molecules, such as Bcl-2 (Domen and Weissman 2000). Directevidence for the involvement of apoptosis in HSC number regulation came fromthe findings that overexpression of the anti-apoptotic gene bcl-2 led to increasednumbers of Thy-1.1low, Sca-1+, c-kit+, Lin- cells, a population",
+    "15Apoptosis is caused by the activation of the caspase cascade, which isinitiated by two signaling routes (stress-induced death and death-domainreceptor-induced death) (Domen 2001). This process can be prevented by anti-apoptotic molecules, such as Bcl-2 (Domen and Weissman 2000). Directevidence for the involvement of apoptosis in HSC number regulation came fromthe findings that overexpression of the anti-apoptotic gene bcl-2 led to increasednumbers of Thy-1.1low, Sca-1+, c-kit+, Lin- cells, a population",
+    "Apoptosis modulating genesApopotosis or programmed cell death is associated withalterations in cell morphology, particularly the nucleus, withendonucleatytic cleavage of DNA into nucleosomal lengthfragments.Apoptosis may resultfrom withdrawalofgrowth signals.Fas, a transmembrane protein of the nerve growth factor/tumor necrosis factor receptor family signals apoptotic de-ath signals apoptotic death in some cell types. Fas but notbel-2 gene expression is negatively regulated by TSH (Ka-wakami et al., 1996),"
+  ],
+  "answer": "Apoptosis is a form of programmed cell death that is essential for normal cell turnover and various biological processes [1]. It is characterized by a sequence of well-defined events leading to cell destruction [2]. Apoptosis occurs in a controlled manner, resulting in the elimination of cells by macrophages without causing secondary damage to surrounding cells [3]. This process can be initiated through several pathways, including Bcl-2 activation of caspases, signals from death receptors on the plasma membrane, or induction by granzyme B secreted from cytotoxic T cells [1].",
+  "question": "What is apoptosis?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_4 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_4
new file mode 100644
index 0000000..a42f857
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_4
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2010 - Do different neurons age differently Direct genome-wide analysis of aging in single identified cholinergic neurons.pdf",
+    "2020 - Whole-genome sequencing of Chinese.pdf",
+    "2011 - EXPLOITING NATURAL AND INDUCED GENETIC VARIATION TO STUDY HEMATOPOIESIS.pdf",
+    "2000 - Genome-wide study of aging and oxidative stress.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2020 - A multidimensional systems biology.pdf",
+    "2008 - Combining transcriptional profiling and genetic linkage analysis to uncover gene networks operating in hematopoietic stem cells and their progeny.pdf",
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2012 - Genome-Environment Interactions That Modulate.pdf"
+  ],
+  "extraction_id": [
+    "81c68113-aa96-5af3-b4fc-5898fa20e379",
+    "0d3deffe-1f4d-5a6b-9acb-56d56141ad60",
+    "2b1a11ea-1574-5df6-b73a-a34052098751",
+    "ac5d00c0-f445-5c6a-b248-12c82c985d9a",
+    "52c67b46-63f2-54ae-a78e-e9d54a55f6e4",
+    "9d1656aa-32d2-5094-8232-4817655b1cbd",
+    "bf7b1e3c-bb4f-5a88-9167-a8c3b90cd68a",
+    "d59d7882-333d-5576-86ab-3cfa6354b946",
+    "d59d7882-333d-5576-86ab-3cfa6354b946",
+    "a01ca925-4ccf-5863-a162-7bd4c754fe89"
+  ],
+  "document_id": [
+    "153b070f-0291-5ed4-ad33-edea5e3fa8f7",
+    "9ac921c7-3991-579b-bd53-7966b91e3aae",
+    "6f250b15-61b3-57ed-8900-5aa4a173fa8c",
+    "3fc2266c-d677-54f9-b3a2-5129eedf214a",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "d040bfe3-e409-5b5c-b8f8-f3dd4fc060e3",
+    "af6e0103-849d-542f-bca7-0251082bc0b3",
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec"
+  ],
+  "id": [
+    "chatcmpl-ADZVSukRfQ2bwSsJtuTxllhMDtRvP",
+    "6f04401a-b938-5a60-8b69-d37f9086748c",
+    "02b405a4-71d7-5b85-9138-8a97c537601c",
+    "8f8848f4-d5fb-5f8c-a6b1-0f965f2abbc6",
+    "b58deffd-3cd3-5b7b-893d-b9cfc880830b",
+    "8fd5ab85-67ed-55e6-bbfa-09436c4fdbfb",
+    "61baeaa5-d65a-54b5-bfee-9bab8bbf1985",
+    "2eb33321-d0fe-5fc4-aab0-7184f2b397e0",
+    "b719fbc0-94e4-5df0-abb7-0d13fc36214c",
+    "413f8f54-b5cc-5089-9f5c-d9e3b8bcf594",
+    "3c369292-4b9c-5156-a80f-4b3301026f30"
+  ],
+  "contexts": [
+    "OTHER AGING RELATED GENES",
+    "ation of the process of aging. Studies revealed from 300 to 750 genes related to longev- ity that are critically involved in a variety of life activities, such as growth and developme nt, energy metabolism, oxi- dative stress, genomic stability maintenance, and neurocog- nition [ 4]. These candidate genes include mainly APOE, a gene involved in lipoprotein metabolism [ 5,6]. Others are those involved in cell cycle regulation, cell growth and signal transduction, the maintenance of genome stability,",
+    "down-regulated during aging were genes involved in DNA repair and chromatin remodelling. 55 While these studies revealed thousands of age-regulated genes,  the ultimate causes of these expression perturbations remain unknown. Analyzing age-dependent gene expression changes using multi-dimensional genetical genomics could bring the identification of genes causing the age-induced alterations  and thereby future therapeutic intervention strategies  one step  closer.  Adding the dimension of epigenetics",
+    "dam-age, as well as genes involved in inducing apoptosis (10, 11). Theaging process is also accompanied by changes in the expressionpatterns of a number of genes (1214). How the regulation ofgene expression in aging correlates with that in response tooxidative stress, however, is understood poorly.",
+    "www.ncbi.nlm.nih.gov/homologene)  of genes strongly asso-ciated with aging in model organisms. Also included are genesin which mutations result in segmental progeroid syndromes,such as the Werners syndrome gene, as well as genes criticalin pathways previously related to aging, such as the insulin/insulin-like signalling pathway (de Magalhes   et     al  ., 2005a). The",
+    "genes driving cellular senescence, and perform various integrative analyses. Genes inducing cellular senescence tend to be overexpressed with age in human tissues and are significantly overrepresented in anti-longevity and tumor-suppressor genes, while genes inhibiting cellular senescence overlap with pro-longevity and oncogenes. Furthermore, cellular senescence genes are strongly conserved in mammals but not in invertebrates. We also build",
+    "those down-regulated during aging were genes involved in DNA repair and chromatin remodelling (Chambers et al. 2007b ). While these studies revealed thousands of age- regulated genes, the ultimate causes of these expressionperturbations remain unknown. Analyzing age-dependent gene expression changes using multidimensional genetical genomics could bring the identification of genes causingthe age-induced alterations and thereby future therapeutic intervention strategies one step closer.",
+    "lar signatures of mammalian aging. Some of the genes",
+    "overexpressed with age seem to be a response to aging,in that they have been previously found to have protec-tive functions (de Magalha es et al., 2009b). As such,these genes may help organisms manage aging andcould be targets for manipulation. Likewise, gene ex-pression analysis of CR has been conducted to identifyassociated genes (Lee et al., 1999, 2000). A number ofmolecular signatures have emerged from such studiesthat could be useful to identify candidate processes andpathways that affect aging,",
+    "al., 2009; Stanfel et al., 2009). Many of these genesmodulate the response to environmental signals, such asfood availability, and act in signaling pathways that ifunderstood can be targeted (Fig. 1). The genetic regula-tion of aging is therefore an emerging field with multipleapplications in the human nutrition, cosmetic, and phar-maceutical industries. AGING GENES AS TARGETS FOR DRUG DISCOVERY 91"
+  ],
+  "answer": "Several genes are involved in the aging process, as identified in the provided context:  1. **APOE**: This gene is involved in lipoprotein metabolism and is one of the candidate genes related to longevity [2]. 2. **Genes involved in cell cycle regulation, cell growth, and signal transduction**: These genes play a role in maintaining genome stability and are implicated in the aging process [2]. 3. **Genes involved in DNA repair and chromatin remodeling**: These genes are down-regulated during aging, indicating their involvement in the aging process [3], [7]. 4. **Genes inducing apoptosis**: These genes are also associated with the aging process [4]. 5. **Werners syndrome gene**: Mutations in this gene result in segmental progeroid syndromes, which are related to aging [5]. 6. **Genes in the insulin/insulin-like signaling pathway**: These genes are critical in pathways previously related to aging [5]. 7. **Genes driving cellular senescence**: These genes tend to be overexpressed with age in human tissues and are significantly overrepresented in anti-longevity and tumor-suppressor genes [6].  These genes collectively contribute to various aspects of the aging process, including genomic stability, cellular senescence, and response to oxidative stress.",
+  "question": "which genes are involved in the aging process"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_5 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_5
new file mode 100644
index 0000000..55613f3
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_5
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2005 - Genomes Optimize Reproduction Aging as a Consequence of the Developmental Program.pdf",
+    "2007 - Genome Dynamics and Transcriptional Deregulation.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf",
+    "2005 - Genomes Optimize Reproduction Aging as a Consequence of the Developmental Program.pdf",
+    "2005 - Aging and Genome Maintenance.pdf",
+    "2001 - The genetics of aging.pdf",
+    "2009 - Genomic instability and DNA damage responses in progeria arising.pdf",
+    "2017 - An integrative metabolomics.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf"
+  ],
+  "extraction_id": [
+    "ac2b646d-b25b-55d2-b1f9-1180a7f0b7bf",
+    "dd3d8efb-1ce0-532e-9c77-caccb0820944",
+    "e9cec19b-9005-57e0-991e-c8b0125040df",
+    "65bc0695-ec07-50fa-83c7-f36763dc96db",
+    "dd3d8efb-1ce0-532e-9c77-caccb0820944",
+    "05fb83b5-b589-565f-89fa-c7a2fe1ec048",
+    "3f6cd069-806a-513a-a5bf-e547fd1e9737",
+    "02f84d0b-eb00-5252-a647-b2b042f1f7a9",
+    "af201c05-daed-5cba-abc8-e714483e602f",
+    "af6e2449-9591-51e8-bd8a-4f6eab30843e"
+  ],
+  "document_id": [
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "94d264da-5e72-5eb9-9fd9-a81ac2b91b77",
+    "296e3322-a519-5e4f-b955-8cc03c68e78f",
+    "62b635c3-040e-512a-b016-6ef295308a1e",
+    "94d264da-5e72-5eb9-9fd9-a81ac2b91b77",
+    "ecf53e34-4491-5db1-ad37-304671799179",
+    "aa9a9193-b6f3-5ef8-aefd-e01ec44abb46",
+    "b7d96f9f-8ad4-5f8f-94f9-60404806d478",
+    "cb0831f4-540a-5620-b69e-03d6127f84e5",
+    "62b635c3-040e-512a-b016-6ef295308a1e"
+  ],
+  "id": [
+    "chatcmpl-ADZVYj6XpaNI2z12sJiqcMEMc2C8W",
+    "b008df20-7a1c-5cee-9d71-a9758492b256",
+    "7e3dbd4f-cc54-549c-b7f2-d7a57b176185",
+    "b4eea796-bb34-591b-8873-b6d8b773d24f",
+    "1b4ba691-cd5b-52dd-9b50-ac404c6559cd",
+    "37d5273c-5359-598a-ae9a-eb407314774b",
+    "1acd92f0-07a2-53ac-86d8-2195a37e8cab",
+    "228391ea-b709-5d78-96d7-c480802f5cbc",
+    "1f14235a-ae99-58f9-8f39-a6dd29c2f676",
+    "c5c49956-fdf4-5aef-8852-2c6d14fd05b4",
+    "2d467a17-2367-5736-b0dd-20239f85cfa2"
+  ],
+  "contexts": [
+    "in the aging process.",
+    "age-related decline results from damaging by-products of metabolism and/or inefficient repairmechanisms (27, 32). According to this view, dam-agewhich can take on many formsaccumu-lates throughout the life span (38). The exponentialincrease in mortality and the functional declinethat characterize aging, however, only begin aftersexual maturity, whether this occurs at age 13, as inhumans, age 5, as in monkeys, or at less than 2months, as in mice. Therefore, one alternative viewis that aging is perhaps",
+    "of a pro-cess of mutation accumulation in somatic cells. While im-plicated as a general cause of aging, no specic mecha-nism has been proposed as to how mutation accumulationcould ever lead to the multitude of degenerative processesthat comprise aging. We have now demonstrated that alarge variety of mutations accumulate with age at greatlydifferent rates in a tissue-specic manner. More recentlywe have shown that while some organs, such as brain, donot seem to accumulate mutations with age at all,",
+    "this process between proteins and other macromolecules responsible for ageing,  while the theory of free radicals suggests that ageing is the result of inadequate pro- tection against cell and tissue damage by free radicals and oxidative stress through- out life. Finally, the wear-and-tear theory poses that the cumulative damage that  eventually leads to ageing and death is, in fact, the result of the continuous function- ing of vital processes, during which stochastic errors gradually arise.",
+    "Many mechanistic theories of aging argue that",
+    "cell senescence and cell death pathways, are a major cause of aging pheno-types, such as organ atrophy. This would appear to be a pre-programmed cause of aging, since it is a consistent response of a sizable fraction of the cell population. However, cellular responses to damage are unlikely to be the onlyexplanation for aging, since even very old organisms still appear to have am-ple tissue capacity left to function optimally.",
+    "function during aging.",
+    "INTRODUCTION    The aging process represents progressive changes in a  cell or an organism which culminate in death due to accumulated defects in function leading to system failure [1].  These defe cts result in part from  accumulated damage to DNA.  Such damage may result      www.impactaging.com AGING, January 2009, Vol. 1. No 1  Review",
+    "that induce complex molecular changes and, in turn, a deterioration of cellular structures and function. These changes are major causes of age-related diseases like cancer or cardiovascular disorders [1, 2]. The main mo- lecular adaptations occurring during aging are loss ofgenomic stability due to reduced DNA repair capacities [3], loss of proliferative potential caused by increased senescence [1, 4], and age-related alterations in the DNA-methylation patterns that affect cellular plasticity",
+    "cause in turn metabolic and cognitive alterations, resulting in increasing vulnerabil- ity to environmental challenge and a growing risk for disease and death [1]. Since  aging comprises the greatest risk factor for a variety of chronic diseases, includ- ing cancer, cardiovascular disorders, and neurodegenerative diseases [2], one of the  goals of biomedical research is to decipher the molecular mechanism underlying  aging, which in turn might facilitate the development of treatments aimed at delay-"
+  ],
+  "answer": "The aging process is caused by several factors, including:  1. **Accumulated Damage**: Age-related decline results from damaging by-products of metabolism and/or inefficient repair mechanisms, leading to damage that accumulates throughout the lifespan [2]. 2. **Mutation Accumulation**: A process of mutation accumulation in somatic cells, although no specific mechanism has been proposed for how this leads to the multitude of degenerative processes that comprise aging [3]. 3. **Free Radicals and Oxidative Stress**: The theory of free radicals suggests that aging results from inadequate protection against cell and tissue damage by free radicals and oxidative stress throughout life [4]. 4. **Wear-and-Tear**: The wear-and-tear theory posits that cumulative damage from the continuous functioning of vital processes leads to aging and death due to stochastic errors gradually arising [4]. 5. **Cell Senescence and Death Pathways**: Cell senescence and cell death pathways are major causes of aging phenotypes, such as organ atrophy, which appear to be pre-programmed responses of a sizable fraction of the cell population [6]. 6. **Accumulated Defects in Function**: Progressive changes in a cell or organism lead to accumulated defects in function, resulting in system failure and death [8]. 7. **Loss of Genomic Stability**: Loss of genomic stability due to reduced DNA repair capacities, loss of proliferative potential caused by increased senescence, and age-related alterations in DNA-methylation patterns that affect cellular plasticity [9].  These factors collectively contribute to the aging process and the associated decline in physiological functions.",
+  "question": "what causes the aging process"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_6 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_6
new file mode 100644
index 0000000..adf6e59
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_6
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2010 - Do different neurons age differently Direct genome-wide analysis of aging in single identified cholinergic neurons.pdf",
+    "2020 - A multidimensional systems biology.pdf",
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2020 - Whole-genome sequencing of Chinese.pdf",
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2011 - EXPLOITING NATURAL AND INDUCED GENETIC VARIATION TO STUDY HEMATOPOIESIS.pdf",
+    "2000 - Genome-wide study of aging and oxidative stress.pdf",
+    "2007 - Temporal and spatial transcriptional profiles.pdf",
+    "2008 - Evolution of the Aging Brain Transcriptome and Synaptic.pdf"
+  ],
+  "extraction_id": [
+    "81c68113-aa96-5af3-b4fc-5898fa20e379",
+    "9d1656aa-32d2-5094-8232-4817655b1cbd",
+    "d59d7882-333d-5576-86ab-3cfa6354b946",
+    "0d3deffe-1f4d-5a6b-9acb-56d56141ad60",
+    "a01ca925-4ccf-5863-a162-7bd4c754fe89",
+    "52c67b46-63f2-54ae-a78e-e9d54a55f6e4",
+    "2b1a11ea-1574-5df6-b73a-a34052098751",
+    "ac5d00c0-f445-5c6a-b248-12c82c985d9a",
+    "2e42619b-d0b2-5d33-aab8-6f04002ee807",
+    "bab54a5c-0b3c-5c5b-9b2b-5e7a67492a9c"
+  ],
+  "document_id": [
+    "153b070f-0291-5ed4-ad33-edea5e3fa8f7",
+    "d040bfe3-e409-5b5c-b8f8-f3dd4fc060e3",
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "9ac921c7-3991-579b-bd53-7966b91e3aae",
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "6f250b15-61b3-57ed-8900-5aa4a173fa8c",
+    "3fc2266c-d677-54f9-b3a2-5129eedf214a",
+    "38f27ec7-08bf-5397-b2b8-bde95e0dc3f8",
+    "cf413489-3986-5a5f-925d-58f94fa57428"
+  ],
+  "id": [
+    "chatcmpl-ADZVfJ7vrTDhDZNUBDMrr0RnqmSWE",
+    "6f04401a-b938-5a60-8b69-d37f9086748c",
+    "61baeaa5-d65a-54b5-bfee-9bab8bbf1985",
+    "b719fbc0-94e4-5df0-abb7-0d13fc36214c",
+    "02b405a4-71d7-5b85-9138-8a97c537601c",
+    "4d6876c5-9226-587c-8d3e-d4957ee42dba",
+    "8fd5ab85-67ed-55e6-bbfa-09436c4fdbfb",
+    "8f8848f4-d5fb-5f8c-a6b1-0f965f2abbc6",
+    "43abb9e9-5ffb-58d8-b5b9-251c50c1283d",
+    "bf2cd208-273f-5848-b243-df8b95ea7833",
+    "9430a0cd-5e05-536b-9d47-5b0b0674df5d"
+  ],
+  "contexts": [
+    "OTHER AGING RELATED GENES",
+    "genes driving cellular senescence, and perform various integrative analyses. Genes inducing cellular senescence tend to be overexpressed with age in human tissues and are significantly overrepresented in anti-longevity and tumor-suppressor genes, while genes inhibiting cellular senescence overlap with pro-longevity and oncogenes. Furthermore, cellular senescence genes are strongly conserved in mammals but not in invertebrates. We also build",
+    "lar signatures of mammalian aging. Some of the genes",
+    "ation of the process of aging. Studies revealed from 300 to 750 genes related to longev- ity that are critically involved in a variety of life activities, such as growth and developme nt, energy metabolism, oxi- dative stress, genomic stability maintenance, and neurocog- nition [ 4]. These candidate genes include mainly APOE, a gene involved in lipoprotein metabolism [ 5,6]. Others are those involved in cell cycle regulation, cell growth and signal transduction, the maintenance of genome stability,",
+    "genes (http://genomics.senescence.info/genes/), more than700 genes have been identified that regulate lifespan inmodel organisms (de Magalha es et al., 2009a). Many ofthese genes and their associated pathwayssuch as theinsulin/IGF1/GH pathwayhave been shown to affect lon-gevity across different model organisms (Kenyon, 2010).Therefore, at least some mechanisms of aging are evolu-tionarily conserved and may have potential therapeuticapplications (Baur et al., 2006). For example, evidencesuggests the use of",
+    "www.ncbi.nlm.nih.gov/homologene)  of genes strongly asso-ciated with aging in model organisms. Also included are genesin which mutations result in segmental progeroid syndromes,such as the Werners syndrome gene, as well as genes criticalin pathways previously related to aging, such as the insulin/insulin-like signalling pathway (de Magalhes   et     al  ., 2005a). The",
+    "down-regulated during aging were genes involved in DNA repair and chromatin remodelling. 55 While these studies revealed thousands of age-regulated genes,  the ultimate causes of these expression perturbations remain unknown. Analyzing age-dependent gene expression changes using multi-dimensional genetical genomics could bring the identification of genes causing the age-induced alterations  and thereby future therapeutic intervention strategies  one step  closer.  Adding the dimension of epigenetics",
+    "Aging is a biological process universal to eukaryotic organ- isms, and its underlying mechanisms are under intensive study. Genetic analyses of yeast, nematode, fly, and mouse haveuncovered a number of genes, whether mutated or misexpressed,that would increase the lifespans of these organisms (1). These genes include superoxide dismutase , a free-radical scavenger; methuselah , a potential G protein-coupled receptor, in Drosoph- ila melanogaster ; and p66 shc, an oxidative stress-response gene, in",
+    "The multifactorial and temporal features of aging can beanalyzed efficiently by genome-wide transcriptional profiling,which has been conducted in various model organisms and hu-mans (Melov and Hubbard 2004). Aging is associated with alter-ations in transcript levels of many genes, including those in-volved in evolutionarily conserved mitochondrial and protea-somal functions (McCarroll et al. 2004), some of which havebeen shown to be directly involved in regulating lifespan in C.",
+    "5. Jiang CH, Tsien JZ, Schultz PG, Hu Y (2001) The effects of aging on gene expression in the hypothalamus and cortex of mice. Proc Natl Acad Sci U S A 98: 19301934. 6. Lu T, Pan Y, Kao SY, Li C, Kohane I, et al. (2004) Gene regulation and DNA damage in the ageing human brain. Nature 429: 883891. 7. Fraser HB, Khaitovich P, Plotkin JB, Paabo S, Eisen MB (2005) Aging and gene expression in the primate brain. PLoS Biol 3: e274. 8. Zahn JM, Poosala S, Owen AB, Ingram DK, Lustig A, et al. (2007) AGEMAP: a"
+  ],
+  "answer": "Several genes are involved in the aging process, as identified in various studies:  1. **APOE**: This gene is involved in lipoprotein metabolism and is critically related to longevity [4]. 2. **Superoxide dismutase**: A free-radical scavenger gene that has been shown to increase lifespan in model organisms [8]. 3. **Methuselah**: A potential G protein-coupled receptor gene in *Drosophila melanogaster* associated with increased lifespan [8]. 4. **p66shc**: An oxidative stress-response gene involved in aging [8]. 5. **Insulin/IGF1/GH pathway genes**: These genes are part of a pathway that affects longevity across different model organisms [5]. 6. **Genes involved in DNA repair and chromatin remodeling**: These genes are down-regulated during aging [7]. 7. **Genes associated with cellular senescence**: These include genes that induce or inhibit cellular senescence, which are overexpressed with age and are linked to anti-longevity and tumor-suppressor functions [2].  Additionally, studies have identified between 300 to 750 genes related to longevity that are involved in various life activities such as growth, development, energy metabolism, oxidative stress, genomic stability maintenance, and neurocognition [4].",
+  "question": "which genes are involved in aging"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_7 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_7
new file mode 100644
index 0000000..ce4d749
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_7
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2010 - Do different neurons age differently Direct genome-wide analysis of aging in single identified cholinergic neurons.pdf",
+    "2020 - Whole-genome sequencing of Chinese.pdf",
+    "2011 - EXPLOITING NATURAL AND INDUCED GENETIC VARIATION TO STUDY HEMATOPOIESIS.pdf",
+    "2009 - The Human Ageing Genomic Resources online.pdf",
+    "2000 - Genome-wide study of aging and oxidative stress.pdf",
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2012 - Genome-Environment Interactions That Modulate.pdf",
+    "2000 - Genome-wide study of aging and oxidative stress.pdf",
+    "2008 - Genome-wide analysis of aging and learning-related genes.pdf",
+    "2008 - Combining transcriptional profiling and genetic linkage analysis to uncover gene networks operating in hematopoietic stem cells and their progeny.pdf"
+  ],
+  "extraction_id": [
+    "81c68113-aa96-5af3-b4fc-5898fa20e379",
+    "0d3deffe-1f4d-5a6b-9acb-56d56141ad60",
+    "2b1a11ea-1574-5df6-b73a-a34052098751",
+    "52c67b46-63f2-54ae-a78e-e9d54a55f6e4",
+    "ac5d00c0-f445-5c6a-b248-12c82c985d9a",
+    "d59d7882-333d-5576-86ab-3cfa6354b946",
+    "a01ca925-4ccf-5863-a162-7bd4c754fe89",
+    "ac5d00c0-f445-5c6a-b248-12c82c985d9a",
+    "593d5c1f-316a-5eab-814c-4ba150111d3e",
+    "bf7b1e3c-bb4f-5a88-9167-a8c3b90cd68a"
+  ],
+  "document_id": [
+    "153b070f-0291-5ed4-ad33-edea5e3fa8f7",
+    "9ac921c7-3991-579b-bd53-7966b91e3aae",
+    "6f250b15-61b3-57ed-8900-5aa4a173fa8c",
+    "e43cd3b6-ad8e-5422-ba7c-ceb6e66cc529",
+    "3fc2266c-d677-54f9-b3a2-5129eedf214a",
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "b1a1997c-e9df-5dc0-9d12-a3977d0c64ec",
+    "3fc2266c-d677-54f9-b3a2-5129eedf214a",
+    "cae3fe6b-9fe6-561d-8a2f-fe9fbd580f3d",
+    "af6e0103-849d-542f-bca7-0251082bc0b3"
+  ],
+  "id": [
+    "chatcmpl-ADZVlFDc0TCB7W4npOF5tI5Kq0cCy",
+    "6f04401a-b938-5a60-8b69-d37f9086748c",
+    "02b405a4-71d7-5b85-9138-8a97c537601c",
+    "8f8848f4-d5fb-5f8c-a6b1-0f965f2abbc6",
+    "8fd5ab85-67ed-55e6-bbfa-09436c4fdbfb",
+    "b58deffd-3cd3-5b7b-893d-b9cfc880830b",
+    "413f8f54-b5cc-5089-9f5c-d9e3b8bcf594",
+    "3c369292-4b9c-5156-a80f-4b3301026f30",
+    "43abb9e9-5ffb-58d8-b5b9-251c50c1283d",
+    "b284606e-a2db-5151-9d30-b591493b984d",
+    "2eb33321-d0fe-5fc4-aab0-7184f2b397e0"
+  ],
+  "contexts": [
+    "OTHER AGING RELATED GENES",
+    "ation of the process of aging. Studies revealed from 300 to 750 genes related to longev- ity that are critically involved in a variety of life activities, such as growth and developme nt, energy metabolism, oxi- dative stress, genomic stability maintenance, and neurocog- nition [ 4]. These candidate genes include mainly APOE, a gene involved in lipoprotein metabolism [ 5,6]. Others are those involved in cell cycle regulation, cell growth and signal transduction, the maintenance of genome stability,",
+    "down-regulated during aging were genes involved in DNA repair and chromatin remodelling. 55 While these studies revealed thousands of age-regulated genes,  the ultimate causes of these expression perturbations remain unknown. Analyzing age-dependent gene expression changes using multi-dimensional genetical genomics could bring the identification of genes causing the age-induced alterations  and thereby future therapeutic intervention strategies  one step  closer.  Adding the dimension of epigenetics",
+    "www.ncbi.nlm.nih.gov/homologene)  of genes strongly asso-ciated with aging in model organisms. Also included are genesin which mutations result in segmental progeroid syndromes,such as the Werners syndrome gene, as well as genes criticalin pathways previously related to aging, such as the insulin/insulin-like signalling pathway (de Magalhes   et     al  ., 2005a). The",
+    "dam-age, as well as genes involved in inducing apoptosis (10, 11). Theaging process is also accompanied by changes in the expressionpatterns of a number of genes (1214). How the regulation ofgene expression in aging correlates with that in response tooxidative stress, however, is understood poorly.",
+    "overexpressed with age seem to be a response to aging,in that they have been previously found to have protec-tive functions (de Magalha es et al., 2009b). As such,these genes may help organisms manage aging andcould be targets for manipulation. Likewise, gene ex-pression analysis of CR has been conducted to identifyassociated genes (Lee et al., 1999, 2000). A number ofmolecular signatures have emerged from such studiesthat could be useful to identify candidate processes andpathways that affect aging,",
+    "al., 2009; Stanfel et al., 2009). Many of these genesmodulate the response to environmental signals, such asfood availability, and act in signaling pathways that ifunderstood can be targeted (Fig. 1). The genetic regula-tion of aging is therefore an emerging field with multipleapplications in the human nutrition, cosmetic, and phar-maceutical industries. AGING GENES AS TARGETS FOR DRUG DISCOVERY 91",
+    "Aging is a biological process universal to eukaryotic organ- isms, and its underlying mechanisms are under intensive study. Genetic analyses of yeast, nematode, fly, and mouse haveuncovered a number of genes, whether mutated or misexpressed,that would increase the lifespans of these organisms (1). These genes include superoxide dismutase , a free-radical scavenger; methuselah , a potential G protein-coupled receptor, in Drosoph- ila melanogaster ; and p66 shc, an oxidative stress-response gene, in",
+    "nicance of genes that were found to be aected by aging,the most prominent appeared to be involved in processesthat involve cell division, cell death and apoptosis, migra-tion of cells, and dierentiation, all of which are consistentwith changes in the dierent stages of neurogenesis. Thesechanges at the molecular level agree with studies at the cel- lular level that report changes in rate of migration, dieren- tiation and neurogenesis with aging ( Seki & Arai, 1995;",
+    "those down-regulated during aging were genes involved in DNA repair and chromatin remodelling (Chambers et al. 2007b ). While these studies revealed thousands of age- regulated genes, the ultimate causes of these expressionperturbations remain unknown. Analyzing age-dependent gene expression changes using multidimensional genetical genomics could bring the identification of genes causingthe age-induced alterations and thereby future therapeutic intervention strategies one step closer."
+  ],
+  "answer": "Several genes are involved in the aging process, each playing a role in various biological activities and pathways. Key genes and their associated functions include:  1. **APOE**: Involved in lipoprotein metabolism [2]. 2. **Superoxide dismutase**: A free-radical scavenger [8]. 3. **Methuselah**: A potential G protein-coupled receptor in *Drosophila melanogaster* [8]. 4. **p66shc**: An oxidative stress-response gene [8]. 5. **Genes involved in DNA repair and chromatin remodeling**: These genes are down-regulated during aging [3], [10]. 6. **Genes involved in cell cycle regulation, cell growth, and signal transduction**: These genes help maintain genome stability [2]. 7. **Genes involved in inducing apoptosis**: These genes are associated with damage response [5]. 8. **Genes in the insulin/insulin-like signaling pathway**: Critical in pathways previously related to aging [4].  These genes collectively contribute to processes such as growth and development, energy metabolism, oxidative stress response, genomic stability maintenance, and neurocognition [2].",
+  "question": "what genes are involved in  the aging process"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_8 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_8
new file mode 100644
index 0000000..b1bb064
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_8
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Genome-wide association analysis of age-at-onset.pdf",
+    "2007 - Genetic correlates of brain aging on MRI and cognitive test measures a genome-wide association and linkage analysis in the Framingham study.pdf",
+    "2009 - MicroRNA Implications for Alzheimer Disease and other Human CNS.pdf",
+    "2018 - Genomics New Light on Alzheimer?s.pdf",
+    "2012 - Mitochondrial Genomic Analysis of Late Onset.pdf",
+    "2021 - A genome-wide association study with 1,126,563.pdf",
+    "2017 - Genomic Variants, Genes, and Pathways.pdf",
+    "2003 - The application of functional genomics.pdf",
+    "2018 - Cognitive decline and dementia in diabetes mellitus.pdf",
+    "2012 - Genome-wide association study of Alzheimer?s disease.pdf"
+  ],
+  "extraction_id": [
+    "2a2e5ce1-cc56-579c-bf79-f9057f4c9671",
+    "b545e588-2876-5928-9c01-710c1371b44e",
+    "4b383c2a-f0de-5420-af8d-07060b8874f3",
+    "64f3adb4-e745-5738-af28-43c2a870c086",
+    "b826d64d-9d95-5522-8179-82f79d957c03",
+    "c4d63e5d-36ac-572e-8269-f9efd9c0437e",
+    "7cff03ac-de86-5e70-bbcb-dadc2fa447c3",
+    "171377f4-24a7-5cde-adff-c9c7096edc75",
+    "d2b4b131-b7c9-595b-813a-b0940c4e87c0",
+    "1f9039f1-91e9-538b-b709-a1880cf47007"
+  ],
+  "document_id": [
+    "9874359e-5f5d-5e6a-9844-cd9a1d2cae24",
+    "56b25b5a-fc9e-5d61-8502-1c110466ba16",
+    "70d08119-d16d-5e9b-89ed-ec8547be125a",
+    "940593d2-04c3-59b9-a5bf-976febbc6f71",
+    "5404a17c-34a9-5881-8b1a-2acacdc996a8",
+    "77fafe7f-6f8f-5779-9d49-77213f149512",
+    "a5bf6a11-3ed5-5222-bc4d-d5149188cdbd",
+    "ec5f7b7d-6bd2-5580-bf3e-3c8b64a74169",
+    "fc7027a3-f885-55b8-b56d-bb8117e2a2f1",
+    "7e7a8526-ff6e-5c83-ae72-e45509e3b788"
+  ],
+  "id": [
+    "chatcmpl-ADZVsDdzFGdoBGkCMl1sqg3Cv85XM",
+    "6ac3f90f-ec8a-55c7-a3f7-d597d9d7cb2e",
+    "4014c984-d6d9-5eb2-a25e-9e9fe15d1b41",
+    "cd8f4d4a-2b1a-542f-a3f3-364a83fb10fb",
+    "f24834c0-1862-5d9f-bdb6-2af38505aa5c",
+    "064a3510-9a3d-5b93-b848-69478e02e013",
+    "794ac337-0c08-52ca-9603-fe840fa343af",
+    "fc44e06f-a727-5544-ad7c-6ba3632552b7",
+    "5e6b9b4b-1e03-585a-af52-18a054e1e603",
+    "784e7626-1d9c-521b-84f3-965965435366",
+    "dac0ab78-d01f-5f95-a129-559cbe6791ec"
+  ],
+  "contexts": [
+    "Introduction Alzheimers disease (AD), a devastating neurodegen- erative disease, is the most common form of dementiaamong the elderly. Genetically, AD is a complex and multifactorial disease with the possible involvement of multiple genes. The rare early-onset form of the diseaseusually follows an autosomal-dominant inheritance pattern and to date three genes have been identified: amyloid precursor protein ( APP) and presenilin 1 and 2(PSEN1 andPSEN2 ). The common late-onset form of",
+    "Background Age-related neurological diseases such as stroke and dementia represent a substantial population burden, and one in three persons will develop either stroke or demen- tia in their lifetime [1]. Twin studies suggest that 3778% of the variance in the age of onset of Alzheimer's disease (AD), the most common cause of dementia in the elderly, can be attributed to additive genetic effects [2,3]. Con- versely, cognitively healthy aging also has a substantial",
+    "cognitive status in Alzheimer's disease. Neurobiol. Aging  1996 , 17:  921-933.  [3]  Ertekin-Taner, N. Genetics of Alzheimer's disease: a centennial  review. Neurol. Clin.  2007 , 25: 611-667.  [4]  Bernardi, L., Tomaino, C., Anfossi, M., Gallo, M., Geracitano, S.,  Puccio, G., Colao, R., Frangipane, F., Mirabelli, M., Smirne, N.,  Giovanni Maletta, R., Bruni, A.C. Late onset familial Alzheimer's  disease: novel presen ilin 2 mutation and PS1 E 318G polymor- phism. J. Neurol.  2008 , 255: 604-606.",
+    "Keywords: alzheimers disease; genomics; GWAS; genetic risk factors; epigenetic modication; aging 1. Introduction Alzheimers disease (AD) is the most common cause of dementia, accounting for approximately 6080% of dementia cases, followed by vascular dementia (approximately 10%), Lewy Body or Parkinsons disease-related dementia, and alcohol-mediated dementia [ 1]. Mild cognitive impairment, one of the representative early symptoms of AD, makes this disease distinguishable from other types",
+    "14. Heyman A, Wilkinson WE, Hurwitz BJ, Schmechel D, Sigmon AH, et al. (1983) Alzheimers disease: genetic aspects and associated clinical disorders. AnnNeurol 14: 507515. 15. Farrer LA, Myers RH, Connor L, Cupples LA, Growdon JH (1991) Segregation analysis reveals evidence of a major gene for Alzheimer disease. Am J HumGenet 48: 10261033. 16. Duara R, Lopez-Alberola RF, Barker WW, Loewenstein DA, Zatinsky M, et al. (1993) A comparison of familial and sporadic Alzheimers disease. Neurology 43: 13771384.",
+    "(2016).  3. DeTure, M. A. & Dickson, D. W . The neuropathological diagnosis of Alzheimers disease. Mol. Neurodegener. 14, 32 (2019).  4. Gatz, M. et al. Heritability for Alzheimers disease: the study of dementia in Swedish twins. J. Gerontol. A Biol. Sci. Med. Sci. 52, M117M125 (1997).  5. Gatz, M. et al. Role of genes and environments for explaining Alzheimer disease. Arch. Gen. Psychiatry 63, 168174 (2006).",
+    "Lett 379(3):199204. Avramopoulos D. 2009. Genetics of Alzheimers disease: Recent advances. Genome Med 1(3):34. Bachman DL, Wolf PA, Linn R, Knoefel JE, Cobb J, Belanger A, DAgostino RB, White LR. 1992. Prevalence of dementia and probable seniledementia of the Alzheimer type in the Framingham study. Neurology42(1):115119. Barral S, Cheng R, Reitz C, Vardarajan B, Lee J, Kunkle B, Beecham G,",
+    "[11] and the exclusion of cerebrovascular factors as inherentetiopathogenic determinants of neuronal deathin AD, taking into account that in patients olderthan 70 years of age the vast majority of caseswith dementia show a clear cerebrovascular com-promise  [12]. In addition, most studies attempt- ing to correlate clinical features with singlegenotypes are partially biased due to heterogene-ity and inaccuracy in phenotype recruitment.Furthermore, 6080% of the therapeutic fail-ures in AD",
+    "associated with Alzheimers disease neuropathology. J. Alzheimers Dis. 60, 10351043 (2017). 63. Gottesman, R. F. etal. Association between midlife vascular risk factors and estimated brain amyloid  deposition. JAMA 317, 14431450 (2017). 64. Moran, C. etal. T ype 2 diabetes mellitus and  biomarkers of neurodegeneration. Neurology 85,  11231130 (2015). 65. Vemuri, P . etal. Age, vascular health, and Alzheimer disease biomarkers in an elderly sample. Ann. Neurol.   82, 706718 (2017).",
+    "Introduction Alzheimers disease (AD), the most common form of dementia, is highly heritable (heritability of up to 76%) but genetically complex.1Neuropatho- logically, the disease is characterized by extracellular senile plaques containing b-amyloid (A b) and intra- cellular neurofibrillary tangles containing hyperpho-sphorylated tau protein. 1Before 2009, four genes had been definitively implicated in its aetiology. Muta- tions of the amyloid precursor protein (APP) gene"
+  ],
+  "answer": "The genotypes related to Alzheimer's disease (AD) and dementia that have commonalities with those for aging include several key genetic factors:  1. **Amyloid Precursor Protein (APP)**: Mutations in the APP gene are implicated in the early-onset form of Alzheimer's disease. This gene is also associated with the aging process, as the accumulation of amyloid plaques is a hallmark of both AD and age-related cognitive decline [1], [10].  2. **Presenilin 1 and 2 (PSEN1 and PSEN2)**: These genes are also involved in the early-onset form of AD and are linked to the production of amyloid-beta, a protein that accumulates in the brains of individuals with AD and is also associated with aging [1].  3. **APOE \u03b54 Allele**: The APOE \u03b54 allele is a well-known genetic risk factor for late-onset Alzheimer's disease. It is also associated with aging, as it influences lipid metabolism and is linked to age-related cognitive decline [2], [4].  4. **Genetic Variance in Age of Onset**: Twin studies suggest that 37-78% of the variance in the age of onset of Alzheimer's disease can be attributed to additive genetic effects, indicating a significant overlap between the genetic factors influencing AD and those affecting the aging process [2].  These genotypes highlight the genetic commonalities between Alzheimer's disease, dementia, and the aging process, emphasizing the multifactorial nature of these conditions.",
+  "question": "Describe the genotypes related to Alzheimers and dementia which have commonalities with those for aging."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_9 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_9
new file mode 100644
index 0000000..6ec3e5f
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_aging_9
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2007 - Genetic correlates of brain aging on MRI and cognitive test measures a genome-wide association and linkage analysis in the Framingham study.pdf",
+    "2009 - MicroRNA Implications for Alzheimer Disease and other Human CNS.pdf",
+    "2012 - Genome-wide association analysis of age-at-onset.pdf",
+    "2003 - The application of functional genomics.pdf",
+    "2018 - Cognitive decline and dementia in diabetes mellitus.pdf",
+    "2012 - Genome-wide association study of Alzheimer?s disease.pdf",
+    "2018 - Genomics New Light on Alzheimer?s.pdf",
+    "2012 - Mitochondrial Genomic Analysis of Late Onset.pdf",
+    "2003 - Results of a high-resolution genome screen.pdf",
+    "2017 - Genomic Variants, Genes, and Pathways.pdf"
+  ],
+  "extraction_id": [
+    "b545e588-2876-5928-9c01-710c1371b44e",
+    "4b383c2a-f0de-5420-af8d-07060b8874f3",
+    "2a2e5ce1-cc56-579c-bf79-f9057f4c9671",
+    "171377f4-24a7-5cde-adff-c9c7096edc75",
+    "d2b4b131-b7c9-595b-813a-b0940c4e87c0",
+    "1f9039f1-91e9-538b-b709-a1880cf47007",
+    "64f3adb4-e745-5738-af28-43c2a870c086",
+    "b826d64d-9d95-5522-8179-82f79d957c03",
+    "053cb638-e8ee-593b-9e3c-fe745534adfc",
+    "7cff03ac-de86-5e70-bbcb-dadc2fa447c3"
+  ],
+  "document_id": [
+    "56b25b5a-fc9e-5d61-8502-1c110466ba16",
+    "70d08119-d16d-5e9b-89ed-ec8547be125a",
+    "9874359e-5f5d-5e6a-9844-cd9a1d2cae24",
+    "ec5f7b7d-6bd2-5580-bf3e-3c8b64a74169",
+    "fc7027a3-f885-55b8-b56d-bb8117e2a2f1",
+    "7e7a8526-ff6e-5c83-ae72-e45509e3b788",
+    "940593d2-04c3-59b9-a5bf-976febbc6f71",
+    "5404a17c-34a9-5881-8b1a-2acacdc996a8",
+    "37eda0c5-73f9-5615-be6f-7016071ec1f4",
+    "a5bf6a11-3ed5-5222-bc4d-d5149188cdbd"
+  ],
+  "id": [
+    "chatcmpl-ADZVxBA3IhLP4BoAeswrss7YVgqiu",
+    "4014c984-d6d9-5eb2-a25e-9e9fe15d1b41",
+    "cd8f4d4a-2b1a-542f-a3f3-364a83fb10fb",
+    "6ac3f90f-ec8a-55c7-a3f7-d597d9d7cb2e",
+    "5e6b9b4b-1e03-585a-af52-18a054e1e603",
+    "784e7626-1d9c-521b-84f3-965965435366",
+    "dac0ab78-d01f-5f95-a129-559cbe6791ec",
+    "f24834c0-1862-5d9f-bdb6-2af38505aa5c",
+    "064a3510-9a3d-5b93-b848-69478e02e013",
+    "f95a098d-6950-551a-8854-2c4b956cb10b",
+    "fc44e06f-a727-5544-ad7c-6ba3632552b7"
+  ],
+  "contexts": [
+    "Background Age-related neurological diseases such as stroke and dementia represent a substantial population burden, and one in three persons will develop either stroke or demen- tia in their lifetime [1]. Twin studies suggest that 3778% of the variance in the age of onset of Alzheimer's disease (AD), the most common cause of dementia in the elderly, can be attributed to additive genetic effects [2,3]. Con- versely, cognitively healthy aging also has a substantial",
+    "cognitive status in Alzheimer's disease. Neurobiol. Aging  1996 , 17:  921-933.  [3]  Ertekin-Taner, N. Genetics of Alzheimer's disease: a centennial  review. Neurol. Clin.  2007 , 25: 611-667.  [4]  Bernardi, L., Tomaino, C., Anfossi, M., Gallo, M., Geracitano, S.,  Puccio, G., Colao, R., Frangipane, F., Mirabelli, M., Smirne, N.,  Giovanni Maletta, R., Bruni, A.C. Late onset familial Alzheimer's  disease: novel presen ilin 2 mutation and PS1 E 318G polymor- phism. J. Neurol.  2008 , 255: 604-606.",
+    "Introduction Alzheimers disease (AD), a devastating neurodegen- erative disease, is the most common form of dementiaamong the elderly. Genetically, AD is a complex and multifactorial disease with the possible involvement of multiple genes. The rare early-onset form of the diseaseusually follows an autosomal-dominant inheritance pattern and to date three genes have been identified: amyloid precursor protein ( APP) and presenilin 1 and 2(PSEN1 andPSEN2 ). The common late-onset form of",
+    "[11] and the exclusion of cerebrovascular factors as inherentetiopathogenic determinants of neuronal deathin AD, taking into account that in patients olderthan 70 years of age the vast majority of caseswith dementia show a clear cerebrovascular com-promise  [12]. In addition, most studies attempt- ing to correlate clinical features with singlegenotypes are partially biased due to heterogene-ity and inaccuracy in phenotype recruitment.Furthermore, 6080% of the therapeutic fail-ures in AD",
+    "associated with Alzheimers disease neuropathology. J. Alzheimers Dis. 60, 10351043 (2017). 63. Gottesman, R. F. etal. Association between midlife vascular risk factors and estimated brain amyloid  deposition. JAMA 317, 14431450 (2017). 64. Moran, C. etal. T ype 2 diabetes mellitus and  biomarkers of neurodegeneration. Neurology 85,  11231130 (2015). 65. Vemuri, P . etal. Age, vascular health, and Alzheimer disease biomarkers in an elderly sample. Ann. Neurol.   82, 706718 (2017).",
+    "Introduction Alzheimers disease (AD), the most common form of dementia, is highly heritable (heritability of up to 76%) but genetically complex.1Neuropatho- logically, the disease is characterized by extracellular senile plaques containing b-amyloid (A b) and intra- cellular neurofibrillary tangles containing hyperpho-sphorylated tau protein. 1Before 2009, four genes had been definitively implicated in its aetiology. Muta- tions of the amyloid precursor protein (APP) gene",
+    "Keywords: alzheimers disease; genomics; GWAS; genetic risk factors; epigenetic modication; aging 1. Introduction Alzheimers disease (AD) is the most common cause of dementia, accounting for approximately 6080% of dementia cases, followed by vascular dementia (approximately 10%), Lewy Body or Parkinsons disease-related dementia, and alcohol-mediated dementia [ 1]. Mild cognitive impairment, one of the representative early symptoms of AD, makes this disease distinguishable from other types",
+    "14. Heyman A, Wilkinson WE, Hurwitz BJ, Schmechel D, Sigmon AH, et al. (1983) Alzheimers disease: genetic aspects and associated clinical disorders. AnnNeurol 14: 507515. 15. Farrer LA, Myers RH, Connor L, Cupples LA, Growdon JH (1991) Segregation analysis reveals evidence of a major gene for Alzheimer disease. Am J HumGenet 48: 10261033. 16. Duara R, Lopez-Alberola RF, Barker WW, Loewenstein DA, Zatinsky M, et al. (1993) A comparison of familial and sporadic Alzheimers disease. Neurology 43: 13771384.",
+    "disease. Nat. Genet. ,19, 321 322. 7. Bergem, A.L., Engedal, K. and Kringlen, E. (1997) The role of heredity in late-onset Alzheimer disease and vascular dementia. A twin study. Arch. Gen. Psychiat. ,54, 264 270. 8. Payami, H., Grimslid, H., Oken, B., Camicioli, R., Sexton, G., Dame, A., Howieson, D. and Kaye, J. (1997) A prospective study of cognitive health inthe elderly (Oregon Brain Aging Study): effects of family history andapolipoprotein E genotype. Am. J. Hum. Genet. ,60, 948 956.",
+    "Lett 379(3):199204. Avramopoulos D. 2009. Genetics of Alzheimers disease: Recent advances. Genome Med 1(3):34. Bachman DL, Wolf PA, Linn R, Knoefel JE, Cobb J, Belanger A, DAgostino RB, White LR. 1992. Prevalence of dementia and probable seniledementia of the Alzheimer type in the Framingham study. Neurology42(1):115119. Barral S, Cheng R, Reitz C, Vardarajan B, Lee J, Kunkle B, Beecham G,"
+  ],
+  "answer": "The genotypes related to Alzheimer's disease (AD) and dementia that have commonalities with those for aging include several key genetic factors:  1. **Apolipoprotein E (APOE)**: The APOE gene, particularly the APOE \u03b54 allele, is a well-known genetic risk factor for late-onset Alzheimer's disease. It is also associated with cognitive decline in aging populations [9]. Studies have shown that individuals carrying the APOE \u03b54 allele have an increased risk of developing Alzheimer's disease and exhibit more rapid cognitive decline as they age [8].  2. **Amyloid Precursor Protein (APP)**: Mutations in the APP gene are implicated in the early-onset form of Alzheimer's disease. While these mutations are rare, they contribute to the understanding of the disease's pathology, which overlaps with mechanisms involved in aging, such as amyloid plaque formation [3], [6].  3. **Presenilin 1 and 2 (PSEN1 and PSEN2)**: Mutations in these genes are also associated with early-onset Alzheimer's disease. These mutations affect the production of amyloid-beta, a key component of the plaques found in the brains of Alzheimer's patients. The accumulation of amyloid-beta is a process that is also observed in the aging brain, albeit to a lesser extent in cognitively healthy individuals [3], [6].  4. **Genetic Variance in Age of Onset**: Twin studies suggest that a significant portion of the variance in the age of onset of Alzheimer's disease can be attributed to genetic factors, indicating a genetic overlap between aging and the development of Alzheimer's disease [1].  5. **Vascular Health and Aging**: Vascular health is another important factor that links aging with Alzheimer's disease. Midlife vascular risk factors, such as hypertension and diabetes, are associated with increased amyloid deposition in the brain, which is a hallmark of Alzheimer's disease. This suggests that the genetic factors influencing vascular health and aging may also play a role in the development of Alzheimer's disease [5].  In summary, the genotypes related to Alzheimer's disease and dementia that share commonalities with those for aging include APOE, APP, PSEN1, and PSEN2, as well as genetic factors influencing vascular health and the age of onset of the disease [1], [3], [5], [6], [8], [9].",
+  "question": "Describe the genotypes related to Alzheimer's and dementia which have commonalities with those for aging."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_1 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_1
new file mode 100644
index 0000000..4b5b06e
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_1
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2015 - Pharmacogenetics and individual responses to treatment of hyperglycemia.pdf",
+    "2015 - Pharmacogenetics and individual responses to treatment of hyperglycemia.pdf",
+    "2021 -  Monogenic diabetes a gateway to precision medicine.pdf",
+    "2020 - Precision Medicine in Diabetes.pdf",
+    "2014 - Heritability of variation in glycaemic response to metformin.pdf",
+    "2015 - Pharmacogenetics and individual responses to treatment of hyperglycemia.pdf",
+    "2012 - Type 2 Diabetes Genetics Beyond GWAS.pdf",
+    "2018 - Human Genetics of Obesity and Type 2 Diabetes Mellitus.pdf",
+    "2010 - Pharmacogenetics of Anti-Diabetes Drugs.pdf",
+    "2011 - Inherited destiny Genetics and gestational diabetes mellitus.pdf"
+  ],
+  "extraction_id": [
+    "026d2a7d-a7b7-5342-981a-2664a998c79b",
+    "026d2a7d-a7b7-5342-981a-2664a998c79b",
+    "baea9ac6-7ff9-5724-87ed-81b17e2469cd",
+    "c27447b1-5f7e-5b8b-9172-baba74ffc29b",
+    "90ea6bd5-5140-5c73-ace7-fd5030e83c6d",
+    "026d2a7d-a7b7-5342-981a-2664a998c79b",
+    "a3a875fa-e55b-52d0-b9bf-72b96330c393",
+    "e18fd615-3cde-5dc2-ab7d-a9e17d4c8ed6",
+    "a1359f6d-8f61-51ca-8b02-45420e345946",
+    "48c3e4a4-db23-5fca-9c46-775e80894655"
+  ],
+  "document_id": [
+    "46081466-a50f-59d8-893d-8b8883b38507",
+    "46081466-a50f-59d8-893d-8b8883b38507",
+    "0b6ff786-6a7b-5d24-ba5e-7a61fee7757f",
+    "0ad5b2de-d782-5d43-b294-bff5c7befd2d",
+    "458da117-3235-5852-aff2-529c0bf16074",
+    "46081466-a50f-59d8-893d-8b8883b38507",
+    "d59a38d7-889b-51b5-b896-c305c82a2169",
+    "2083de31-17c6-5d1e-9aa6-2efc6c1d9ac2",
+    "ffeebaf9-ff76-5751-9b8b-7a2a4a4f1dc3",
+    "6d341cd2-ae56-5807-9aff-39298efc4d06"
+  ],
+  "id": [
+    "chatcmpl-ADZQAXp2EmWZCiBbiRu4ySm4isUy8",
+    "6aa611a9-aa5b-5dc5-a760-eaf1f95b8109",
+    "4352b950-a365-523c-b704-9eb4eddaf448",
+    "db50a759-ac52-5e02-a5c1-5c898f16bd27",
+    "c372d094-ceb2-56d1-82f3-c63f65e5c5c1",
+    "f187dbbd-3380-566a-ab25-18fc923e2263",
+    "9ced327e-3feb-5b7b-a938-30ad544113e2",
+    "b4516514-f107-5b15-b73d-0d3d89dce5a8",
+    "6707ac07-6096-5eaa-b6c4-315faa4c2813",
+    "c2b8b8a1-d19e-5f7e-aa22-a421367e4fdd",
+    "35d3fc6c-28a8-53fe-9574-e92d87f01c19"
+  ],
+  "contexts": [
+    "interindividual variation in responses to antidiabetic treatment and may provide the foundation for future genotype-based treatment standards. Pharmacogenetics and Genomics 25:475 484 Copyright  2015 Wolters Kluwer Health, Inc. All rights reserved. Pharmacogenetics and Genomics 2015, 25:475 484 Keywords: antidiabetic treatment, diabetes type 2, disease progression, genotype, pharmacogenetics aSection of Metabolic Genetics, Novo Nordisk Research Foundation Center for",
+    "treatment guidelines. Yet, the interindividual response to therapy and slope of disease progression varies markedly among patients with type 2 diabetes. Gene gene, gene environment, and gene treatment interactions may explain some of the variation in disease progression. Several genetic variants have been suggested to beassociated with response to antidiabetic drugs. Some are present in drug receptors or drug metabolizers ( OCT genes, KCNJ11 ,ABCC8 , and CYP2C9 ). Numerous type 2 diabetes",
+    "mic control in the majority of insulin-treated  patients. Diabet Med . 2009;26(4):437441.  20. Pearson ER, et al. Sensitivity to sulphonylureas  in patients with hepatocyte nuclear factor-1alpha  gene mutations: evidence for pharmacogenetics  in diabetes. Diabet Med . 2000;17(7):543545.  21. Pearson ER, et al. Genetic cause of hypergly- caemia and response to treatment in diabetes.  Lancet . 2003;362(9392):12751281.  22. Fantasia KL, Steenkamp DW. Optimal glycemic",
+    "When considering etiological varia- tion, recent work partitioning diabe-tes-associated genetic variants by theirpresumed etiological process (parti-tioned polygenic scores) (6,42,101)may de ne genetically driven dominant processes. These processes, such asb-cell dysfunction, lipodystrophy, or obe- sity, could respond differently to drugsthat act on these pathways, such assulfonylureas, glucagon-like peptide 1 re- ceptor agonist (GLP-1RA), DPP4i, and thiazolidinediones.",
+    "source of such variation might help to identify patients most likely not to respond to metformin and could help to develop more e  ective agents by providing insight into  the biological mechanism of metformin. As with other complex traits, glycaemic response to  metformin is probably determined by the interplay between genetic and environmental factors. Clinical variables such as BMI, drug adherence, and dosing only account for part of the variation. 3 Pharmacogenetic",
+    "Pharmacogenetics and individual responses to treatment of hyperglycemia in type 2 diabetes Line Engelbrechtsena, Ehm Anderssona, Soeren Roepstorffb, Torben Hansenaand Henrik Vestergaarda The aim of this study was to summarize current knowledge and provide perspectives on the relationships between human genetic variants, type 2 diabetes, antidiabetic treatment, and disease progression. Type 2 diabetes is a complex disease with clear-cut diagnostic criteria and",
+    "Genomics. 2010; 20:3844. [PubMed: 19898263] 168. Jablonski KA, McAteer JB, de Bakker PI, Franks PW, Pollin TI, et al. Common variants in 40 genes assessed for diabetes incidence and response to metformin and lifestyle intervention in the diabetes prevention program. Diabetes. 2010; 59:26722681. [PubMed: 20682687] 169. Wolford JK, Yeatts KA, Dhanjal SK, Black MH, Xiang AH, et al. Sequence variation in PPARG may underlie differential response to troglitazone. Diabetes. 2005; 54:33193325. [PubMed: 16249460]",
+    "10.1007/s00125-017-4227-1.  42. Hattersley AT, et al. Precision diabetes: learning from monogenic diabetes. Diabetologia. 2017;60:769777. doi: 10.1007/s00125-017-4226-2.  43. Florez JC. The pharmacogenetics of metformin. Diabetologia.  2017;60:16481655. doi: 10.1007/s00125-017-4335-y.  44. Maruthur NM, et al. The pharmacogenetics of type 2 diabetes: a system-atic review. Diabetes Care. 2014;37:876886. doi: 10.2337/dc13-1276.  45. Zhou K, et al. Variation in the glucose transporter gene SLC2A2 is associ-",
+    "typically based on efficacy, yet favorable respon ses to such therapeutics are oftentimes  variable and difficult to pred ict. Characterization of drug  response is expected to  substantially enhance our ability to provide patients with the most effective treatment  strategy given their indivi dual backgrounds, yet pharmacogenetic study of diabetes  medications is still in its infancy. To date, major pharmacogenetic studies have focused on",
+    "treatment or adverse effects  and dosing of medications  are not likely to be adversely affected by environmental  exposures and tend to have large effect sizes [95]. There fore, some of the variability in response or dosing could  be due to genetic variation. Pharmacogenetics in the area  of diabetes is still in its infancy, although there have been  studies examining KCNJ11  and sulfonylurea therapy for  both rare [96,97] and common [98,99] variants and res"
+  ],
+  "answer": "The genetic bases for the varying efficacy of diabetes treatments among individuals are multifaceted and involve several genetic variants that influence drug response. Key points include:  1. **Genetic Variants in Drug Receptors and Metabolizers**: Variations in genes such as OCT, KCNJ11, ABCC8, and CYP2C9 have been associated with differing responses to antidiabetic drugs. These genes are involved in drug receptors or drug metabolism, which can affect how well a patient responds to a particular treatment [2].  2. **Gene-Gene and Gene-Environment Interactions**: The interplay between different genes and between genes and environmental factors can also contribute to the variability in treatment efficacy. This includes how genetic predispositions interact with lifestyle factors and other environmental influences [2].  3. **Specific Genetic Mutations**: For instance, mutations in the hepatocyte nuclear factor-1alpha gene have been linked to sensitivity to sulphonylureas, highlighting a specific pharmacogenetic interaction in diabetes treatment [3].  4. **Genetically Driven Dominant Processes**: Certain genetic variants may drive dominant processes such as beta-cell dysfunction, lipodystrophy, or obesity, which in turn can influence the effectiveness of drugs targeting these pathways, such as sulfonylureas, GLP-1 receptor agonists, DPP4 inhibitors, and thiazolidinediones [4].  5. **Pharmacogenetic Studies**: Although pharmacogenetic research in diabetes is still developing, studies have shown that genetic variations can significantly impact drug response. For example, variations in the glucose transporter gene SLC2A2 have been associated with differential responses to metformin [8].  6. **Complex Interplay of Factors**: The glycaemic response to treatments like metformin is likely determined by a complex interplay of genetic and environmental factors, with clinical variables such as BMI, drug adherence, and dosing also playing roles [5].  These genetic factors underscore the importance of personalized medicine in diabetes treatment, where understanding an individual's genetic makeup can help tailor more effective therapeutic strategies.",
+  "question": "What are the genetic bases for the varying efficacy of diabetes treatments among individuals?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_10 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_10
new file mode 100644
index 0000000..149e49a
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_10
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2004 - Diabetes Genes a.pdf",
+    "2018 - Animal models of obesity.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2021 -  Monogenic diabetes a gateway to precision medicine.pdf",
+    "2003 - A functional polymorphism in the promoterenhancer region of the FOXP3Scurfin gene associated with type 1 diabetes.pdf",
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "1997 - Development of non-insulin-dependent diabetes mellitus.pdf",
+    "2017 - Differentiation of Diabetes by Pathophysiology.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "1984 - A Polymorphic Locus.pdf"
+  ],
+  "extraction_id": [
+    "0d4b1f7a-5757-5129-835a-6ed0334a1e0e",
+    "f3ffcc7b-7d0c-5f1a-9a5a-762cf6c85403",
+    "5364714e-1909-58f7-a15d-fb60df4e7bee",
+    "9b3c9e69-82ff-5bfc-9b5b-77038a59d8e9",
+    "aacbb5a1-c294-5568-ba02-3d4342091e86",
+    "d5004507-475d-5ac1-9aa6-b5fd664b8bf7",
+    "e27a0679-e4b0-5e9f-8d98-1cd4e8b08b00",
+    "a9accd40-eb89-5595-bf27-b6b82b49f4d4",
+    "4beabe81-e24e-535c-9df3-bfaa9cfdde90",
+    "57f307f8-2493-5438-ad08-b4d85288b94e"
+  ],
+  "document_id": [
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "9e9af9c7-814f-562e-a04d-878528a38002",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "0b6ff786-6a7b-5d24-ba5e-7a61fee7757f",
+    "4a3964a4-0aea-58ee-b749-33e0d8c62228",
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "8598a406-5a79-5f9e-8a1b-bf69daf071bf",
+    "9cfaef1e-fb60-5c2b-94f0-632c89b2eb16",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "df2478dd-6970-5d8d-99e5-bb23f46bcdb1"
+  ],
+  "id": [
+    "chatcmpl-ADZR8nkJvPePGl3BA4Ofgj0JbQyIv",
+    "c3e7c5c2-d8d3-55ba-ba38-b8ca5eb2487a",
+    "24aac9a6-38ed-5a54-b60a-0604d31e5972",
+    "65469d6d-578f-5c78-97cf-d06b2f483dcf",
+    "92749b8d-6c7a-5f2b-a873-a04904cc247b",
+    "18ba0b3f-51a8-5af0-98ec-3b45f1e3219a",
+    "2454130e-8098-5c7f-944b-c5933a8409f8",
+    "856c7a02-c233-5b00-ae1c-55a5e2b1a2ed",
+    "3313b0de-44f4-5cb5-9735-2fefd5ebf0bb",
+    "1e84a9e4-7bd1-51ad-80b5-3a371c090151",
+    "3b29472a-7875-5761-86d5-cbc57c20db85"
+  ],
+  "contexts": [
+    "two broad etiopathogenetic groups. In one group (type I diabetes), the cause  is an absolute deficiency of insulin secretion. Individuals at increased risk of  developing this type of diabetes can often be identified by serological  evidence of an autoimmune process of the pancreatic islets and by genetic  markers. In the second and more prevalent group (type 2 diabetes), the cause  is a combination of resistance to insulin action with inadequate  compensatory insulin secretory response.",
+    "Diabetes mellitus.  Type1 diabetes mellitus (T1DM) and  T2DM have different causes, but both ultimately lead to  pancreatic -cell dysfunction. Damaging the pancreas  chemically or mechanically can induce experimental  diabetes mellitus. Pancreatic damage can be achieved by  surgically removing parts of or all of the pancreatic tissue  (pancreatectomy) to reduce or fully ablate endogenous  insulin production282. The benefit of this method is the  lack of toxic adverse effects (compared with diabetogenic",
+    "Diabetes is a disorder of carbohydrate metabolism charac-terized primarily by hyperglycemia resulting from ineffec-tive uptake of glucose by tissues. Type 1 diabetes is an autoimmune disease that typically occurs early in life and results in total loss of insulin production, whereas type 2 diabetes develops over time as tissues develop a resistance to insulin, and insulin release from the pancreas slowly diminishes. As carbohydrates have the greatest effect on blood glucose of all macronutrients, their",
+    "diabetes but a rare cause of diabetes diag - nosed in childhood or adulthood. Diabetes .  2008;57(4):10341042.  152. Molven A, et al. Mutations in the insulin gene can  cause MODY and autoantibody-negative type 1  diabetes. Diabetes . 2008;57(4):11311135.  153. Gloyn AL, et al. Mutations in the genes encoding  the pancreatic beta-cell KATP channel subunits  Kir6.2 (KCNJ11) and SUR1 (ABCC8) in diabe - tes mellitus and hyperinsulinism. Hum Mutat.  2006;27(3):220231.",
+    "Type 1 diabetes is an autoimmune disease caused by T-cell-mediated destruction of insulin-producing beta cellsin the pancreatic islets of Langerhans (Atkinson andMaclaren 1994). Various aberrations in immune regula-tion have been described in both human patients andanimal models of type 1 diabetes (Rosmalen et al. 2002).A recent study has demonstrated that the disturbance ofcentral and/or peripheral tolerance mechanisms existed indiabetes-prone humans and animals (Sakaguchi 2000).With respect to the",
+    "disorder caused by different factors characterized by a chronic high level of blood sugar with distur-bances to carbohydrate, fat, and protein metabo-lism resulting from defects in insulin secretion, insulin action, or both [  83 ]. Scientists have  divided diabetes into three different types: Type 1 F. Assah and J.C. Mbanya",
+    "(Fig. 1), indicating that insulin resistance and insulin secretory defect played a cooperative role in the development and exac- erbation of diabetes, even though neither was strong enough alone to cause overt diabetes. From another point of view, even if genetically determined insulin resistance itself might not be sufficient for the development of diabetes, insulin resis- tance results in diabetes if pancreatic /H9252 cell function is im- paired genetically (this study) or nongenetically. Development",
+    "tors, and other environmental factors that trigger isletautoimmunity and/or type 1 diabetes. Type 2 Diabetes Type 2 diabetes develops when b-cells fail to secrete suf- cient insulin to keep up with demand, usually in the context of increased insulin resistance. A minority of peo- ple diagnosed with type 2 diabetes also have evidence ofislet autoimmunity (57,58). Obesity is a major risk factor for type 2 diabetes (59,60) with complex genetic and en- vironmental etiology.",
+    "have environmental (islet-injuring drugs or a particular diet) and/or genetic  (monogenic or polygenic) causes. We have grouped the models by cause  and type of diabetes. While this grouping is reasonable and instructive, it  can over-emphasize distinctions. For example, it is believed that beta cell  failure (and/or poor islet regeneration) contributes to type 2 diabetes, but in  their pure, severe form these processes cause type I diabetes.  MODELS OF INSULIN-DEFICIENT DIABETES",
+    "Diabetes mellitus comprises a heterogenous group of disorders that have been classified as either insulin-dependent (IDDM) or non-insulin-depend- ent (NIDDM).1 Their causes are poorly understood but appear to involve some form of interaction between ge- netic and environmental factors.2-4 Some of the environmen- tal factors that can contribute to IDDM include viral infections and chemicals, while obesity is a common predisposing fac- tor for NIDDM. Genes that confer susceptibility or can cause"
+  ],
+  "answer": "Diabetes can be caused by different factors depending on the type.   1. **Type 1 Diabetes**: This type is primarily caused by an absolute deficiency of insulin secretion due to an autoimmune process that destroys insulin-producing beta cells in the pancreatic islets [1], [5]. Genetic markers and serological evidence can often identify individuals at increased risk [1].  2. **Type 2 Diabetes**: This type is caused by a combination of resistance to insulin action and an inadequate compensatory insulin secretory response [1]. It develops over time as tissues become resistant to insulin, and the pancreas's ability to release insulin diminishes [3]. Obesity is a major risk factor for type 2 diabetes, with both genetic and environmental factors playing a role [8].  3. **General Factors**: Both types of diabetes ultimately lead to pancreatic beta-cell dysfunction [2]. Environmental factors, such as viral infections and chemicals, can contribute to the development of diabetes, particularly type 1, while obesity is a common predisposing factor for type 2 diabetes [10].  In summary, diabetes is caused by a combination of genetic, autoimmune, and environmental factors that affect insulin production and action [1], [3], [8], [10].",
+  "question": "What causes diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_11 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_11
new file mode 100644
index 0000000..97776c1
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_11
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2004 - Diabetes Genes a.pdf",
+    "2016 - NIH working group report using genomic information to guide weight management  From universal.pdf",
+    "2012 - Diabetes and Hypertension Is There a Common Metabolic Pathway.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "2004 - Diabetes Genes b.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "2005 - Metabolic Syndrome A Comprehensive Perspective Based  on Interactions Between Obesity Diabetes and Inflammation.pdf",
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "2018 - Type 2 Diabetes in adolescents and young adults.pdf"
+  ],
+  "extraction_id": [
+    "ad5bdba5-b3c6-50ac-a4b3-3089e7bed0da",
+    "bbaf5afd-56e0-5ded-bf17-c8c36d67122c",
+    "e0a47978-ae34-5905-baff-36a3364d21af",
+    "424d7751-3dbf-5e10-83ca-12101841d17c",
+    "cef57178-c218-52d3-b049-aa6ca097fd73",
+    "ad5bdba5-b3c6-50ac-a4b3-3089e7bed0da",
+    "c6cfb382-639a-5dd4-a9c8-c8f57b6daabc",
+    "f7fe5916-4f25-5740-8737-f668f216575d",
+    "4657f231-5e0c-5572-ad75-22c74f55a70f",
+    "6c730685-6ec0-52a4-8f33-671a39616a86"
+  ],
+  "document_id": [
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "0ee28c8a-3618-559e-be0a-30f2579a0d1f",
+    "37b08243-09de-5a78-b2bb-1eade3c714af",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "da2f2624-e3e6-5e2d-b406-941db2fe7671",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "de2aa54c-eb0f-5dc3-ac92-23ee3215dd2a",
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "cc708325-df0a-55ec-9e9b-2bf97835c992"
+  ],
+  "id": [
+    "chatcmpl-ADZRC8tLslwOrPHuuXyMSXCo1Prfp",
+    "5479de8e-2994-5b99-a0a7-915840f1de0d",
+    "cdd1e1cb-6b89-5045-96e2-280f6d615ab4",
+    "50f8c1de-8641-5cb6-8080-620f15810922",
+    "9da658e9-223d-527d-a913-b1d8eac31de2",
+    "e317892f-8310-5414-869e-b759258b2eeb",
+    "a3060853-46b0-506a-b3ed-9e85c2c450da",
+    "8c4e8b2c-6730-541c-8a2e-22fbd7ddb487",
+    "09070d01-4946-559c-9b44-f502c7b066c3",
+    "eb818d5f-6b01-53ef-8343-1823c449f779",
+    "e08c0b4a-24f9-576e-b5cf-74641fe81fd0"
+  ],
+  "contexts": [
+    "2 diabetes suggest that regular exercise might play an important role in  decreasing the very high incidence of premature coronary artery disease.  Although there are no randomized controlled trials assessing reduction  in cardiovascular events induced by physical activity in type 2 diabetes,  available evidence is consistent with the concept that physical activity may play  an important role in reducing cardiovascular risk in type 2 diabetes. 44 Large",
+    "tern of weight change impact health. For example, in the DiabetesPrevention Program (DPP; described in more detail later), both short- and intermediate-term weight loss were associated with reduced diabetes risk and intermediate cardiometabolic risk factor levels, whereas weight cycling (defined as number of 5 lb [2.25 kg] weight cycles) raised diabetes risk, fasting glucose levels, insulinresistance, and systolic blood pressure. Initial (baseline to 1 month)",
+    "sclerosis Risk in Communities (ARIC) study, the highestquartile of leisure activity (primarily cycling and walking)had a 34% lower odds of developing hypertension over 6 years compared to the least active [ 107]. Thus, physical activity reduces the risk of developing diabetes and hyper- tension. The mechanism involves changes in body weight and glucose tolerance, as well as other factors [ 107]. The effect of obesity susceptibility genes on the onset of",
+    "exercise can reduce the incidence of type 2 diabetes. Tuomilehto and  coworkers demonstrated that the individuals on a consistent diet and exercise  program had 10% incidence of diabetes during 4 years of follow-up  compared to 22% for patients in the control group, who met only once a year  with the dietician and the physician.40 A six-year randomized trial conducted  by Pan and colleagues demonstrated that exercise resulted in 46% reduction",
+    "Exercise  Exercise has been shown to prevent development of Type 2  diabetes in high-risk groups. A number of studies have looked at the effect  of insulin on delaying the onset of diabetes. In a study of 5990 male  alumni from an American university followed over 10 years, 202 pts (3.3  percent) developed Type 2 diabetes mellitus. The relative risk was lower  in patients who exercised regularly even when adjusted for obesity,  hypertension, and a family history of diabetes. The benefit was greatest in",
+    "nonrandomized studies of both men and women with type 2 diabetes and  impaired glucose tolerance have found that physical activity is associated with  a decreased risk for cardiovascular disease. It also appears that the amount of  physical activity is inversely associated with coronary events.5354  RISK OF EXERCISE IN PATIENTS WITH DIABETES  The risks associated with exercise can be divided into metabolic,  vascular, neurologic and musculoskeletal (Table 4).",
+    "74 The mechanism underlying this effect of exercise is not known;however, it is noteworthy that lifestyle change is a very effectiveway to reduce the rate of development of diabetes in a predia-betic population, as shown by the diabetes prevention study. 75,76 Both a reduction in macronutrient intake and exercise cause areduction in inflammation. References 1. Reaven GM. Banting lecture 1988. Role of insulin resistance in human disease. Diabetes . 1988;37:15951607.",
+    "physical training on carbohydrate metabolism and associated cardiovascular risk factors in patients with diabetes. Diabetes Rev. 1995;3:378407.      23.    Rogers MA, Yamamoto C, King DS, Harberg JM,  Ensani AA, Holloszy JO. Improvement in glucose tolerance after one week of exercise in patients with mild NIDDM. Diabetes Care. 1988;11:6138.      24.    Eriksson KF, Lindgarde F. Prevention of type 2 dia- betes mellitus by diet and physical exercise. Diabetologia. 1991;34:8918.",
+    "migrant and other observational studie!f86970 and prospective studies in subjects  at high risk for developing type 2 diabetes.717273 Recently, large interventional  trials have reinforced the benefits of exercise in reducing the risk for type 2  diabetes. These include the Malmo study from Sweden45, the Da Quing study  from China74 and the recently concluded Finnish Diabetes Prevention Study.75  These prospective but not randomized studies show a reduction in the risk of  560",
+    "reduce systolic blood pressure, reduce total cholesterol, raise HDL cholesterol, and improve endothelial function in overweight patients with young-onset type 2 diabetes. 47  However, any potential benefits to the cardiovascular disease risk profile are lost within 36 months after cessation of exercise training, and do not confer protection against later cardiovascular events. 47,121 Additionally,  reviews49,121,122 of the limited number of studies done to"
+  ],
+  "answer": "Yes, cycling, as a form of physical activity, can reduce the risk of diabetes. According to the context provided:  - The Atherosclerosis Risk in Communities (ARIC) study found that higher levels of leisure activity, which primarily included cycling and walking, were associated with a 34% lower odds of developing hypertension over six years compared to the least active group. This study also indicated that physical activity reduces the risk of developing diabetes and hypertension through mechanisms involving changes in body weight and glucose tolerance [3]. - Additionally, other studies have shown that exercise, including activities like cycling, can prevent the development of Type 2 diabetes in high-risk groups [5].  Thus, engaging in regular cycling can contribute to a reduced risk of developing diabetes.",
+  "question": "Does cycling reduce risk of diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_12 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_12
new file mode 100644
index 0000000..4425d61
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_12
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2008 - Glossary of Genetics Genomics Terms.pdf",
+    "2015 - The genetics of diabetic complications.pdf",
+    "2009 - From Disease Association to Risk Assessment.pdf",
+    "2014 - Identification of novel risk genes associated with type 1 diabetes mellitus.pdf",
+    "2017 - Machine Learning and Data Mining Methods in Diabetes Research.pdf",
+    "1994 - Genetic Predisposition to Diabetic Nephropathy.pdf",
+    "2007 - Network-Based Analysis.pdf",
+    "2007 - Network-Based Analysis.pdf",
+    "2007 - Network-Based Analysis.pdf",
+    "2008 - High-Density Single Nucleotide Polymorphism.pdf"
+  ],
+  "extraction_id": [
+    "53e868dd-b318-5cf3-8b2e-98a548aab7cf",
+    "27de21d5-8e86-5233-8196-ff09c1916eb8",
+    "6f819601-6eea-54a4-ab88-27e1b0602287",
+    "cce6eb13-6c59-5916-a108-477128ed6912",
+    "46f1cae6-a01f-5445-b20f-0eadf892f8bf",
+    "9f693d00-d331-5924-a0bc-8ec8614ccd6e",
+    "295a5916-ff2d-54b2-a0fe-4b279e71b5ad",
+    "334686b0-71fb-5820-9649-3cdf355f1dfe",
+    "63fadd0c-3522-5339-b534-807144901fa9",
+    "e04a055d-630a-50b5-a468-560e25bf1d40"
+  ],
+  "document_id": [
+    "c66d2572-071d-5aaf-829c-b3ca6cf6d697",
+    "5bc1f058-caf2-5cb4-9623-b1d04b074a3c",
+    "a61066d0-0d1a-5f10-96c3-aa96bacdad5e",
+    "97fe33b0-a6c7-59b6-bd34-05528e77293f",
+    "e2dcbb80-5ad7-5441-b170-9b46607445b0",
+    "e9dce475-d40f-5cda-a53d-3e722191d447",
+    "1f23601c-2dab-570a-a2ca-039283831b17",
+    "1f23601c-2dab-570a-a2ca-039283831b17",
+    "1f23601c-2dab-570a-a2ca-039283831b17",
+    "ce79f562-c274-5cbe-bae2-e5b688348b04"
+  ],
+  "id": [
+    "chatcmpl-ADZRGfPVf97ITQ8FTyJHnwW98zLJf",
+    "54ff4672-bf7f-5158-b228-ca3d45e0cb0d",
+    "8a7d2ffb-20b3-572a-99af-ec120e268bd3",
+    "506f423f-23f7-5d72-b614-1ccc9b38e853",
+    "d84c2221-5d5b-5df9-bafd-2dd17e9fb132",
+    "069b11b5-0785-599a-b92e-543e133c1c65",
+    "f0848c4e-5c55-5e13-8ac5-75065aaed286",
+    "9c16d623-9eb8-57fe-8ae8-48009f766d64",
+    "9fd6df7d-c275-573d-8c8e-afe69ec5c544",
+    "5005ed0b-8b17-540f-8106-94593c601084",
+    "81a23927-18e2-54fe-94c2-6b64cc3c7020"
+  ],
+  "contexts": [
+    "Genetic factors appear to play a role in determining an individuals risk of developing diabetes. It is hoped",
+    "Diabetes (GoKinD) study: a genetics collection  available for identifying genetic susceptibility  factors for diabetic nephropathy in type1  diabetes. J. Am. Soc. Nephrol.  17, 17821790  (2006). 137. Scott, R.A. etal. Large-scale association  analyses identify new loci influencing glycaemic  traits and provide insight into the underlying  biological pathways. Nat. Genet.  44, 9911005  (2012). Author contributions All authors researched the data for the article,",
+    "identifying genetic susceptibility factors for diabetic nephropathy in type 1 diabetes. J Am Soc Nephrol 17: 17821790. 44. Manolio TA, Rodriguez LL, Brooks L, Abecasis G, Ballinger D, et al. (2007) New models of collaboration in genome-wide association studies: the Genetic Association Information Network. Nat Genet 39: 10451051. 45. Mailman MD, Feolo M, Jin Y, Kimura M, Tryka K, et al. (2007) The NCBI dbGaP database of genotypes and phenotypes. Nat Genet 39: 11811186.",
+    "in Diabetes (GoKinD) study: a genetics collection availablefor identifying genetic susceptibility factors for diabeticnephropathy in type 1 diabetes. J Am Soc Nephrol 2006; 177: 1782 1790. 10. Pezzolesi MG, Poznik GD, Mychaleckyj JC, et al. Genome- wide association scan for diabetic nephropathysusceptibility genes in type 1 diabetes. Diabetes 2009; 586: 14031410. 11. Paterson AD, Lopes-Virella MF, Waggott D, et al.",
+    "beta cell function, insulin mode of action, glucose  metabolism and/or other risk  factors.  It is a fact that advances in genotyping technology, over the past few years,  have facilitated rapid progress in large-scale gene tic studies.   Identification of a large  number of novel genetic variants increasing suscept ibility diabetes and related traits  opened up opportunities, not existing thus far, to associate this genetic information",
+    "DISCUSSION The findings of previous epidemiological and family studies suggest that diabetic nephropathy results from an interaction between metabolic abnormalities that are typical of poorlycontrolled IDDM and predisposing genetic factors (4,5). Thenature of the genetic factors, however, has remained un- known (22). Using a candidate gene approach, we have found in this",
+    "PLoS Genetics | www.plosgenetics.org June 2007 | Volume 3 | Issue 6 | e96 0963 Type 2 Diabetes Network-Based Analysis",
+    "PLoS Genetics | www.plosgenetics.org June 2007 | Volume 3 | Issue 6 | e96 0971 Type 2 Diabetes Network-Based Analysis",
+    "PLoS Genetics | www.plosgenetics.org June 2007 | Volume 3 | Issue 6 | e96 0967 Type 2 Diabetes Network-Based Analysis",
+    "High-Density Single Nucleotide Polymorphism Genome-Wide Linkage Scan for Susceptibility Genes forDiabetic Nephropathy in Type 1 Diabetes Discordant Sibpair Approach John J. Rogus,1,2G. David Poznik,1Marcus G. Pezzolesi,1,2Adam M. Smiles,1Jonathon Dunn,1 William Walker,1Krzysztof Wanic,1,2Dariusz Moczulski,1,2,3Luis Canani,1,2,4Shinichi Araki,1,2,5 Yuichiro Makita,1,2,6James H. Warram,1and Andrzej S. Krolewski1,2 OBJECTIVE Epidemiological and family studies have demon-"
+  ],
+  "answer": "GeneNetwork can assist in identifying genetic factors involved in diabetes by leveraging large-scale genetic studies and databases that focus on genetic susceptibility factors. For instance, the Diabetes (GoKinD) study provides a genetics collection specifically aimed at identifying genetic susceptibility factors for diabetic nephropathy in type 1 diabetes [2], [4]. Additionally, advances in genotyping technology have facilitated rapid progress in large-scale genetic studies, enabling the identification of numerous novel genetic variants that increase susceptibility to diabetes and related traits [5]. These resources and technological advancements can be utilized within GeneNetwork to analyze genetic data and identify potential genetic factors involved in diabetes.",
+  "question": "How can GeneNetwork assist in identifying genetic factors involved in diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_13 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_13
new file mode 100644
index 0000000..8a4a369
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_13
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2020 - Gene network a completely updated tool for systems genetics analyses.pdf",
+    "2010 - Systems genetics analyses predict a transcription role for P2P-R Molecular confirmation that P2P-R is a transcriptional co-repressor.pdf",
+    "2008 - Genetic Analysis of Posterior Medial Barrel Subfield Size.pdf",
+    "2011 - Peroxisomal L-bifunctional enzyme (Ehhadh) is essential for the production of medium-chain dicarboxylic acids.pdf",
+    "2009 - Metabolomics Applied to Diabetes Research.pdf",
+    "2020 - GeneNetwork a toolbox for systems genetics.pdf",
+    "2017 - GeneNetwork a toolbox for systems genetics.pdf",
+    "2011 - Using the PhenoGen Website for \u201cIn Silico\u201d Analysis of Morphine-Induced Analgesia Identifying Candidate Genes.pdf",
+    "2015 - Cell cycle gene expression networks discovered using systems biology Significance in carcinogenesis.pdf",
+    "2013 - Pathways, Networks and Systems Medicine Conferences.pdf"
+  ],
+  "extraction_id": [
+    "858f630f-9443-5f13-ac40-8e16eadd9ba1",
+    "ec624ebb-489a-5437-a721-f01cf981d0a7",
+    "66aad1b1-a76d-58a8-aa40-76a6b58c4964",
+    "a8b40857-7ae8-512a-9817-bea1ae3345ba",
+    "380e9a2e-8f9f-5f9e-ba20-3695b1c60fda",
+    "4ca2fc9e-7d42-5ea3-b1b7-a296bfbc6a09",
+    "7dd82b3f-58bd-5915-9eea-250f11412ff2",
+    "0e3a5e40-06b0-58d4-b495-3093954ed17b",
+    "5b6d04d2-3aa2-5a43-814a-b13e60e3bb1d",
+    "9ca6d444-064c-5743-b029-9d634685f11b"
+  ],
+  "document_id": [
+    "128224f1-3545-52c3-93cb-77c3cf4ec70a",
+    "e4d1e2e9-f267-5814-8c7b-dc11d7eec9bf",
+    "76a715a4-8222-598b-8e65-6d5b6e807989",
+    "ac61753e-bcb2-55c3-804b-e821e3d1a4ad",
+    "a6ae2fb6-88ae-588f-a98d-b6092f886ed9",
+    "d11a87ca-4989-59af-95e3-ab90af7d9212",
+    "682c3a51-0aa5-54a3-a6e7-a09b81c0e8b6",
+    "eb266fa1-8dec-5c56-a3d5-b508bd6bd448",
+    "6f354254-4f4d-52ad-bed7-9356f43c0b20",
+    "b50a9732-7d01-5d4d-8f33-a9d43dbc7df3"
+  ],
+  "id": [
+    "chatcmpl-ADZRLVC30o2qvIhM1bclRsts27OFA",
+    "c63cfaee-749e-547b-9c0a-086266f10670",
+    "a9508122-3b14-5365-979c-ba580bdcb78f",
+    "21936758-94b1-506f-9229-77e26001ae44",
+    "8b8a24da-a175-5cb8-91bd-8966fca5d344",
+    "418060c8-fafb-5010-a512-55819ed36a3d",
+    "7ce6c0fe-8b0a-5ce9-83d1-6e6b99b4f24d",
+    "30e2423f-2b2b-5c7d-8808-b025242fa0c7",
+    "fa07b1bf-94e6-515b-8400-cf3afa8b8741",
+    "dcb29dfe-ba22-54bc-91f7-af3261a18fd2",
+    "f163b61d-987b-50eb-aef2-ee0dc0eddb9f"
+  ],
+  "contexts": [
+    "GeneNetwork have reinvigorated it, including the addition  of data from  10 species, multi -omics  analysis, updated code, and new tools. The new GeneNetwork is now an exciting resource for  predictive medicine and systems genetics, which is constantly being maintained and improved.    Here, we give a brief overview of the process  for carrying out some of the most common  functions on GeneNetwork, as a gateway to deeper analyses , demonstrating how a small",
+    "GeneNetwork http://www.genenetwork.org is anexample of a bioinformatics tool that can be used to explore systems genetics data. The importance of defining biological networks and predicting molecular interactions has been emphasized by several reports [1,2]. Such studies emphasize that when knowledge about DNA variation within popula- tions is interfaced with data on gene expression, protein interactions and DNA-protein binding, biological networks can be constructed that are predictive of the",
+    "GeneNetwork provides users with an array of analyticaltools to compare a given trait with a number of data setsavailable from other experimenters. Microarray data ofgene expression in the brain and data of other phenotypes are two such examples of possible tools. For this study, we",
+    "subnetworks   GeneNetwork (www.genenetwork.org) is a depository of data- sets and tools for use in complex systems biology approaches in  order to generate or predict higher order gene function ( 23, 24 ).",
+    "of these tools to diabetes andmetabolic disease research at the cellular, animal model,and human disease levels are summarized, with a partic-ular focus on insights gained from the more quantitativetargeted methodologies. We also provide early examplesof integrated analysis of genomic, transcriptomic, andmetabolomic datasets for gaining knowledge about meta-bolic regulatory networks and diabetes mechanisms andconclude by discussing prospects for future insights.",
+    "including correlation and network analysis to compare associations  between tissues and between other rodent or human data sets[32] Many of the Data Sets are amenable to systems genetics mapping and other methods and are accessible at GeneNetwork. The Description and Usage column provides details about the data set and potential",
+    "including correlation and network analysis to compare associations  between tissues and between other rodent or human data sets[32] Many of the Data Sets are amenable to systems genetics mapping and other methods and are accessible at GeneNetwork. The Description and Usage column provides details about the data set and potential",
+    "data are entered into GeneNetwork after they have been shepherded through a system like PhenoGen that has extensive capabilities for normalization and quality control. A comparison of the brain gene expression datasets and some of the tools for data analysis available on PhenoGen and GeneNetwork is shown in Table 3, and more detailed information on features provided by each site is outlined in the Supplementary DiscussionHoffman et al. Page 5 Addict Biol . Author manuscript; available in PMC 2012 July 1.",
+    "of importance in the emergence of precision medicine ( Curtis, 2015 ; Desautels et al., 2014 ;  Glade Bender et al., 2015 ; Jorgensen, 2015 ; Kummar et al., 2015 ; Marquet et al., 2015 ;  Rubin, 2014 ) wherein therapeutic strategies need to be aligned with specific properties of  tumors. Methods GeneNetwork and WebGestalt GeneNetwork is an open access, online data analysis resource for systems biology and  systems genetics. It contains a large number of microarray datasets from multiple tissues of",
+    "results in applying the method  to type 2 diabetes mellitus suggest it may hold  promise as a useful research tool for complex diseases .  Further details on the  methodol ogy is available from the following paper: Liu M, Liberzon A, Kong  SW, Lai WR, Park PJ et al (2007) Network -based analysis of affected biological  processes in type 2 diabetes models.  PLoS Genet 3(6):e96.   doi:10.1371/journal.pgen.0030096."
+  ],
+  "answer": "GeneNetwork offers several specific tools that are particularly useful for diabetes research:  1. **Multi-Omics Analysis**: GeneNetwork includes data from multiple species and supports multi-omics analysis, which integrates genomic, transcriptomic, and metabolomic datasets. This integration is crucial for understanding the complex regulatory networks and mechanisms underlying diabetes [1], [5].  2. **Correlation and Network Analysis**: These tools allow researchers to compare associations between tissues and between rodent or human datasets. This can help identify key molecular interactions and pathways involved in diabetes [6], [7].  3. **Systems Genetics Mapping**: Many datasets within GeneNetwork are amenable to systems genetics mapping, which can be used to link genetic variations to phenotypic traits related to diabetes. This helps in identifying genetic factors that contribute to the disease [6].  4. **Microarray Data Analysis**: GeneNetwork provides access to a large number of microarray datasets from multiple tissues. This data can be used to study gene expression patterns and identify genes that are differentially expressed in diabetic conditions [3], [9].  5. **Predictive Medicine and Systems Genetics**: GeneNetwork is designed as a resource for predictive medicine and systems genetics, which are essential for developing personalized treatment strategies for diabetes [1].  These tools are applied in diabetes research to gain insights into the genetic and molecular basis of the disease, identify potential biomarkers, and develop targeted therapeutic strategies. For example, integrated analysis of genomic, transcriptomic, and metabolomic datasets can reveal metabolic regulatory networks and mechanisms specific to diabetes [5]. Correlation and network analysis can help in understanding how different tissues interact and contribute to the disease pathology [6], [7].",
+  "question": "What specific tools within GeneNetwork are most useful for diabetes research, and how are they applied?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_14 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_14
new file mode 100644
index 0000000..f56cff3
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_14
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Integrative Analysis of a Cross-Loci Regulation Network.pdf",
+    "2019 - IRS1\u2010 rs10498210 GA and CCR5\u201059029 AG polymorphisms in patients with type 2 diabetes in Kurdistan.pdf",
+    "2014 - Pathophysiology and treatment of type 2 diabetes.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2000 - Pathophysiology and Pharmacological Treatment.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "2004 - Diabetes Genes a.pdf"
+  ],
+  "extraction_id": [
+    "63fe12a0-20b1-5f8b-9fd6-adaecaa66eeb",
+    "5b74e0f4-8b7a-5ef2-ab41-99819fe185cc",
+    "b4a31e40-c59e-525c-afcf-6f1efae2ef3a",
+    "4cf7634b-caa6-589c-939d-3bf8d9410e46",
+    "4cf7634b-caa6-589c-939d-3bf8d9410e46",
+    "3faeb0aa-9894-58e7-a2a6-c5f9ceb5cd22",
+    "8bbb1489-4c01-5d58-99f3-c007c9e4713a",
+    "ae5e854d-a344-56f3-a597-fb8e1e745af7",
+    "77878a83-0136-583f-a237-5443ce967a1a",
+    "a7e3dee0-551a-57ea-8a22-04b5f84252da"
+  ],
+  "document_id": [
+    "de5b6ecc-8a65-5dc7-9a4a-49e6df22472e",
+    "18afbfee-ddee-54b3-88cc-342812a65d09",
+    "ab9288ab-e3ad-58f1-b5ba-183ee17ce4bd",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "27206293-61c6-5a22-a342-6dfe1a759f04",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa"
+  ],
+  "id": [
+    "chatcmpl-ADZRS3ZXELoaA2dkAlgxXgB5ABMRV",
+    "1da7c09b-640d-5791-b411-4f87dfc2e03c",
+    "7b13b8a2-7d7f-5675-b3f0-c9aa48566b45",
+    "089f1f4d-d9d9-5245-b7ca-cf6e2ff87111",
+    "dd5661a5-6349-5abf-95c6-4abe424ada78",
+    "4f477c98-cd75-5955-ac3e-72a44d53f45e",
+    "2ace6cbd-084e-57e5-b6ed-1f42527b59b9",
+    "eafb2d74-ca99-52c6-a93d-89711b063a53",
+    "63d20d73-d6a3-56cb-91f6-34c66f439c7f",
+    "39b2638a-6bc8-5d16-bde4-4abd573cca47",
+    "8240c5cc-7d4f-51b4-935a-eb47314b3ec0"
+  ],
+  "contexts": [
+    "Figure 3. Schematics view of insulin regulation. Elevated glucose level by either food intake or liver glycogenolysis is sensed by islet and leads to insulin secretion to the bloodstream. The increased insulin stimulates peripheral tissues to absorb glucose, and as a consequence, the glucose le vel",
+    "plays an important role in regulating insulin secretion in beta cells of the pancreas. It has been shown that glucosestimu-lated insulin secretion may be triggered by the autocrine ac-tivation of the insulin signaling pathway, including insulin receptor phosphorylation, tyrosine phosphorylation in IRS1, and the activation of PI3Kinase. Putting together these data leads to the hypothesis that a single molecular impairment in the pathway of insulin signaling, including an incomplete interaction between",
+    "(A) Insulin interacts in the liver to suppress glucose production, and in muscle and adipose tissue to stimulate uptake of glucose, aminoacids, and fatty acids. The amount of insulin released to maintain normal glucose homoeostasis is established by prevailing insulin sensitivity. This feedback is probably mediated through neuronal and humoral mechanisms, but exact mediators are still not known. (B) When insulin resistance develops in insulin-sensitive tissues, feedback to  cells ensures that the cells",
+    "Insulin Action   In healthy, normal individuals, blood glucose concentra- tion is maintained within a narrow range. After an over-night fast or between meals, blood glucose normally falls within the range of 3.5  5.5   mM. Immediately after a meal containing carbohydrate, blood glucose concentration rises to a peak of 6  10   mM followed by a sharp decline back to baseline within 60 minutes. This exquisite control is achieved by a   ne balance between glucose absorption",
+    "from the gut, glucose production by the liver, and glucose extraction from the blood into the cells and tissues.   Insulin plays a central role in the regulation of blood",
+    "glucose transport into the cell. Concomitantly, insulin stimulates intracellular utili-zation of glucose by many other tissues as well. In the fasting state, the main physiological function of insulin is to suppress glucose production by the liver and prevent uncontrolled lipolysis and ketogenesis, without which dia-betic ketoacidosis would quickly develop. Hence, if either of these aspects of insulin action is impaired, then periph-eral or liver hepatic insulin resistance or both are said to be present.",
+    "and suppression ofglucose production are regulated by insulin.",
+    "the pancreas in response to an increase in blood glucose, such as that which follows a carbohydrate - containing meal. Insulin acts to decrease blood glucose levels by increasing glucose uptake by tissues and by decreasing gluconeogenesis by the liver. To increase tissue uptake, insulin triggers the translocation of GLUT4 receptors to the cell surface in skeletal muscle and adipose tissue. Insulin also stimulates each of the regulatory enzymes in the glycolytic pathway, while also inhibiting the key",
+    "insulin suppresses both hepatic and renal glucose release, 3031 and stimulates  glucose uptake exogenous insulin administration causes systemic glucose  utilization to exceed systemic glucose release so that plasma glucose  concentrations decrease.  As the plasma glucose levels decrease there is a characteristic  hierarchy of responses (Figure 1 ). Reduction of insulin secretion, the first in  the cascade of hypoglycemia counterregulation, 2 derepresses glucose",
+    "Counter-regulatory hormones antagonize the glucose lowering action  of insulin, and act to raise the blood glucose level. Glucagon, a potent  counter-regulatory hormone inhibited by insulin, is secreted from pancreatic  alpha cells when cells perceive low glucose. In diabetes, pancreatic insulin  levels are reduced and glucagon is chronically elevated. In DKA, in addition  to low insulin action, there is the cellular perception of low glucose , which"
+  ],
+  "answer": "Insulin plays a central role in the regulation of blood glucose levels by facilitating the uptake of glucose into peripheral tissues and suppressing glucose production by the liver. When blood glucose levels rise, such as after food intake, insulin is secreted by the pancreas into the bloodstream [1]. This secretion stimulates peripheral tissues, including muscle and adipose tissue, to absorb glucose, thereby lowering blood glucose levels [1], [3], [8].   Insulin also acts in the liver to suppress glucose production and gluconeogenesis, further contributing to the reduction of blood glucose levels [3], [6], [8]. Additionally, insulin triggers the translocation of GLUT4 receptors to the cell surface in skeletal muscle and adipose tissue, enhancing glucose uptake by these tissues [8]. This finely tuned balance between glucose absorption, production, and extraction ensures that blood glucose levels are maintained within a narrow range [4], [5].",
+  "question": "What role does insulin play in the regulation of blood glucose levels?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_15 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_15
new file mode 100644
index 0000000..45ce004
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_15
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2017 - Regular exercise participation improves genomic stability in diabetic patients an exploratory study to analyse telomere length and DNA damage.pdf",
+    "2017 - Age at natural menopause and risk of type 2 diabetes a prospective cohort study.pdf",
+    "2018 - Type 2 Diabetes in adolescents and young adults.pdf",
+    "2012 - Meta-Analysis of the Relationship between Common.pdf",
+    "2011 - Interaction Between Exercise and Genetics.pdf",
+    "2017 - Differentiation of Diabetes by Pathophysiology.pdf",
+    "2018 - Type 2 Diabetes in adolescents and young adults.pdf",
+    "2018 - Type 2 Diabetes in adolescents and young adults.pdf",
+    "2018 - Type 2 Diabetes in adolescents and young adults.pdf",
+    "2015 - A Chromosome 13 locus is associated with male-specific mortality in mice.pdf"
+  ],
+  "extraction_id": [
+    "0e53122e-a308-55f7-8ee8-a0857ac9c52f",
+    "7d48e6ab-4fe7-539f-beee-b65ee92cb187",
+    "ec07ba94-177b-599e-8104-92b2e0b943e6",
+    "a1598191-38cd-5ece-ad1f-24f100ce43a4",
+    "3c217c69-d626-541a-a846-3277ddbf3e12",
+    "47e113ad-dfd3-5a57-be05-7cbb1457f0b6",
+    "ec07ba94-177b-599e-8104-92b2e0b943e6",
+    "18efcee1-5d44-55fc-9d04-e8caaf1a13ab",
+    "ea61883c-6516-509f-b336-e50da1272f69",
+    "5cc56e3b-53ab-5299-814d-014e2ed31d2f"
+  ],
+  "document_id": [
+    "dcaf7b09-2d54-5cbf-b061-e3c4e6c6c518",
+    "7457a250-85c1-520b-a177-f7775a60b14f",
+    "cc708325-df0a-55ec-9e9b-2bf97835c992",
+    "0135b1a8-aa80-5d2f-a64c-daf7806868bd",
+    "c36db75e-4b76-540d-9efb-d0e156e61541",
+    "9cfaef1e-fb60-5c2b-94f0-632c89b2eb16",
+    "cc708325-df0a-55ec-9e9b-2bf97835c992",
+    "cc708325-df0a-55ec-9e9b-2bf97835c992",
+    "cc708325-df0a-55ec-9e9b-2bf97835c992",
+    "ad8f2626-87fb-520e-8cef-ee9a9cc3ab0b"
+  ],
+  "id": [
+    "chatcmpl-ADZRXnhQYKwt6bdryLveg63fKEuX4",
+    "a2125212-46c1-5020-ab62-196db673c2f5",
+    "df41de1c-0fa9-5a3b-a610-20c98a3f2b2d",
+    "c03fd521-abb7-53c7-a2f9-61fea38ca541",
+    "f2086a09-1ba9-5cc7-8573-3a3c654ee2c6",
+    "dee77427-41b1-5ebb-a50c-8d29bc152f88",
+    "765594a9-3100-57fb-a061-3d72b771fce7",
+    "2747746a-5ede-5c96-aa43-8b71a29c1ce0",
+    "c0c7ac1b-e32e-5971-a63c-7d2ee7b48d5f",
+    "91e2be43-bd18-5304-9755-3ee18dcf0040",
+    "13911a7c-bc0a-5643-bbe6-f4ec9a9a2292"
+  ],
+  "contexts": [
+    "The biological processes linking aging and disease risk are poorly understood. Still, aging is considered to date as  one of the main factors responsible for several complex diseases including cancer, cardiovascular diseases, and diabetes. Particularly, type 2 diabetes (T2D) has become very prevalent all over the world, with a projected increas- ing growth rate for the years ahead 1. The pathophysiological mechanism that underlines diabetic complications",
+    "unclear whether age at menopause is associated with risk of type2d i a b e t e s[ 3,4]. Data from cross-sectional studies examining the association between age at menopause and type 2 diabetes are contradictory, with a few studies reporting no association and some other reporting higher odds of having type 2 diabetes with early onset of menopause [ 57]. Recently, a nested case cohort study reported that an increased risk of type 2 diabetes is associ-",
+    "The mechanisms leading to development of type 2 diabetes in young people are similar to those in older patients; however, the speed of onset, severity, and interplay of reduced insulin sensitivity and defective insulin secretion might be different in patients who develop the disease at a younger age. 18 In adolescents with type 2 diabetes, as in  later onset type 2 diabetes, the initial deterioration in -cell function is characterised by loss of first-phase nutrient-stimulated insulin secretion.",
+    "anincreased risk of developing type 2 diabetes (T2D) later in their",
+    "T2D is associated with age, and Western populations are aging rapidly. The second major explanation is our lifestyles have changed dramatically in recent years. Epidemiological studies have identified strong T2D risk relationships for obesity, sedentary behavior [24], and diets rich in energy [5], processed carbohydrates [6], and animal fats [7]. Collectively, these lifestyle factors impede the actions of insulin and raise hepatic glucose production, which can result in the diminution of endog-enous",
+    "tion. Many people with type 2 diabetes ultimately requirei n s u l i nt h e r a p y ,w h i c hr e ects long-standing type 2 diabetes and greatly diminished b-cell function but also likely includes individuals who have slowly progressingautoimmune diabetes with adult onset (LADA) or otherambiguous forms of diabetes. Age. Data from randomized controlled trials in people with type 2 diabetes under the age of 18 years or over the age of 65 years are scarce. Bene cial effects of tight",
+    "strong inverse association between BMI and age at diagnosis of type 2 diabetes. When type 2 diabetes presents in later life, the severity of insulin resistance is often greater among individuals with a history of protracted and severe obesity, particularly with excess visceral adiposity. 28",
+    "patients with young-onset type 2 diabetes than in patients without diabetes, whereas the risk of myocardial infarction was much less (typically 24 times higher) in patients with  type 2 diabetes presenting in middle and later life. 106 In  Hong Kong, where 20% of type 2 diabetes diagnosed since  1995 occurs in people aged 40 years or younger, a 7-year prospective study 107 showed that when adjusted for age,",
+    "type 2 diabetes, the major predisposing risk factors are obesity, family history, and sedentary lifestyle. Onset of diabetes at a younger age (defined here as up to age 40 years) is associated with longer disease exposure and increased risk for chronic complications. Young-onset type 2 diabetes also affects more individuals of working age, accentuating the adverse societal effects of the disease. Furthermore, evidence is accumulating that young-onset type 2 diabetes has a more aggressive disease phenotype,",
+    "pathophysiology of type 2 diabetes. Diabetes 60(10):26242634. doi:10.2337/db11-0415Aging Clin Exp Res 123"
+  ],
+  "answer": "Aging is considered one of the main factors responsible for the development of type 2 diabetes (T2D) [1]. As populations in Western countries are aging rapidly, the prevalence of T2D is increasing [5]. This is partly due to lifestyle changes that impede insulin action and increase hepatic glucose production, which are more pronounced in older adults [5]. Additionally, the severity of insulin resistance tends to be greater in older individuals, especially those with a history of prolonged and severe obesity [7]. Therefore, aging significantly increases the risk of developing type 2 diabetes.",
+  "question": "How does aging affect the risk of developing type 2 diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_16 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_16
new file mode 100644
index 0000000..baa98fa
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_16
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "2018 - Global aetiology and epidemiology of type 2 diabetes mellitus and its complications.pdf",
+    "2014 - Pathophysiology and treatment of type 2 diabetes.pdf",
+    "2011 - Lifestyle and Genetics in Obesity and type 2 Diabetes.pdf",
+    "2010 - Interactions of Dietary Whole-Grain Intake.pdf",
+    "2008 - Public Health Genomics Approach to Type 2 Diabetes.pdf",
+    "2009 - Zinc and Diabetes - clinical links and molecular mechanisms.pdf",
+    "2011 - Type 2 diabetes across generations from pathophysiology to prevention and management.pdf",
+    "2007 - Physical activity modifies the effect of SNPs in the SLC2A2 (GLUT2).pdf",
+    "2011 - Lifestyle and Genetics in Obesity and type 2 Diabetes.pdf"
+  ],
+  "extraction_id": [
+    "e6158348-e782-5e6d-9d89-3169b8fa630f",
+    "b534ab93-c837-5d89-809d-92062b1d49a4",
+    "35936d60-f8db-502e-be2c-4fe39f60fddd",
+    "93638ea5-6d1f-5b6a-9629-798804de24dd",
+    "6283c124-b479-5050-86ca-dc42390147a1",
+    "12668f1a-1631-5cce-bb6a-80b4de3fbb9e",
+    "ef8e6aa1-b7e0-5988-b9fb-a339317f9a66",
+    "de689016-3a4c-53b2-b3bf-a25ccbcbbb02",
+    "65609b08-1113-5a7f-9117-73476bcf50de",
+    "93638ea5-6d1f-5b6a-9629-798804de24dd"
+  ],
+  "document_id": [
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "8bc8f3d4-968f-5252-ab4c-832b92e9ec0d",
+    "ab9288ab-e3ad-58f1-b5ba-183ee17ce4bd",
+    "a16d3328-039c-530a-bfe5-f6f80ecf2ad0",
+    "e4d4a19e-18a0-5a08-9ab7-537f31b7cdc1",
+    "47186d35-9c05-5b0a-b8cd-21d2e0e688d8",
+    "72ab8458-928b-56b9-9547-1ba4b59dfab9",
+    "0f49b102-1d7e-5702-af30-35e5f2ed93a6",
+    "6f5ced46-b777-563a-b644-432f4e7e2644",
+    "a16d3328-039c-530a-bfe5-f6f80ecf2ad0"
+  ],
+  "id": [
+    "chatcmpl-ADZRbty26XP7vi2KOPG4Rh8fHX6iY",
+    "4e079c08-9095-5ec2-8c19-c6d0b222891e",
+    "b81dd6ab-e06e-55a2-bc0a-c89c5e883d3a",
+    "d090cda1-cf6a-5f2f-899a-3c7c763d0c8c",
+    "2b361786-7027-54e1-825d-34abc3a3fe98",
+    "89339b65-325f-588f-9f25-761124f0012f",
+    "74ec2f7f-a933-53b3-a78a-c69b9796c1c5",
+    "e6e5b010-d608-5a19-ae74-d571499fbb7b",
+    "2dc0e0fa-b061-5c09-8af3-02a44811042e",
+    "0e465787-e5b0-5f33-88cf-9bd1d0624f68",
+    "4d08d1ea-03a2-53d9-bb9d-df46c3fc2dcb"
+  ],
+  "contexts": [
+    "of Type 2 Diabetes   The lifestyle intervention using physical exercise  and modi  cation of nutrition is ef  cient in pre- venting type 2 diabetes in patients with impaired glucose tolerance [  99 ]. Clinical trials con  rm  that lifestyle interventions (dietary modi  cation  and increased physical activity) reduce the risk of progressing from impaired glucose tolerance to type 2 diabetes [  105 ]. Assessing T2D risk accord- ing to FINDRISK scale [ 106 ] is quite common in",
+    "Major clinical trials have demonstrated that diet and lifestyle modifications are  effective in preventing T2DM in high-risk individuals.  T2DM management strategies including lifestyle modifications, social support and  ensuring medication adherence are key to reducing the incidence of diabetes  mellitus complications. REVIEWS NATURE REVIEWS | ENDOCRINOLOGY  VOLUME 14 | FEBRUARY 2018 | 89",
+    "focused on people with impaired glucose tolerance or impaired fasting glucose because of their high risk of development of type 2 diabetes. Several studies have examined the ability of lifestyle modi  cation and drugs to slow progression to diabetes (table 2). Findings from these trials have nearly all shown a bene  t, with lifestyle modi  cations being more e   cacious than any drug, with  the exception of the thiazolidinedione anti diabetics. 163175",
+    "no or just minor weight loss was achieved, diabetes incidence was also reduced (   Pan et al., 1997   ;    Ramachandran et al., 2006   ). In addition, on the long term weight was partially or totally regained in all of the studies (   Knowler et al., 2009   ;    Li et al., 2008   ;    Lindstrom et al., 2006   ;    Lindstrom et al., 2003   ). Despite this regain T2DM risk remained low or decreased further, thus the e  ect of lifestyle is unlikely to be solely due to",
+    "proven particularly effective for preven-tion and management of type 2 diabetes.For example, improvement in dietaryquality, in conjunction with other lifestylemodications like increased physical ac-tivity, was shown to be more effectivethan pharmacological treatment in pre-vention of diabetes in individuals at highrisk (1). Further, lifestyle modicationmay mitigate the risk associated with thestrongest known diabetes risk loci (2).While the existence of environmental in-uences on genetic risk (and vice",
+    "spite of our incomplete knowledge of the genetics of type 2diabetes today, the burden of type 2 diabetes can be amelio-rated at the population level. Recent studies have found thatlifestyle changes through diet and exercise can prevent or",
+    "Lifestyle modification including exercise, nutrition and behavioral changes is the cornerstone to prevent and treat type 2 diabetes. Oral antidiabetic medication either as single agent or combination therapy is frequently required to maintain metabolic control, as assessed by monitoring ofglycated hemoglobin A 1C(HbA 1C) levels. Eventually, asignificant proportion of patients with type 2 diabetes require the exogenous administration of insulin [40].",
+    "diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 2001; 344: 134350. 114 Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in  the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002; 346: 393403. 115 Ramachandran A, Snehalatha C, Mary S, Mukesh B, Bhaskar AD,",
+    "type 2 diabetes. Physical activity, favorable dietary changes,and weight reduction were essential components of a success-ful lifestyle intervention in two large randomized controlled trials on the prevention of type 2 diabetes in high-risk individ-uals with impaired glucose tolerance (IGT), including theFinnish Diabetes Prevention Study (DPS) (44) and the Diabe-tes Prevention Program (DPP) (22). In the DPS, increasedphysical activity was associated with a decreased risk of type",
+    "demonstrate that lifestyle modi  cation comprising higher levels  of PA and prudent food consumption may be e  ective in obesity  and T2DM prevention. The positive e  ect of lifestyle on body  weight seems somewhat transient, whereas the e  ect on T2DM  is sustained for longer periods. Furthermore, lifestyle modi  ca- tion appears to have an e  ect on diabetes risk independently of  body weight and even of weight loss.      Lifestyle and Genetics in Obesity and Type 2 Diabetes"
+  ],
+  "answer": "Lifestyle changes, particularly those involving diet and physical activity, have been shown to be effective in managing and even reversing type 2 diabetes (T2DM). Several sources in the provided context support this:  1. **Prevention and Management**: Lifestyle interventions, including dietary modifications and increased physical activity, have been proven effective in preventing the progression from impaired glucose tolerance to type 2 diabetes [1], [2], [3], [5], [9]. These interventions are more efficacious than pharmacological treatments in some cases [5].  2. **Sustained Effects**: The positive effects of lifestyle changes on type 2 diabetes risk are sustained over longer periods, even if weight is partially or totally regained [4], [10]. This suggests that the benefits of lifestyle modifications are not solely dependent on weight loss.  3. **Cornerstone of Treatment**: Lifestyle modification, including exercise, nutrition, and behavioral changes, is considered the cornerstone for both preventing and treating type 2 diabetes [7].  In summary, lifestyle changes can indeed play a significant role in reversing type 2 diabetes, as evidenced by multiple clinical trials and studies [1], [2], [3], [4], [5], [7], [9], [10].",
+  "question": "Can lifestyle changes reverse type 2 diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_2 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_2
new file mode 100644
index 0000000..20550b1
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_2
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2008 - Glossary of Genetics Genomics Terms.pdf",
+    "2013 - Genetic association of ADIPOQ gene variants with type 2 diabetes, obesity.pdf",
+    "2015 - The Association of Type 2 Diabetes Loci.pdf",
+    "2016 - Genome-Wide Association Studies of Type 2 Diabetes.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2019 - Genetic Risk Scores for Diabetes Diagnosis.pdf",
+    "2012 - Gene-Environment Interactions in the Development of Type 2 Diabetes.pdf",
+    "2018 - Quantitative Relationship Between Cumulative Risk Alleles Based.pdf",
+    "2015 - Diabetes mellitus The epidemic of the century.pdf",
+    "2012 - The Pathogenesis and Natural History of Type 1 diabetes.pdf"
+  ],
+  "extraction_id": [
+    "53e868dd-b318-5cf3-8b2e-98a548aab7cf",
+    "c2875fb7-31e1-51f2-87b8-f2c21d597d08",
+    "8703f848-f3bc-58b2-932a-a49b1f0fb002",
+    "c92c81bb-ede1-5e01-af7d-e244214fc856",
+    "eb3de845-98db-505c-bb7f-c0f3259875fc",
+    "a8162fba-c5da-504f-a018-b6242a026bc5",
+    "b961664b-5008-547c-a302-ee8c719f68fd",
+    "6db9f25e-36fd-51c0-be36-6dfacd963b1b",
+    "b1c7a0c1-90a0-54fe-a662-9113e44e2c9f",
+    "b797dd19-b8f4-5dc9-93ee-ace7594bf3bf"
+  ],
+  "document_id": [
+    "c66d2572-071d-5aaf-829c-b3ca6cf6d697",
+    "6a2afe9a-51c0-52a6-be40-c034fb45c69a",
+    "a2abccec-e5cb-56ae-93b9-3040bc09f148",
+    "185aad8a-6a5b-5b18-81c4-ef251edef5e7",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "8c66aca1-d4ba-534d-a037-4273de340ee1",
+    "ea9601ed-ad83-506e-b1b7-e7211671ff73",
+    "d585896e-1c32-51cb-827d-e4fd3b3943f3",
+    "e114dd28-fd39-56df-bdeb-8806474a6c10",
+    "acad2a9b-1149-539b-b335-661176d631f1"
+  ],
+  "id": [
+    "chatcmpl-ADZQIhRURTB7PnDm4Bf2cVOJhSbs0",
+    "54ff4672-bf7f-5158-b228-ca3d45e0cb0d",
+    "0ccd2114-85e2-5aa3-85b5-3ae4b202037a",
+    "bf4247f8-f82c-5c40-b5af-3a68476f54bf",
+    "4b289db2-bda2-51d1-8f65-1cda62a4e40f",
+    "9fc663d2-2833-51e7-ae6a-55b007a6e27c",
+    "a67fe95c-11ac-5d06-8757-209f9abd0fd8",
+    "14608f3c-f5fa-52d6-b2c7-6ce6fd40985f",
+    "32b978f9-4bce-5f39-a655-09685b0d0f1f",
+    "74ab0f97-7758-5b01-b178-afee23d2e6cc",
+    "10d30a80-821a-5d09-988b-60bc71eae43c"
+  ],
+  "contexts": [
+    "Genetic factors appear to play a role in determining an individuals risk of developing diabetes. It is hoped",
+    "ger, will develop diabetes because the prevalence of diabetes increases with age. In order to circumvent this problem, age was adjusted for in2 K. Ramya et al. / Gene xxx (2013) xxx xxx Please cite this article as: Ramya, K., et al., Genetic association of ADIPOQ gene variants with type 2 diabetes, obesity and serum adiponectin levels in south Indian population, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.09.012",
+    "elderly population. PLoS One 9: e100548. doi: 10.1371/journal.pone.0100548 PMID: 24959828 23. Strawbridge RJ, Dupuis J, Prokopenko I, Barker A, Ahlqvist E, Rybin D, et al. (2011) Genome-wide association identifies nine common variants associated with fasting proinsulin levels and provides new insights into the pathophysiology of type 2 diabetes. Diabetes 60: 2624 2634. doi: 10.2337/db11-0415 PMID: 21873549",
+    "information for diabetes risk prediction - differences according to sex, age, family history and obesity. PloS One 8(5):e64307. doi: 10.1371/journal.pone.0064307 Neel JV (1962) Diabetes mellitus: a thrifty genotype rendered detrimental by progress? Am J Hum Genet 14:353362 Neel JV (1999) The thrifty genotype in 1998. Nutr Rev 57(5 Pt 2):S2S9 Palmer ND, McDonough CW, Hicks PJ, Roh BH, Wing MR, An SS, Hester JM, Cooke JN,",
+    "insulin resistance, hypertension, and dyslipidemia (Obesity Education Initiative Expert Panel,  1998 ). Insulin resist-ance increases with age, and the incidence of diabetes rises sharply in the elderly (American Diabetes Association,  2010a ).   In a few patients, genetic mutations appear to be associ- ated with T2D (Roche  et al.  ,  2005 ; American Diabetes  Association,  2010a ). For example, recent work using the DPP data has led to the identi  cation of 27 single nucle-",
+    "early-onset diabetes in some pedigrees, but it also maybe observed in individuals who retain normal glucose tolerance into late adulthood and beyond ( ). Studying  individuals from  HNF A-MODY families, Lango Allen et al. () found that a -SNP T Dr s P S was signi cantly associated with earlier age of diabetes diagnosis, with each additional risk allele accelerating diagnosis by ~ months. Clinical application of predictive scores",
+    "12. de Miguel-Yanes JM, Shrader P, Pencina MJ, Fox CS, Manning AK, et al. 2011. Genetic risk reclassi- cation for type 2 diabetes by age below or above 50 years using 40 type 2 diabetes risk single nucleotide polymorphisms. Diabetes Care 34:12125 13. Dempe A, Scherag A, Hein R, Beckmann L, Chang-Claude J, Schafer H. 2008. Gene-environment interactions for complex traits: denitions, methodological requirements and challenges. Eur. J. Hum. Genet. 16:116472",
+    "diabetes risk genes predicts impaired glucose tolerance in female andobese individuals. PLoS One . 2012;7:e38224 . 74. Stevens JW, Khunti K, Harvey R, et al. Preventing the progression to type 2 diabetes mellitus in adults at high risk: a systematic review and network meta-analysis of lifestyle, pharmacological and surgicalinterventions. Diabetes Res Clin Pract . 2015;107:320 331(in eng).Cumulative Risk Alleles and Type 2 Diabetes Mellitus 18jJ Epidemiol 2018;28(1):3-18",
+    "and protects against oxidative stress-induced insulin-deficient  diabetes. PLoS One  2014; 9: e87941 [PMID: 24498408 DOI:  10.1371/journal.pone.0087941] 23 Maahs DM , West NA, Lawrence JM, Mayer-Davis EJ. Epidemiology  of type 1 diabetes. Endocrinol Metab Clin North Am  2010; 39:  481-497 [PMID: 20723815 DOI: 10.1016/j.ecl.2010.05.011] 24 Daneman D . Type 1 diabetes. Lancet  2006; 367: 847-858 [PMID:  16530579 DOI: 10.1016/S0140-6736(06)68341-4]",
+    "Sosenko JM, Skyler JS, Krischer JP , Greenbaum CJ, Mahon J, Rafkin LE, Cuthbertson D, Cowie C, Herold K, Eisen-barth G, et al. 2010. Glucose excursions between states of glycemia with progression to type 1 diabetes in the diabetes prevention trial-type 1 (DPT-1). Diabetes 59: 23862389. Steck AK, Armstrong TK, Babu SR, Eisenbarth GS. 2011. Type 1 Diabetes Genetics Consortium. Stepwise or linear decrease in penetrance of type 1 diabetes with lower-risk HLA genotypes over the past 40 years. Diabetes 60:"
+  ],
+  "answer": "Protective genetic factors against diabetes in elderly populations can be understood through several key insights from the provided context:  1. **Genome-Wide Association Studies (GWAS)**: Research has identified common genetic variants associated with fasting proinsulin levels, which provide new insights into the pathophysiology of type 2 diabetes. These findings suggest that certain genetic variants may play a protective role by influencing insulin regulation and glucose metabolism in the elderly [3].  2. **Single Nucleotide Polymorphisms (SNPs)**: Studies have identified specific SNPs that are associated with diabetes risk. For example, some genetic mutations are linked to early-onset diabetes, but individuals with certain protective alleles may retain normal glucose tolerance into late adulthood. This indicates that specific genetic configurations can delay or prevent the onset of diabetes in elderly populations [6].  3. **Gene-Environment Interactions**: The interaction between genetic factors and environmental influences is crucial in understanding diabetes risk. Protective genetic factors may interact with lifestyle and environmental factors to reduce the likelihood of developing diabetes in older adults [7].  4. **Oxidative Stress Protection**: Certain genetic factors may protect against oxidative stress-induced insulin-deficient diabetes. This suggests that genetic mechanisms that enhance the body's ability to manage oxidative stress could be protective against diabetes in the elderly [9].  In summary, protective genetic factors against diabetes in elderly populations include specific genetic variants that influence insulin regulation, SNPs that delay diabetes onset, gene-environment interactions, and genetic mechanisms that protect against oxidative stress. These factors collectively contribute to a reduced risk of developing diabetes in older adults.",
+  "question": "Explain Protective Genetic Factors Against Diabetes in Elderly Populations"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_3 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_3
new file mode 100644
index 0000000..5caf9a4
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_3
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2011 - Lifestyle and Genetics in Obesity and type 2 Diabetes.pdf",
+    "2011 - Lifestyle and Genetics in Obesity and type 2 Diabetes.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "2016 - Epigenetics and aging.pdf",
+    "2005 - Metabolic Syndrome A Comprehensive Perspective Based  on Interactions Between Obesity Diabetes and Inflammation.pdf",
+    "2011 - Lifestyle and Genetics in Obesity and type 2 Diabetes.pdf",
+    "2021 - Gene-by-environment modulation of lifespan and weight gain in the murine BXD family.pdf",
+    "2020 - Precision Medicine in Diabetes.pdf",
+    "2011 - Type 2 diabetes across generations from pathophysiology to prevention and management.pdf"
+  ],
+  "extraction_id": [
+    "93638ea5-6d1f-5b6a-9629-798804de24dd",
+    "93638ea5-6d1f-5b6a-9629-798804de24dd",
+    "3bf4c712-4a5a-5a67-9e2a-d83fba8c1cb4",
+    "bc31e1f8-f149-50c4-82c1-86e2d465202c",
+    "4fb7ef96-fe5a-5d81-bf28-c756656f1cbb",
+    "c6cfb382-639a-5dd4-a9c8-c8f57b6daabc",
+    "551087b1-8e80-5a7b-839a-304f566a6417",
+    "bca61863-81b3-5ef7-850d-10cc9577a9e1",
+    "68183d3e-4c95-5363-92b8-891dccf7e3d6",
+    "de689016-3a4c-53b2-b3bf-a25ccbcbbb02"
+  ],
+  "document_id": [
+    "a16d3328-039c-530a-bfe5-f6f80ecf2ad0",
+    "a16d3328-039c-530a-bfe5-f6f80ecf2ad0",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "71b206ec-81bd-5194-8b21-ae522f8cbc2d",
+    "de2aa54c-eb0f-5dc3-ac92-23ee3215dd2a",
+    "a16d3328-039c-530a-bfe5-f6f80ecf2ad0",
+    "4d082da4-fa48-5170-8147-c4fea47a5d4b",
+    "0ad5b2de-d782-5d43-b294-bff5c7befd2d",
+    "0f49b102-1d7e-5702-af30-35e5f2ed93a6"
+  ],
+  "id": [
+    "chatcmpl-ADZQPOsxOK9DJcrr7qBEh29WBnCmr",
+    "4d08d1ea-03a2-53d9-bb9d-df46c3fc2dcb",
+    "be87703d-e7b2-5db5-9983-5412e09a57ba",
+    "5c99d3b9-8b1a-5be4-8689-97662557dac4",
+    "4c5eb67d-3bdd-58d7-bf5e-d1d08a47118d",
+    "3fd5d259-8fd4-5b0d-bb64-134424baeef2",
+    "8c4e8b2c-6730-541c-8a2e-22fbd7ddb487",
+    "6f12fbd4-284d-5d41-9d60-54aa268a635d",
+    "06c32067-10ea-599a-9af2-9413ad8c8984",
+    "57012499-8167-5e51-8cb5-b436460e24a2",
+    "2dc0e0fa-b061-5c09-8af3-02a44811042e"
+  ],
+  "contexts": [
+    "demonstrate that lifestyle modi  cation comprising higher levels  of PA and prudent food consumption may be e  ective in obesity  and T2DM prevention. The positive e  ect of lifestyle on body  weight seems somewhat transient, whereas the e  ect on T2DM  is sustained for longer periods. Furthermore, lifestyle modi  ca- tion appears to have an e  ect on diabetes risk independently of  body weight and even of weight loss.      Lifestyle and Genetics in Obesity and Type 2 Diabetes",
+    "suggested to attenuate its negative e  ect on metabolic pro  le,  body weight, and diabetes risk (   Franks et al., 2007   ;    Kilpelainen et al., 2008   ;    Lindi et al., 2002   ;    Ruchat et al., 2010   ) (             Table 1   ).  The notion that lifestyle modi  cation can eliminate the increased  risk for development of T2DM in subjects with genetic suscepti-bility is also supported by  ndings of    Barwell et al. (2008)    who",
+    "M., Bray, G. A. et al (2006). Effect of weight loss withlifestyle intervention on risk of diabetes. Diabetes Care, 29 , 21022107. Herder, C., Peltonen, M., Koenig, W., Sutfels, K., Lindstrom, J. et al (2009). Anti-inammatory effect oflifestyle changes in the Finnish Diabetes PreventionStudy. Diabetologia, 52 , 433442. Hung, J., McQuillan, B. M., Thompson, P . L., and Beilby,",
+    "22        Medications for Diabetes  Prevention   Even in the most successful of the randomized  controlled trials, the risk reduction for incident diabetes following lifestyle intervention was ~60 % [  48  51 ]. That raises the argument as to",
+    "SRT2104 extend the life span of obese mice and protect against age- related changes in multiple tissues ( 215). The antidiabetic drug metformin also induces effects similar to CR (216). Diabetes is considered an age-associated disease, and disturbances in insulin signaling and carbohydrate homeostasis may essentially lead toother age-related complications, including cancer, if untreated. Along with its antidiabetic properties, metformin supplementation has been",
+    "74 The mechanism underlying this effect of exercise is not known;however, it is noteworthy that lifestyle change is a very effectiveway to reduce the rate of development of diabetes in a predia-betic population, as shown by the diabetes prevention study. 75,76 Both a reduction in macronutrient intake and exercise cause areduction in inflammation. References 1. Reaven GM. Banting lecture 1988. Role of insulin resistance in human disease. Diabetes . 1988;37:15951607.",
+    "uals, but also for low-risk lean individuals (   Kriska et al., 2003   ;    Meisinger et al., 2005   ;    Schulze et al., 2006   ). Furthermore, health-ier lifestyle has been shown to be associated with decreased incidence of obesity- and T2DM-related complications such as hypertension and cardiovascular disease (   Manson et al., 2002   ;    Stampfer et al., 2000   ).     Evidence from randomized controlled trails   The e  cacy of lifestyle changes in obesity and T2DM prevention",
+    "extends lifespan. Cell Rep. 20, 451463 (2017). [PubMed: 28700945]  64. Barzilai N & Ferrucci L Insulin resistance and aging: A cause or a protective response? J.  Gerontol. Ser. A 67, 13291331 (2012). 65. Holmes MV , Ala-Korpela M & Smith GD Mendelian randomization in cardiometabolic disease:  challenges in evaluating causality. Nat. Rev. Cardiol. 14, 577590 (2017). [PubMed: 28569269]  66. Holmes MVet al.Mendelian randomization of blood lipids for coronary heart disease. Eur. Heart J.",
+    "70. Knowler WC, Barrett-Connor E, Fowler SE,et al.; Diabetes Prevention Program ResearchGroup. Reduction in the incidence of type 2diabetes with lifestyle intervention or metfor-min. N Engl J Med 2002;346:393 403 71. Crandall J, Schade D, Ma Y, et al.; DiabetesPrevention Program Research Group. The in-uence of age on the effects of lifestyle mod-",
+    "diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 2001; 344: 134350. 114 Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in  the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002; 346: 393403. 115 Ramachandran A, Snehalatha C, Mary S, Mukesh B, Bhaskar AD,"
+  ],
+  "answer": "Lifestyle modifications, including increased physical activity (PA) and prudent food consumption, have been shown to be effective in preventing obesity and type 2 diabetes mellitus (T2DM) [1]. These modifications can have a sustained positive effect on diabetes risk, even independently of body weight and weight loss [1]. This suggests that the benefits of lifestyle changes extend beyond mere weight management.  Furthermore, lifestyle modifications can attenuate the negative effects on metabolic profiles, body weight, and diabetes risk, even in individuals with genetic susceptibility to T2DM [2]. This indicates that lifestyle changes can potentially eliminate the increased risk of developing T2DM in genetically predisposed individuals.  The effectiveness of lifestyle changes in reducing the rate of diabetes development has been demonstrated in prediabetic populations, where both reduced macronutrient intake and increased exercise have been shown to lower inflammation, a key factor in diabetes progression [6]. Additionally, randomized controlled trials have shown that lifestyle interventions can reduce the incidence of T2DM by approximately 60% [4].  In summary, lifestyle modifications play a crucial role in reducing the risk of aging-associated diabetes by improving metabolic health, reducing inflammation, and mitigating genetic risks, thereby offering a comprehensive approach to diabetes prevention [1], [2], [4], [6].",
+  "question": "Explain Effect of Lifestyle Modifications on Aging-Associated Diabetes Risk"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_4 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_4
new file mode 100644
index 0000000..fc0bdbd
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_4
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2015 - Pharmacogenetics and individual responses to treatment of hyperglycemia.pdf",
+    "2010 - Genome-wide association study (GWAS)-identified disease risk alleles do not compromisehuman longevity.pdf",
+    "2011 - Genomics of human longevity.pdf",
+    "2021 - Gene-by-environment modulation of lifespan and weight gain in the murine BXD family.pdf",
+    "2016 - Whole-Genome Sequencing of a Healthy Aging Cohort.pdf",
+    "2017 - Four Genome-Wide Association Studies Identify New.pdf",
+    "2008 - Glossary of Genetics Genomics Terms.pdf",
+    "2019 - Bioinformatic prediction of critical genes and pathways.pdf",
+    "2011 - Genomics of human longevity.pdf",
+    "2019 - Genetic Risk Scores for Diabetes Diagnosis.pdf"
+  ],
+  "extraction_id": [
+    "32275a81-cd67-525e-b6c1-c68dc441ab62",
+    "680423ed-71cc-5049-a80f-c78fe86e35ff",
+    "7c183ae5-f10e-5f0c-962e-32135887b3bd",
+    "bca61863-81b3-5ef7-850d-10cc9577a9e1",
+    "c55b4a12-6cc8-5594-87d4-53e4f8f023d1",
+    "a6075268-c86f-536b-a6b4-d2e18be9f117",
+    "53e868dd-b318-5cf3-8b2e-98a548aab7cf",
+    "4109e561-4721-5f4e-b4d5-4353f8d1741d",
+    "7c183ae5-f10e-5f0c-962e-32135887b3bd",
+    "a8162fba-c5da-504f-a018-b6242a026bc5"
+  ],
+  "document_id": [
+    "46081466-a50f-59d8-893d-8b8883b38507",
+    "200c2966-b647-552f-8504-0d6fb7f50bfa",
+    "2e038219-fdaa-506f-9cd3-51379054130e",
+    "4d082da4-fa48-5170-8147-c4fea47a5d4b",
+    "3a287979-e475-545b-99e6-4c1925653a79",
+    "c10653f6-b3d7-5b92-9271-ab8fcc7905a7",
+    "c66d2572-071d-5aaf-829c-b3ca6cf6d697",
+    "01201944-11f2-52d9-ac3e-7af685d4a4c4",
+    "2e038219-fdaa-506f-9cd3-51379054130e",
+    "8c66aca1-d4ba-534d-a037-4273de340ee1"
+  ],
+  "id": [
+    "chatcmpl-ADZQVK9rNW7qGGShVvwBLR6uFNp9v",
+    "849d5eca-38a4-553e-83da-a967ba81614c",
+    "260a4030-b151-5afd-ae06-86246ee73a7a",
+    "558acee9-89ff-599a-8502-bc181bc94995",
+    "06c32067-10ea-599a-9af2-9413ad8c8984",
+    "19faf41b-7716-5244-a9c3-196c2e5cd477",
+    "369b0a64-a439-573a-99dd-67d911026c37",
+    "54ff4672-bf7f-5158-b228-ca3d45e0cb0d",
+    "a45fa299-f675-5050-a510-dfa6d0954a25",
+    "cfe4eab8-fb34-5d0b-ae67-79c3d9993e15",
+    "a67fe95c-11ac-5d06-8757-209f9abd0fd8"
+  ],
+  "contexts": [
+    "Longitudinal Study of Aging. The natural history of progression from normalglucose tolerance to type 2 diabetes in the Baltimore Longitudinal Study of Aging. Diabetes 2003; 52:1475 1484. 22 Hornbak M, Allin KH, Jensen ML, Lau CJ, Witte D, Jrgensen ME ,e ta l .A combined analysis of 48 type 2 diabetes genetic risk variants shows nodiscriminative value to predict time to first prescription of a glucose lowering drug in Danish patients with screen detected type 2 diabetes. PLoS One 2014; 9:e104837.",
+    "A set of currently known alleles increasing the risk for coronary artery disease, cancer, and type 2 diabetes as identi ed by genome- wide association studies was tested for compatibility with human longevity. Here, we show that nonagenarian siblings from long- lived families and singletons older than 85 y of age from the general population carry the same number of disease risk alleles as young controls. Longevity in this study population is not compromised by",
+    "52561.x ) 17 Atzmon, G., Schechter, C., Greiner, W ., Davidson, D., Rennert, G. & Barzilai, N. 2004 Clinical phenotype of families with longevity. J. Am. Geriatr. Soc. 52, 274 277. ( doi:10.1111/j.1532-5415.2004.52068.x ) 18 Rozing, M. P . et al. 2009 Human insulin/IGF-1 and familial longevity at middle age. Aging (Albany NY )1, 714722. 19 Rozing, M. P . et al. 2010 Favorable glucose tolerance and lower prevalence of metabolic syndrome in",
+    "extends lifespan. Cell Rep. 20, 451463 (2017). [PubMed: 28700945]  64. Barzilai N & Ferrucci L Insulin resistance and aging: A cause or a protective response? J.  Gerontol. Ser. A 67, 13291331 (2012). 65. Holmes MV , Ala-Korpela M & Smith GD Mendelian randomization in cardiometabolic disease:  challenges in evaluating causality. Nat. Rev. Cardiol. 14, 577590 (2017). [PubMed: 28569269]  66. Holmes MVet al.Mendelian randomization of blood lipids for coronary heart disease. Eur. Heart J.",
+    "et al., 2012 ), possibly due to the indirect and/or a mixed relation- ship between individual genetic disease risk loci and exceptional longevity (as discussed by Fortney et al., 2015 ) versus the poten- tially more direct relationship between aging in the absence of disease and overall genetic disease risk. On the other hand, no difference in genetic risk is observed for type 2 diabetes genetic risk and cancer. Some of these ndings (type 2 diabetes, colon, and lung cancer) can be explained by the",
+    "5. Garagnani P, Giuliani C, Pirazzini C, etal. Centenarians as super-controls to assess the biological relevance of genetic risk factors for common age-related diseases: a proof of principle on type 2 diabetes. Aging (Albany NY). 2013;5:373385. doi:10.18632/aging.100562  6. Sebastiani P, Nussbaum L, Andersen SL, Black MJ, Perls TT. Increasing  sibling relative risk of survival to older and older ages and the importance",
+    "Genetic factors appear to play a role in determining an individuals risk of developing diabetes. It is hoped",
+    "The pursuit of longevity has been the goal of humanity since ancient times. Genetic alterations have been demonstrated to affect lifespan. As increasing numbers of pro-longevity genes and anti-longevity genes have been discovered in Drosophila, screening for functionally important genes among the large number of genes has become difficult. The aim of the present study was to explore critical genes and pathways affecting longevity in Drosophila melanogaster. In this study, 168 genes associated with",
+    "offspring without diabetes mellitus of nonagenariansiblings: the Leiden Longevity Study. J. Am. Geriatr. Soc. 58, 564569. ( doi:10.1111/j.1532-5415.2010. 02725.x ) 20 Suh, Y . et al. 2008 Functionally signicant insulin-like growth factor I receptor mutations in centenarians.Proc. Natl Acad. Sci. USA 105, 34383442. ( doi:10. 1073/pnas.0705467105 ) 21 Heijmans, B. T ., Beekman, M., Houwing-Duistermaat, J. J., Cobain, M. R., Powell, J., Blauw, G. J., van der",
+    "early-onset diabetes in some pedigrees, but it also maybe observed in individuals who retain normal glucose tolerance into late adulthood and beyond ( ). Studying  individuals from  HNF A-MODY families, Lango Allen et al. () found that a -SNP T Dr s P S was signi cantly associated with earlier age of diabetes diagnosis, with each additional risk allele accelerating diagnosis by ~ months. Clinical application of predictive scores"
+  ],
+  "answer": "Longevity genes play a significant role in protecting against diabetes by influencing various biological pathways that contribute to better glucose tolerance and metabolic health. Here are some key points explaining their role:  1. **Favorable Glucose Tolerance and Lower Prevalence of Metabolic Syndrome**: Research has shown that individuals from long-lived families tend to have favorable glucose tolerance and a lower prevalence of metabolic syndrome, which are critical factors in reducing the risk of type 2 diabetes [3].  2. **Insulin/IGF-1 Pathway**: The insulin/IGF-1 signaling pathway is crucial in regulating lifespan and metabolic processes. Mutations in the insulin-like growth factor I receptor (IGF-1R) have been found in centenarians, suggesting that these mutations may contribute to longevity by enhancing insulin sensitivity and glucose metabolism, thereby protecting against diabetes [9].  3. **Genetic Risk and Longevity**: Studies have indicated that nonagenarians and centenarians carry the same number of disease risk alleles for type 2 diabetes as younger controls, yet they do not exhibit the disease. This suggests that longevity genes may counteract the effects of diabetes risk alleles, possibly through mechanisms that enhance metabolic health and insulin sensitivity [2], [5].  4. **Use of Centenarians as Super-Controls**: Centenarians are often used as super-controls in genetic studies to assess the biological relevance of genetic risk factors for common age-related diseases, including type 2 diabetes. This approach helps identify genetic factors that contribute to both longevity and protection against diabetes [6].  In summary, longevity genes contribute to protecting against diabetes by promoting favorable glucose tolerance, enhancing insulin sensitivity, and counteracting the effects of diabetes risk alleles through various genetic and metabolic pathways [2], [3], [5], [6], [9].",
+  "question": "Explain The Role of Longevity Genes in Protecting Against Diabetes"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_5 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_5
new file mode 100644
index 0000000..024af91
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_5
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "2010 - Epidemiology of diabetes.pdf",
+    "2015 - Diabetes mellitus The epidemic of the century.pdf",
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "2016 - Dissecting diabetes metabolic disease.pdf",
+    "2011 - Interaction Between Exercise and Genetics.pdf",
+    "2015 - Recent Trends in Therapeutic Approaches for Diabetes Management  A Comprehensive Updat.pdf",
+    "2015 - Recent Trends in Therapeutic Approaches for Diabetes Management  A Comprehensive Updat.pdf",
+    "2017 - Machine Learning and Data Mining Methods in Diabetes Research.pdf",
+    "2023 - Childhood adiposity and novel subtypes of adult-onset diabetes a Mendelian randomisation and genome-wide genetic correlation study.pdf"
+  ],
+  "extraction_id": [
+    "d5004507-475d-5ac1-9aa6-b5fd664b8bf7",
+    "4307e79a-c3ae-51d7-8510-820375d2c4ca",
+    "6a734fb4-5ce1-5f11-b1fb-288e38ef9a6c",
+    "b2cd4ee5-81b3-5701-8cd1-8dbea4242cc1",
+    "998a92ba-e7fc-5553-b629-7b5797fbfafe",
+    "ed6dcfee-8273-5512-8fb4-fc51a9c921da",
+    "b8e47ab6-95e0-5fbb-bc40-fa9e46c0b1dc",
+    "e4d87eba-dfd4-51e5-a560-1ad46924edf1",
+    "81a02908-ff22-5136-be83-d53e04a81541",
+    "f0e064be-81a0-5ee9-88da-2a7049c65520"
+  ],
+  "document_id": [
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "7f1cb121-3a35-571e-81c9-96a3afd66448",
+    "e114dd28-fd39-56df-bdeb-8806474a6c10",
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "eee2f79d-e093-52fb-871a-798fd859235e",
+    "c36db75e-4b76-540d-9efb-d0e156e61541",
+    "ec4921c2-af14-56cc-aed3-65f8ea236bde",
+    "ec4921c2-af14-56cc-aed3-65f8ea236bde",
+    "e2dcbb80-5ad7-5441-b170-9b46607445b0",
+    "fff2bd78-2ac2-5672-b8fd-ed82ab7c910b"
+  ],
+  "id": [
+    "chatcmpl-ADZQdCy9515POnQgOjqu9IhwdWHwq",
+    "2454130e-8098-5c7f-944b-c5933a8409f8",
+    "6ba4950a-304f-5257-bd31-3e83a2f52df1",
+    "008aa60f-789b-519b-b81d-f437042c3df8",
+    "4660d51a-178a-5a14-a27a-2eeef1b0bf95",
+    "64fa332d-1415-584b-8b7c-43e8e3e698dc",
+    "3ef149b8-30fa-533b-b950-fc4122586080",
+    "ecc77a70-68dc-51a8-92a3-50f417deb98e",
+    "b169069b-43f2-5c24-8431-adfcaad27942",
+    "ae1db826-0202-53c9-a251-0fc9216bbf5c",
+    "ddc1154f-5406-5028-bacb-47a2ee6fbcf4"
+  ],
+  "contexts": [
+    "disorder caused by different factors characterized by a chronic high level of blood sugar with distur-bances to carbohydrate, fat, and protein metabo-lism resulting from defects in insulin secretion, insulin action, or both [  83 ]. Scientists have  divided diabetes into three different types: Type 1 F. Assah and J.C. Mbanya",
+    "Type 1 and type 2 diabetes are the two main types, with type 2 diabetesaccounting for the majority ( >85%) of total diabetes prevalence. Both",
+    "classical classification of diabetes as proposed by the  American Diabetes Association (ADA) in 1997 as type  1, type 2, other types, and gestational diabetes mellitus  (GDM) is still the most accepted classification and  adopted by ADA[1]. Wilkin[8] proposed the accelerator  hypothesis that argues type 1 and type 2 diabetes  are the same disorder of insulin resistance set against  different genetic backgrounds[9]. The difference bet - ween the two types relies on the tempo, the faster",
+    "41 diabetes mellitus (formerly insulin- dependent  diabetes mellitus  IDDM) or type 1 diabetes is also known as juvenile onset diabetes. Type 2 diabetes mellitus (non-insulin-dependent diabe-tes mellitus (formerly non-insulin- dependent dia-betes, NIDDM) or type 2 diabetes  adult-onset diabetes) is found in individuals who are insulin-resistant and who usually have relative insulin de ciency. Gestational diabetes mellitus (GDM),  the third type, is de  ned as any degree of glucose",
+    "Diabetes is a metabolic disease characterized by uncontrolled hyper-glycemia resulting from the variable combination of dysfunctional in-sulin secretion by pancreatic beta cells and insulin resistance. It is generally classi ed into monogenic diabetes (maturity onset diabetes of the young [MODY], neonatal diabetes, mitochondrial diabetes[54,55] , syndromes of insulin resistance) [56], type 1 diabetes (T1D) and type 2 diabetes (T2D). The metabolic syndrome is a combination of",
+    "Diabetes mellitus is a group of metabolic diseases characterized by hyperglycemia (elevated levels of glucose in the blood) resulting from defects in insulin secretion, insulin action, or both. There are two major types of diabetes mellitus: type 1 (T1D) and T2D, although several other rarer forms also exist [13]. T1D is an autoimmune disease that usually occurs in childhood, but the onset may occur at any age. T1D results from a cellular-mediated autoimmune destruction of the beta-cells in the pancreatic",
+    "2. Classification of Diabetes On the basis of insulin deficiency, diabetes can be classifiedintothefollowingtypesasfollows.2.1. Insulin Dependent Diabetes Mellitus (IDDM). It is also known as juvenile onset diabetes or type 1 diabetes, which accounts for 510% of the patients, resulting from cellular-mediated autoimmune destruction of the pancreatic cells. Thediseasecanaffectpeopleofallagesbutusuallyoccursin childrenoryoungadults.Regularsupplyofinsulininjections",
+    "2 JournalofDiabetesResearch Type I diabetes  IDDM Type II diabetes  NIDDM  Gestational  diabetesPancreas Islet of Langerhans-glucagon beta cells: insulin Genomic mutationsadministration for  survival sugar levels Insulin  resistance Defective insulin  production Increased  mortalityY ounger  populationGlobal  pandemicHuman body  and diabetes  pregnancy, it needs complete care and  glucose monitorin g glycemic status individual level identification/development of  lead moleculesRegular insulin Exercise",
+    "However, there are two major clinical types, type 1  diabetes (T1D) and type 2  diabetes (T2D), according to the etiopathology of t he disorder.  T2D appears to be the",
+    "SIDD Severe insulin-deficient diabetes SIRD Severe insulin-resistant diabetes Introduction In 2018, a ground-breaking study identified five novel subtypes of adult-onset diabetes: severe autoimmune diabetes (SAID, including type 1 diabetes and latent autoimmune diabetes in adults [LADA]) and four subtypes of type 2 diabetes (severe insulin-deficient diabetes [SIDD], severe insulin-resistant diabetes [SIRD], mild obesity-related diabetes [MOD] and mild age-"
+  ],
+  "answer": "Diabetes is classified into several types based on different criteria. The main types of diabetes are:  1. **Type 1 Diabetes (T1D)**: Also known as insulin-dependent diabetes mellitus (IDDM) or juvenile onset diabetes, it is an autoimmune disease that usually occurs in childhood but can onset at any age [4], [6], [7].  2. **Type 2 Diabetes (T2D)**: Also known as non-insulin-dependent diabetes mellitus (NIDDM) or adult-onset diabetes, it is characterized by insulin resistance and relative insulin deficiency [4], [6].  3. **Gestational Diabetes Mellitus (GDM)**: This type occurs during pregnancy and is defined as any degree of glucose intolerance with onset or first recognition during pregnancy [4].  4. **Other Types**: These include monogenic diabetes (such as maturity onset diabetes of the young [MODY], neonatal diabetes, mitochondrial diabetes), and syndromes of insulin resistance [5].  Additionally, a 2018 study identified five novel subtypes of adult-onset diabetes, which include severe autoimmune diabetes (SAID, including type 1 diabetes and latent autoimmune diabetes in adults [LADA]), severe insulin-deficient diabetes (SIDD), severe insulin-resistant diabetes (SIRD), mild obesity-related diabetes (MOD), and mild age-related diabetes [10].",
+  "question": "What are the types of diabetes"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_6 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_6
new file mode 100644
index 0000000..9527cf4
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_6
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2010 - Epidemiology of diabetes.pdf",
+    "2015 - Diabetes mellitus The epidemic of the century.pdf",
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "2023 - Childhood adiposity and novel subtypes of adult-onset diabetes a Mendelian randomisation and genome-wide genetic correlation study.pdf",
+    "2015 - Genetics, genomics and personalized medicine in Type 2 Diabetes.pdf",
+    "2017 - Machine Learning and Data Mining Methods in Diabetes Research.pdf",
+    "2018 - Novel subgroups of adult-onset diabetes and their association.pdf",
+    "2021 - Genomic Medicine in Diabetes Improving the Diagnostic Rate of Monogenic Diabetes.pdf",
+    "2007 - Bioethnic Conscription Genes, Race.pdf",
+    "2017 - Painting a new picture of personalised medicine for diabetes.pdf"
+  ],
+  "extraction_id": [
+    "4307e79a-c3ae-51d7-8510-820375d2c4ca",
+    "6a734fb4-5ce1-5f11-b1fb-288e38ef9a6c",
+    "b2cd4ee5-81b3-5701-8cd1-8dbea4242cc1",
+    "f0e064be-81a0-5ee9-88da-2a7049c65520",
+    "670074e5-275c-5999-9fb2-2370a1ce3dbf",
+    "81a02908-ff22-5136-be83-d53e04a81541",
+    "20a6e2db-c742-5f28-a310-62f3bf58d92a",
+    "499fe6d8-73ba-5835-91a7-af3376d1651b",
+    "d824748c-69ce-5124-8a76-99c3cf221f8a",
+    "2ee5d7fa-babf-5feb-b40a-fd453b4b3f31"
+  ],
+  "document_id": [
+    "7f1cb121-3a35-571e-81c9-96a3afd66448",
+    "e114dd28-fd39-56df-bdeb-8806474a6c10",
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "fff2bd78-2ac2-5672-b8fd-ed82ab7c910b",
+    "d8b85c3e-62f3-5e67-99b0-d0a2f225aff0",
+    "e2dcbb80-5ad7-5441-b170-9b46607445b0",
+    "c9a39a25-de31-5553-941b-bf1298cf1693",
+    "e315a891-ba59-57e9-856b-602544375324",
+    "d90126d9-fd87-5b38-87f7-08415f690836",
+    "e226b2b1-0bc4-5d79-b931-ad47f21be045"
+  ],
+  "id": [
+    "chatcmpl-ADZQhFOO3LRPtv9Lg1g6L8gDOic6T",
+    "6ba4950a-304f-5257-bd31-3e83a2f52df1",
+    "008aa60f-789b-519b-b81d-f437042c3df8",
+    "4660d51a-178a-5a14-a27a-2eeef1b0bf95",
+    "ddc1154f-5406-5028-bacb-47a2ee6fbcf4",
+    "945f57d6-b790-5c1b-a94b-c3076ab28adc",
+    "ae1db826-0202-53c9-a251-0fc9216bbf5c",
+    "191582b1-0a31-5791-b123-4e1fa2672962",
+    "ee7614a8-89a2-503a-9da2-4207c22225bc",
+    "13ab2950-2bdc-57d2-840a-042157d2b9e8",
+    "6a7f929c-ba32-51ea-93e1-2b760bcb156d"
+  ],
+  "contexts": [
+    "Type 1 and type 2 diabetes are the two main types, with type 2 diabetesaccounting for the majority ( >85%) of total diabetes prevalence. Both",
+    "classical classification of diabetes as proposed by the  American Diabetes Association (ADA) in 1997 as type  1, type 2, other types, and gestational diabetes mellitus  (GDM) is still the most accepted classification and  adopted by ADA[1]. Wilkin[8] proposed the accelerator  hypothesis that argues type 1 and type 2 diabetes  are the same disorder of insulin resistance set against  different genetic backgrounds[9]. The difference bet - ween the two types relies on the tempo, the faster",
+    "41 diabetes mellitus (formerly insulin- dependent  diabetes mellitus  IDDM) or type 1 diabetes is also known as juvenile onset diabetes. Type 2 diabetes mellitus (non-insulin-dependent diabe-tes mellitus (formerly non-insulin- dependent dia-betes, NIDDM) or type 2 diabetes  adult-onset diabetes) is found in individuals who are insulin-resistant and who usually have relative insulin de ciency. Gestational diabetes mellitus (GDM),  the third type, is de  ned as any degree of glucose",
+    "SIDD Severe insulin-deficient diabetes SIRD Severe insulin-resistant diabetes Introduction In 2018, a ground-breaking study identified five novel subtypes of adult-onset diabetes: severe autoimmune diabetes (SAID, including type 1 diabetes and latent autoimmune diabetes in adults [LADA]) and four subtypes of type 2 diabetes (severe insulin-deficient diabetes [SIDD], severe insulin-resistant diabetes [SIRD], mild obesity-related diabetes [MOD] and mild age-",
+    "7 American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care  37(Suppl. 1), S81S90  (2014). 8 Daneman D. Type 1 diabetes. Lancet  367(9513), 847858  (2006). 9 Kahn SE, Cooper ME, Del Prato S. Pathophysiology and treatment of Type 2 diabetes: perspectives on the past, present, and future. Lancet  383(9922), 10681083 (2014). \t Describes\tthe\tpathophysiology\tof\tType\t2\tdiabetes\t(T2D)\tin \t detail\twith\tprospective\tof\t -cell\tdysfunction\tand\tpotential",
+    "However, there are two major clinical types, type 1  diabetes (T1D) and type 2  diabetes (T2D), according to the etiopathology of t he disorder.  T2D appears to be the",
+    "type 1 diabetes, 723 (53%) had LADA, 162 (12%) had secondary diabetes (coexisting pancreatic disease), and 519 (38%) were unclassifiable because of missing data. The remaining 12  112 (883%) patients were considered  to have type 2 diabetes (appendix). To classify patients into novel diabetes subgroups, first",
+    "4   monogenic diabetes not only provides opportunities for etiology- based treatment of the  minority of individuals with highly penetrant variants, but also informs broader  understanding of diabetes etiology. Types of monogenic diabetes   Maturity onset diabetes of the young (MODY)   MODY comprises most  monogenic diabetes cases, with classical characteristics",
+    "19 RACIALIZED ETIOLOGIES OF DIABETES Diabetes is not one disease but many. More than 90 percent of all diabetics",
+    "with young-onset diabetes. Diabetologia 55:1265 1272 13. Schwartz SS, Epstein S, Corkey BE, Grant SF, Gavin JR 3rd, Aguilar RB (2016) The time is right for a new classification system for diabetes: rationale and implications of the -cell-centric classi- fication schema. Diabetes Care 39:179 186 14. Gale EAM (2006) Declassifying diabetes. Diabetologia 49:1989  1995 15. V oight BF, Scott LJ, Steinthorsdottir V et al (2010) Twelve type 2"
+  ],
+  "answer": "There are several types of diabetes. The classical classification includes type 1, type 2, other types, and gestational diabetes mellitus (GDM) [2]. Additionally, there are subtypes of adult-onset diabetes, including severe autoimmune diabetes (SAID), severe insulin-deficient diabetes (SIDD), severe insulin-resistant diabetes (SIRD), mild obesity-related diabetes (MOD), and mild age-related diabetes [4]. Monogenic diabetes, such as Maturity Onset Diabetes of the Young (MODY), also exists [8]. Therefore, the main types of diabetes are:  1. Type 1 diabetes 2. Type 2 diabetes 3. Other types of diabetes 4. Gestational diabetes mellitus (GDM) 5. Monogenic diabetes (e.g., MODY)  In summary, there are at least five main types of diabetes, with additional subtypes identified in recent studies.",
+  "question": "How many types of diabetes exist?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_7 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_7
new file mode 100644
index 0000000..2ec5b11
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_7
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2017 - Regular exercise participation improves genomic stability in diabetic patients an exploratory study to analyse telomere length and DNA damage.pdf",
+    "2009 - Antidiabetic drug metformin (GlucophageR) increasesbiogenesis of Alzheimer\u2019s amyloid peptides viaup-regulatingBACE1transcription.pdf",
+    "2016 - The dog aging project translational geroscience in companion.pdf",
+    "2018 - Type 2 Diabetes in adolescents and young adults.pdf",
+    "2004 - Diabetes Mellitus and Risk of Alzheimer Disease and Decline in Cognitive Function.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2016 - Whole-Genome Sequencing of a Healthy Aging Cohort.pdf",
+    "2010 - Genetics, pathogenesis and clinical interventions in type\u20091 diabetes.pdf",
+    "2016 - The genetic architecture of type 2 diabetes.pdf",
+    "2016 - Genetic predisposition for beta cell fragility underlies type 1 and type 2 diabetes.pdf"
+  ],
+  "extraction_id": [
+    "0e53122e-a308-55f7-8ee8-a0857ac9c52f",
+    "660377a1-3bd9-5628-ba52-4603b485267a",
+    "0d62683a-9b2d-535a-9464-15df3187bff3",
+    "ec07ba94-177b-599e-8104-92b2e0b943e6",
+    "d301fa41-68c5-5d02-94cc-ef3f83f2df8a",
+    "eb3de845-98db-505c-bb7f-c0f3259875fc",
+    "c55b4a12-6cc8-5594-87d4-53e4f8f023d1",
+    "6d537deb-a8fa-59cf-a961-aa3da56b18da",
+    "8b8b572d-68f5-5470-b5ed-ec5c6219dd5e",
+    "77681744-9c95-530f-afec-248e183fd78c"
+  ],
+  "document_id": [
+    "dcaf7b09-2d54-5cbf-b061-e3c4e6c6c518",
+    "78818ecc-83cb-5189-8eba-2ed7cb6d3a7f",
+    "e841c6bd-78b8-56e1-b3dd-e2bcc8a0f590",
+    "cc708325-df0a-55ec-9e9b-2bf97835c992",
+    "a6fca397-1f0f-5a8d-846d-78e92ef5c088",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "3a287979-e475-545b-99e6-4c1925653a79",
+    "46ac7ad7-a672-5d91-abf8-9c2f6307cd38",
+    "d7e2a9de-46f1-5191-9cb0-dd68eb9f365a",
+    "5b239c51-7b4c-58e0-acca-2061593fe317"
+  ],
+  "id": [
+    "chatcmpl-ADZQlbMdmyuTz6o9831V2evAJSxPf",
+    "a2125212-46c1-5020-ab62-196db673c2f5",
+    "962238ce-db77-5c91-8f41-33640d0bf501",
+    "6da64757-8c4d-5ec8-9c70-8da1be37af81",
+    "2747746a-5ede-5c96-aa43-8b71a29c1ce0",
+    "916717da-d554-5e4c-95d5-780d96c8bad2",
+    "9fc663d2-2833-51e7-ae6a-55b007a6e27c",
+    "19faf41b-7716-5244-a9c3-196c2e5cd477",
+    "5bc52c12-3339-542b-82a2-b839203370b9",
+    "13ca56ac-b751-5bc8-b557-e7a7a12a1b04",
+    "652c144e-94d8-519b-8d1f-1bcb2bf1b7b3"
+  ],
+  "contexts": [
+    "The biological processes linking aging and disease risk are poorly understood. Still, aging is considered to date as  one of the main factors responsible for several complex diseases including cancer, cardiovascular diseases, and diabetes. Particularly, type 2 diabetes (T2D) has become very prevalent all over the world, with a projected increas- ing growth rate for the years ahead 1. The pathophysiological mechanism that underlines diabetic complications",
+    "fects correlate with the functional alterations associated withaging of the brain and with AD pathogenesis (411). The vastmajority of AD cases are late onset and sporadic in origin withaging being the most profound risk factor. Insulin signaling isknown to be involved in the process of brain aging (1220).Insulin dysfunction/resistance in diabetes mellitus (DM) is notonly a common syndrome in the elderly but also considered a riskfactor for AD, especially for vascular dementia (21, 22). The link",
+    "striking similarities to people with respect to age-associ- ated increases in risk for several diseases, the relative risk for individual diseases is not always shared. For example,although the prevalence of type II diabetes in older dogs increases with age, it is still much lower than the current prevalence of type II diabetes in people, and the mostcommon form of diabetes in dogs resembles type I diabetes in people (Nelson and Reusch 2014 ). Whether this reects",
+    "strong inverse association between BMI and age at diagnosis of type 2 diabetes. When type 2 diabetes presents in later life, the severity of insulin resistance is often greater among individuals with a history of protracted and severe obesity, particularly with excess visceral adiposity. 28",
+    "COMMENT In a cohort of more than 800 older persons, we found thatdiabetes mellitus sometime in the study was associated withan increased risk of developing AD during a mean of 5.5years of observation. The risk of incident AD was 65% higherin those with diabetes mellitus than in those without it.Overall, results were similar in analyses restricted to dia-",
+    "insulin resistance, hypertension, and dyslipidemia (Obesity Education Initiative Expert Panel,  1998 ). Insulin resist-ance increases with age, and the incidence of diabetes rises sharply in the elderly (American Diabetes Association,  2010a ).   In a few patients, genetic mutations appear to be associ- ated with T2D (Roche  et al.  ,  2005 ; American Diabetes  Association,  2010a ). For example, recent work using the DPP data has led to the identi  cation of 27 single nucle-",
+    "et al., 2012 ), possibly due to the indirect and/or a mixed relation- ship between individual genetic disease risk loci and exceptional longevity (as discussed by Fortney et al., 2015 ) versus the poten- tially more direct relationship between aging in the absence of disease and overall genetic disease risk. On the other hand, no difference in genetic risk is observed for type 2 diabetes genetic risk and cancer. Some of these ndings (type 2 diabetes, colon, and lung cancer) can be explained by the",
+    "equal number of adults over 18 are thought to develop the disease,although incidence in older people receives less media/research attention. In this review, we discuss our current understanding of the cellular/molecular mechanisms of disease aetiology and progres-sion, the usefulness and limitations of rodent models of spontaneousdiabetes, the factors that are influencing the current increased inci-dence and the clinical opportunities for those affected.",
+    "associated with maturity onset diabetes of the young and early onset-age of  type 2 diabetes. J. Diabetes Complications 26, 343347 (2012). 19. Langenberg, C. et al. Design and cohort description of the InterAct Project:    an examination of the interaction of genetic and lifestyle factors on the incidence of type 2 diabetes in the EPIC Study. Diabetologia 54, 22722282  (2011).",
+    "in the precipitation of diabetes. Saturated fatty acids drive the apoptosis  and senescence of beta cells27,41, with increased oxidative stress42 and  endoplasmic reticulum stress41. As increased body mass index is asso - ciated with earlier onset of T1D43, it is possible that dietary fat is acting  as a sensitizer similar to insHEL, in effect lowering the threshold for  autoimmune stress to precipitate clinical diabetes. The male-specific susceptibility to diabetes in this model is in sharp"
+  ],
+  "answer": "Yes, there is a direct association between aging and susceptibility to having diabetes. Aging is considered one of the main factors responsible for several complex diseases, including diabetes, particularly type 2 diabetes (T2D) [1]. Additionally, insulin resistance, which is a key factor in the development of diabetes, increases with age, and the incidence of diabetes rises sharply in the elderly [6].",
+  "question": "Is there a direct association between aging and susceptibility to having diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_8 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_8
new file mode 100644
index 0000000..619d0b1
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_8
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2008 - Glossary of Genetics Genomics Terms.pdf",
+    "2012 - Predicting Diabetes Our Relentless Quest for Genomic Nuggets.pdf",
+    "2007 - Genome\u2013wide association studies provide new insights into type 2 diabetes aetiology..pdf",
+    "2004 - Diabetes Genes b.pdf",
+    "2013 - Systems Biology Approach Reveals Genome to Phenome Correlation in Type 2 Diabetes.pdf",
+    "2010 - Diabetes in Asia.pdf",
+    "2004 - Diabetic nephropathy Linking histology, cell biology.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "2011 - Type 2 diabetes across generations from pathophysiology to prevention and management.pdf",
+    "2004 - Diabetes Genes a.pdf"
+  ],
+  "extraction_id": [
+    "53e868dd-b318-5cf3-8b2e-98a548aab7cf",
+    "19b662f1-779b-57d2-bb51-e5505479ddc6",
+    "cd034e2b-72bd-5cda-a456-48cf17ead1bf",
+    "35324efd-a618-54aa-b9a5-541a9435134c",
+    "9369222f-e125-58c0-8f2b-cf5daa867f77",
+    "e99fe157-eda9-5e56-9ec9-8f428de2a161",
+    "7d723588-2c9a-5f6f-8ed1-f484fada8d13",
+    "4beabe81-e24e-535c-9df3-bfaa9cfdde90",
+    "baec13ec-c42b-51b4-9974-8ef1c2d10ddc",
+    "5a2221e0-dabc-523c-8358-3e43789e8f7a"
+  ],
+  "document_id": [
+    "c66d2572-071d-5aaf-829c-b3ca6cf6d697",
+    "7ef99c69-e777-521f-bb99-26bcb81de748",
+    "2ad9b6c6-56ed-5ba6-ad88-c1a6777f5196",
+    "da2f2624-e3e6-5e2d-b406-941db2fe7671",
+    "ea7c2799-c259-5d0e-b40b-ecebe0a9fc9f",
+    "0be842b8-7f69-503b-baed-c336e5c834d6",
+    "b0732aa4-4250-563c-b4e7-b98bb8706299",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "0f49b102-1d7e-5702-af30-35e5f2ed93a6",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa"
+  ],
+  "id": [
+    "chatcmpl-ADZQoBhwZsCCyM8hyEreAlAw5KAiT",
+    "54ff4672-bf7f-5158-b228-ca3d45e0cb0d",
+    "ceedd7ed-ae19-5971-89f8-cccc731741b6",
+    "12ee4112-ee8b-584a-bbab-b47863449868",
+    "65084469-bf7c-508c-a211-1f28f7626638",
+    "3bbf736e-7d8b-5e67-a4bf-e1ae28738bf3",
+    "6c14eef8-bb27-503a-9523-9e7a16d71021",
+    "8397e685-13d3-5487-a9c6-856cc119cef9",
+    "44725666-366f-5123-92dd-ab2cf29e88c1",
+    "2aa9f009-ae05-5c93-ac3a-58b1f516d844",
+    "54d5bc85-a2f5-58f6-814f-b511f2e0c4cf"
+  ],
+  "contexts": [
+    "Genetic factors appear to play a role in determining an individuals risk of developing diabetes. It is hoped",
+    "the diabetes epidemic, and its predilection for certain ethnic groups, are unknown. However, interactions between genetic pre-disposition and environmental triggers (or accelerants) are generally presumed to un- derlie the etiology of diabetes (3 5) (Fig. 1). The best known environmental risk factors are dietary habits, physical inactivity, and obesity; interventions that ameliorate theserisk factors prevent the development oftype 2 diabetes (6,7). By contrast, knowledge of the genetic",
+    "increases the risk of type  2 diabetes. Such a strong environmental component to a dis - ease should perhaps have deterred geneticists  from studying the disorder. However, there  are many obese people who do not suffer  from diabetes and many non-obese people  who do, showing that obesity is not the only  factor involved in the aetiology of type   2  diabetes (FIG. 1). In the past 10 years, geneticists have  devoted a large amount of effort to finding type   2 diabetes genes. These efforts have",
+    "future diabetes, however, is not possible on a genetic basis alone. For  example, the concordance rate for identical twins is < 50%, indicating that  either environmental or developmental events (such as T cell development)  affect the progression of diabetes.  The ability of serologic studies to identify individuals at risk for  diabetes in the general population is under investigation. Among relatives of  patients with diabetes, serologic markers can identify patients at high risk.3",
+    "genes relate directly to insulin secretion and indirectly, through collaborating with other genes, to insulin resistance. Thisseems to support the epidemiological evidence that environmentally triggered insulin resistance interacts with geneticallyprogrammed bcell dysfunction to precipitate diabetes. Citation: Jain P, Vig S, Datta M, Jindel D, Mathur AK, et al. (2013) Systems Biology Approach Reveals Genome to Phenome Correlation in Type 2 Diabetes. PLoS ONE 8(1): e53522. doi:10.1371/journal.pone.0053522",
+    "Genetic factors Type 2 diabetes has a strong genetic component and most Asian patients have a   rst-degree relative with diabetes. 48,49 Much progress has been made in our  understanding of the genetics of this disease. Importantly, most of the loci originally associated with diabetes in European populations have been replicated in Asian populations. Whereas monogenic forms of diabetes result from rare genetic mutations with large e  ects,  such as those seen in maturity-onset diabetes of young people,",
+    "literature abounds with evidence for genetic mediation ofthe initiation and progression of diabetic nephropathy.First, there is familial clustering that is not completelyexplained by environmental factors [3947]. Our indexcase and her family are perfect examples of genetic pre-disposition to diabetes and its complications, or, at thevery least, familial clustering. Parving and colleagues es-timated that glycemic control, hypertension, and albu-minuria account for only one-third of the variability",
+    "GENETIC MODELS OF DIABETES  Classically, genetic models of diabetes and obesity have been produced in  two ways. One is serendipitous observation of a spontaneously arising  extreme phenotype, followed by selective breeding to fix the trait. The  resulting model will often be monogenic, i.e. due to a single mutation. The  other approach is by repeated selective breeding of initially normal appearing members of a genetically diverse ( outbred) population that are at",
+    "36 Herder C, Roden M. Genetics of type 2 diabetes: pathophysiologic  and clinical relevance. Eur J Clin Invest 2011; 41: 67992. 37 Dabelea D, Hanson RL, Lindsay RS, et al. Intrauterine exposure  to diabetes conveys risks for type 2 diabetes and obesity: a study of discordant sibships. Diabetes 2000; 49: 220811. 38 Voight BF, Scott LJ, Steinthorsdottir V, et al. Twelve type 2 diabetes  susceptibility loci identi  ed through large-scale association analysis. Nat Genet 2010; 42: 57989.",
+    "Environmental influences interact with genetic factors to determine  susceptibility to type 2 diabetes by affecting either insulin action, insulin  secretion or both. The prevalence of type 2 diabetes has increased markedly  in populations that have rapidly adopted a Western lifestyle (for example the  Pima Indians) and in many populations that have migrated to regions with a  more affluent lifestyle compared to their native country (see Chapter IV.2)."
+  ],
+  "answer": "Genetics plays a significant role in influencing the emergence of diabetes. Several key points highlight this influence:  1. **Genetic Predisposition**: Genetic factors are crucial in determining an individual's risk of developing diabetes [1]. There is a strong genetic component to type 2 diabetes, with many patients having a first-degree relative with the disease [6].  2. **Interaction with Environmental Factors**: The development of diabetes is generally presumed to result from interactions between genetic predisposition and environmental triggers such as dietary habits, physical inactivity, and obesity [2]. This interaction suggests that while genetics set the stage, environmental factors act as accelerants.  3. **Insulin Secretion and Resistance**: Specific genes are directly related to insulin secretion and indirectly to insulin resistance. This supports the idea that genetically programmed beta-cell dysfunction, when combined with environmentally triggered insulin resistance, precipitates diabetes [5].  4. **Familial Clustering**: There is evidence of familial clustering of diabetes, indicating a genetic predisposition to the disease and its complications [7].  5. **Genetic Models**: Genetic models of diabetes have been developed through selective breeding, which often results in monogenic forms of the disease due to single mutations [8].  6. **Concordance in Twins**: The concordance rate for diabetes in identical twins is less than 50%, indicating that while genetics play a significant role, environmental or developmental events also affect the progression of diabetes [4].  In summary, genetics significantly influences the emergence of diabetes by predisposing individuals to the disease, which is then often triggered or exacerbated by environmental factors.",
+  "question": "How does genetics influence the emergency of diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_9 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_9
new file mode 100644
index 0000000..3b2d9f4
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_diabetes_9
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2010 - Pharmacogenetics of Anti-Diabetes Drugs.pdf",
+    "2002 - Genetic Effects on Age-Dependent Onset and Islet Cell.pdf",
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "2004 - Interaction and Association Analysis of a Type 1 Diabetes Susceptibility Locus.pdf",
+    "2008 - Clinical Risk Factors, DNA Variants.pdf",
+    "2011 - Obesity and Type 2 Diabetes What Can Be Unified.pdf",
+    "2017 - Spectrum of mutations in monogenic diabetes genes identified from high-throughput DNA sequencing of 6888 individuals.pdf",
+    "2010 - Genomics, Type 2 Diabetes, and Obesity.pdf",
+    "2015 - Diabetes mellitus The epidemic of the century.pdf",
+    "2015 - Diabetes mellitus The epidemic of the century.pdf"
+  ],
+  "extraction_id": [
+    "73e1aaff-7ef6-5ca2-9c94-23f5674a4f88",
+    "3dbc9428-663d-58ba-9268-3d87957bad4d",
+    "d7bd898b-1d46-557a-b065-f94fc5310b2a",
+    "d71343b2-f7c5-52b4-96f9-bcc98f97fe81",
+    "2643b341-8c50-5cea-af36-86a8b070a80e",
+    "037a2c9e-51be-5bd8-957a-4b3dfcc82e46",
+    "b24927c4-ee83-51a8-b431-b43be7d3b678",
+    "8a28c11f-e0d2-526b-ac85-2f2fbf054fc5",
+    "7cfe9f29-a0ee-56d3-be3b-1b238a43bc07",
+    "7cfe9f29-a0ee-56d3-be3b-1b238a43bc07"
+  ],
+  "document_id": [
+    "ffeebaf9-ff76-5751-9b8b-7a2a4a4f1dc3",
+    "2598ed03-43da-5894-8a62-41c8f53559d1",
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "4246f8d0-69e8-56cf-9674-d379467dfb61",
+    "0018610a-9c86-5e2d-a27d-f66cf4f8519d",
+    "07e715c3-4f95-54f4-8461-5062e62d0fe6",
+    "18a8a000-69ed-5d34-b13f-f5ae016d1067",
+    "3d629777-f1b6-5450-94ef-56736e5a4e10",
+    "e114dd28-fd39-56df-bdeb-8806474a6c10",
+    "e114dd28-fd39-56df-bdeb-8806474a6c10"
+  ],
+  "id": [
+    "chatcmpl-ADZQut2fnFLEnMmGsKclOiiWtUmw9",
+    "437a7129-63b8-5f34-8273-2eef9535e987",
+    "e2cc78dc-8af3-5459-bc88-84939f5aced4",
+    "47e8bd94-fd61-57f2-b1d0-cc139d71936a",
+    "a495dcc8-5cee-58a9-9f15-95be8fbc9b6a",
+    "aa72551a-ac0c-5d7d-8057-34f229f68eb1",
+    "9a42b2f0-cad6-56af-9aa4-0d2f9bedd351",
+    "748d13eb-eec0-5f79-8138-e3227a188b52",
+    "b4efc562-0077-5428-be43-f3eeafeb6847",
+    "d184bcc3-8c38-5969-859a-22db976fec35",
+    "3e22864f-a062-55b2-a9a3-a64cde8bd388"
+  ],
+  "contexts": [
+    "gene are associated with NIDDM in Caucasians. Diabetes 1996 , 45, 825-831.  46.  Tarasov, A.I.; Nicolson, T.J. ; Riveline, J.P.; Taneja, T.K. ; Baldwin, S.A.; Baldwin, J.M.;  Charpentier, G.; Gautier, J.F. ; Froguel, P.; Vaxillaire, M.; et al.  A rare mutation in ABCC8/SUR1  leading to altered ATP-sensitive K+ channel activ ity and beta-cell glucose sensing is associated  with type 2 diabetes in adults. Diabetes 2008 , 57, 1595-1604.",
+    "gene is associated with insulin-dependent diabetes mellitus. Diabetes 33:176 183, 1984 6. Bennett ST, Lucassen AM, Gough SCL, Powell EE, Undlien DE, Pritchard LE, Merriman ME, Kawaguchi Y, Drons eld MJ, Pociot F, Nerup J, Bouzekri N, Cambon-Thomasen A, R nningen KS, Barnett AH, Bain SC, Todd JA: Susceptibility to human type 1 diabetes at IDDM2 is determinedby tandem repeat variation at the insulin gene minisatellite locus. Nat Genet 9:284 292, 1995",
+    "of Diabetes   Results of several genome-wide association stud- ies (GWAS) have linked the following common gene variants with a 1520% increased risk of diabetes: reduced insulin secretion via reduce beta-cell mass (CDKAL1, CDKN2A, CDKN2B) and beta-cell dysfunction (MTNR1B, TCF7L2, KCNJ11) and increased insulin resistance related to obesity (FTO) and unrelated to obesity (IRS1, PPARG) [  11 ]. While most of the early studies",
+    "gene is associated with insulin-dependent diabetes mellitus. Diabetes 33:176 183, 1984 3. Nistico L, Buzzetti R, Pritchard L, Van der Auwera B, Giovannini C, Bosi E, Larrad M, Rios M, Chow C, Cockram C, Jacobs K, Mijovic C, Bain S,Barnett A, Vandewalle C, Schuit F, Gorus F, Tosi R, Pozzilli P, Todd J: TheCTLA-4 gene region of chromosome 2q33 is linked to, and associated with,type 1 diabetes: Belgian Diabetes Registry. Hum Mol Genet 5:1075 1080, 1996",
+    "ly associated with type 2 diabetes: TCF7L2, KCNJ11,   and PPARG . 5-7 However, in 2007, a number of novel  genetic variants ( CDKAL1, IGF2BP2,  the locus on  chromosome 9 close to CDKN2A/CDKN2B, FTO,  HHEX, SLC30A8,  and WFS1)8-14 were shown to in - crease susceptibility to type 2 diabetes in repro - ducible studies. Furthermore, a recent meta-analy - sis identified six novel variants ( JAZF1, CDC123/ CAMK1D, TSPAN8/LGR5, THADA, ADAMTS9, and NOTCH2 ) that are associated with type 2 dia - betes. 15",
+    "date gene approaches now have identified /H1101140 genes as- sociated with type 2 diabetes (17, 18) and a similar num-ber, albeit largely different, with obesity. Most type 2diabetes genes appear to be related to /H9252-cell dysfunction,",
+    "HNF1A ,HNF4A ,HNF1B ,INS,NEUROD1 ,PDX1 ,PAX4 , ABCC8 ,KCNJ11 ,KLF11 ,CEL, and BLK), 6 genes associ- ated with recessive diseases that include diabetes as a phenotype ( WFS1 ,NEUROG3 ,EIF2AK3 ,GLIS3 ,RFX6 , andSLC19A2 ), and 3 genes in which heterozygous mu- tations have been shown to cause diabetes mellitus (PAX6 ,GATA6 , and PPARG ). Our primary objectives were to (1) identify subjects with potentially undiag- nosed monogenic diabetes, (2) compare and contrast the",
+    "4. ORahilly S. Human genetics illumi - nates the paths to metabolic disease. Na - ture 2009;462:307-14. 5. McCarthy MI. Growing evidence for  diabetes susceptibility genes from genome scan data. Curr Diab Rep 2003;3:159-67. 6. Hattersley AT, McCarthy MI. What  makes a good genetic association study? Lancet 2005;366:1315-23. 7. Altshuler D, Hirschhorn JN, Klanne - mark M, et al. The common PPARgamma Pro12Ala polymorphism is associated with decreased risk of type 2 diabetes. Nat Genet 2000;26:76-80.",
+    "genes including interlukin-6 ( IL-6), tumor necrosis  factor- and IL-10 genes were found to be associated  with greater risk of developing type 2 diabetes[171], in  addition to genetic variants in the genes for IL12B ,  IL23R  and IL23A  genes[172]. In a study involving the  hormone sensitive lipase re sponsible for lipolysis  in adipose tissues, a deletion null mutation, which  resulted in the absence of the protein from adipocytes,  was reported to be associated with diabetes[173]. Nine",
+    "2 diabetes[144,149,150], however, not all of these genes  showed consistent and reproducible association with  the disease[151]. Genome wide association studies  (GWAS) in various populations identified 70 loci  associated with type 2 diabetes and revealed positive  linkage of many mutations and SNPs that influence  the expression and physiological impact of the related  proteins and risk to develop type 2 diabetes. One study  involved several thousand type 2 diabetes patients and"
+  ],
+  "answer": "Several genes have been associated with diabetes, both type 1 and type 2, according to the provided context:  1. **Type 1 Diabetes:**    - **Insulin gene minisatellite locus (IDDM2)**: Susceptibility to type 1 diabetes is determined by tandem repeat variation at this locus [2].    - **CTLA-4 gene region of chromosome 2q33**: This gene region is linked to and associated with type 1 diabetes [4].  2. **Type 2 Diabetes:**    - **ABCC8/SUR1**: A rare mutation in this gene affects ATP-sensitive K+ channel activity and beta-cell glucose sensing, leading to type 2 diabetes [1].    - **Common gene variants**: Several genome-wide association studies (GWAS) have linked common gene variants with increased risk of type 2 diabetes, including:      - **CDKAL1, CDKN2A, CDKN2B**: Associated with reduced insulin secretion via reduced beta-cell mass [3].      - **MTNR1B, TCF7L2, KCNJ11**: Associated with beta-cell dysfunction [3].      - **FTO**: Associated with increased insulin resistance related to obesity [3].      - **IRS1, PPARG**: Associated with increased insulin resistance unrelated to obesity [3].    - **Other genes**: Additional genes associated with type 2 diabetes include:      - **TCF7L2, KCNJ11, PPARG, CDKAL1, IGF2BP2, FTO, HHEX, SLC30A8, WFS1** [5].      - **JAZF1, CDC123/CAMK1D, TSPAN8/LGR5, THADA, ADAMTS9, NOTCH2** [5].    - **Genes related to beta-cell dysfunction**: Many type 2 diabetes genes appear to be related to beta-cell dysfunction [6].    - **Monogenic diabetes genes**: Genes such as **HNF1A, HNF4A, HNF1B, INS, NEUROD1, PDX1, PAX4, ABCC8, KCNJ11, KLF11, CEL, BLK** are associated with monogenic forms of diabetes [7].    - **Genes associated with recessive diseases that include diabetes**: **WFS1, NEUROG3, EIF2AK3, GLIS3, RFX6, SLC19A2** [7].    - **Genes with heterozygous mutations causing diabetes mellitus**: **PAX6, GATA6, PPARG** [7].    - **Inflammatory genes**: Genes such as **IL-6, TNF-\u03b1, IL-10, IL12B, IL23R, IL23A** are associated with a greater risk of developing type 2 diabetes [9].  These references provide a comprehensive list of genes associated with both type 1 and type 2 diabetes.",
+  "question": "what genes are associated with diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_1 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_1
new file mode 100644
index 0000000..93c6c86
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_1
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2019 - Shared and distinct genetic risk factors for childhood-onset.pdf",
+    "2008 - Genetic Effects on Environmental Vulnerability to Disease Novartis Foundation Symposium 293.pdf",
+    "2018 - The Genetics and Genomics of Asthma.pdf",
+    "2010 - Recent advances in the genetics and genomics of asthma.pdf",
+    "2019 - Shared and distinct genetic risk factors for childhood-onset.pdf",
+    "2008 - Genetic Effects on Environmental Vulnerability to Disease Novartis Foundation Symposium 293.pdf",
+    "2019 - Leveraging genomics to uncover.pdf",
+    "2018 - The Genetics and Genomics of Asthma.pdf",
+    "2018 - The Genetics and Genomics of Asthma.pdf",
+    "2019 - Shared and distinct genetic risk factors for childhood-onset.pdf"
+  ],
+  "extraction_id": [
+    "306aeeff-1e54-5e8d-9d06-10fc3c995f69",
+    "110ccbf7-ee1f-5326-ac9a-a5ea4a842751",
+    "47d93beb-84d7-55ab-af76-a671ea6dc488",
+    "d398c492-cb9a-5aba-bfd4-4a51bd6eb831",
+    "cfc7a3ec-7c07-5966-a022-decf1c4f8276",
+    "22dfc0a4-f5ac-5fa0-911a-32a5e71c8608",
+    "531fbafe-c4a1-55b2-b832-dfb9a7f67e96",
+    "af4e210e-5d90-5f49-996a-fa177eaf155e",
+    "b95d9c35-eefe-579a-bf1d-12811d1b5a07",
+    "cfc7a3ec-7c07-5966-a022-decf1c4f8276"
+  ],
+  "document_id": [
+    "6f5caedc-c2b4-54e4-8bc6-8b78cab6bb0e",
+    "5d65e407-34e5-5c1c-b394-989b7a09b57d",
+    "47cf2a47-d8d2-583b-8c12-2a7cfbe92e5e",
+    "656c0e8b-d154-551a-bae1-986e418a6aa4",
+    "6f5caedc-c2b4-54e4-8bc6-8b78cab6bb0e",
+    "5d65e407-34e5-5c1c-b394-989b7a09b57d",
+    "5da46d3b-fa82-57f6-b3e5-c82784347881",
+    "47cf2a47-d8d2-583b-8c12-2a7cfbe92e5e",
+    "47cf2a47-d8d2-583b-8c12-2a7cfbe92e5e",
+    "6f5caedc-c2b4-54e4-8bc6-8b78cab6bb0e"
+  ],
+  "id": [
+    "chatcmpl-ADZIhpxkquCUzShWFEamc7p1ntYgw",
+    "21c02be1-f8a1-5d70-abb7-00a866f4a734",
+    "183e3af7-7138-554b-bcda-b76e7eeef30a",
+    "21d3694f-032b-5d8c-93e2-58e85ec92903",
+    "b64a52ee-d2e8-50a7-a101-b255cd905180",
+    "46ee340b-b11c-52ef-a48d-d2c8135b63d8",
+    "1fa74fa4-0f13-55be-8164-ee57dfbc80af",
+    "545f48b6-3b45-5a1a-8def-7a5900ecc40a",
+    "a0a06ce6-5b8f-52ff-8904-901ae666f247",
+    "4d8496c2-5415-58d4-a27d-8f0f7f8d147a",
+    "47361fa0-f820-5114-8074-ad63e0815d81"
+  ],
+  "contexts": [
+    "children is driven more by dysregulated allergy and epithelial barrier function genes, whereas the cause of adult-onset asthma is more lung-centred and environmentally determined, but with immune-mediated mechanisms driving disease progression in both children and adults. Funding US National Institutes of Health. Copyright  2019 Elsevier Ltd. All rights reserved. Introduction Asthma is the most prevalent chronic respiratory disease  worldwide.1 The diagnosis of asthma is based on the",
+    "asthma has increased with alarming frequency in industrialized cities worldwide (e.g. Elias et al 2003). These diseases generally are complex, with clear contribu-tions of genetic background and exposure to environmental stimuli (see Kleeberger & Peden 2005). It is unlikely that the increased incidence in disease can be attributed only to genetics as increases in disease-causing genetic mutations to account for the increase would require multiple generations. Therefore the role of environmental exposures",
+    "living all represent risk factors for asthma, while early farm exposures and breastfeeding confer protective effects. Such observations have been assimilated into the hygiene hypothesis, rst set out in 1989 (136), positing that reduced early microbial exposure and its impacts on immunity underliethe postIndustrial Revolution atopy and asthma epidemic. Responsible for a transformation in our understanding of microbial factors in asthma has been a revolution of a different kind. Only",
+    "tobacco smoke exposure and with early-onset asthma (before age 4) [49/C15/C15]. Further studies of preschool asth- matics have shown the 17q21 variants are associated with an almost two-fold increased risk of developing recurrent wheeze, asthma, asthma exacerbations and bronchial hyper-responsiveness, but are not associated with eczema, rhinitis or allergic sensitization, indicating that they are specic determinants of nonatopic asthma in children [47].",
+    "for childhood-onset asthma supports the widely held idea that asthma in childhood is due to impaired barrier function in the skin and other epithelial surfaces. This model proposes that compromised epithelial barriers promote sensitisation to food and airway allergens and to wheezing illnesses in early life. 46,47 In fact, childhood  onset-specific loci identified in this study have been associated with atopic dermatitis or food allergies, such as FLG on 1q21.3 with the atopic march, 41 atopic",
+    "relation to asthma and other atopic diseases). The prompt in the asthma example came from the observation of the apparent effect of being reared in a farm envi-ronment. Of course, it was crucial to replicate that observation in different social contexts and it was also important to have some leverage on a likely biological mediating pathway (in that case exposure to endotoxins). Similarly, the G  E",
+    "[11] Shaaban R, Zureik M, Soussan D, Neukirch C, Heinrich J, Sunyer J, et al. Rhinitis and onset of asthma: a longitudinal population-based study. Lancet (London, England) 2008;372(9643):104957.  [12] de NijsSB, VenekampLN, BelEH. Adult-onset asthma: is it really different? Eur Respir Rev 2013;22(127):44.  [13] RackemannFM. Intrinsic asthma. J Allergy 1940;11(2):14762.  [14] JarvisD, NewsonR, LotvallJ, HastanD, TomassenP, KeilT, etal. Asthma in adults and its as -",
+    "GG19CH10_Cookson ARI 26 July 2018 9:47 Epigenetic Features of Asthma: Within the Lung A study of the epigenome in primary airway epithelial cells from 74 asthmatic and 41 non-asthmatic adults (111) revealed a regulatory locus on chromosome 17q1221 (the same locus identied by asthma GWASs) associated with asthma risk and epigenetic signatures of specic asthma endo-types. ORMDL3 expression was related to the differentially methylated region at this locus, while",
+    "studies have identied a range of pre-, peri-, and postnatal environmental factors, including modeof delivery, diet, and early lower respiratory tract infection, that confer relative risk or protection. Attempts to map the genetic architecture of asthma have identied a broad spectrum of potential contributory genes. Many of these genes demonstrate inconsistent patterns of replication betweencohorts, most likely reecting a combination of true positive and true negative results and the",
+    "49  Variants at those loci were all associated with earlier age of asthma onset. We further showed that these loci are associated with childhood-onset asthma, even after exclusion of patients with a history of allergic diseases in prespecified analyses, suggesting both a crucial role for the allergic diathesis in the development of asthma in childhood and a shared architecture between allergic disease and childhood-onset asthma. 33,46 By contrast, the enrichment for genes highly expressed"
+  ],
+  "answer": "The most cited environmental factor for the onset of asthma is exposure to tobacco smoke, particularly in early childhood [4]. Additionally, reduced early microbial exposure, as posited by the hygiene hypothesis, is also a significant environmental factor [3].",
+  "question": "What is the most cited environmental factor for the onset of asthma?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_10 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_10
new file mode 100644
index 0000000..d736997
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_10
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2020 - Clinical Genetics and Genomics of Aging.pdf",
+    "2017 - Regular exercise participation improves genomic stability in diabetic patients an exploratory study to analyse telomere length and DNA damage.pdf",
+    "2008 - Telomeres and Aging.pdf",
+    "2002 - Mitochondrial dysfunction leads to telomere attrition.pdf",
+    "2006 - Sex-specific telomere length profiles.pdf",
+    "2018 - Sex Differences in Aging Genomic Instability.pdf",
+    "2008 - Telomeres and Aging.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2006 - Sex-specific telomere length profiles.pdf",
+    "2010 - Roles of Werner syndrome protein in protection of genome integrity.pdf"
+  ],
+  "extraction_id": [
+    "efd18101-9cf2-56b5-8f86-c2aba6caa0bc",
+    "0e53122e-a308-55f7-8ee8-a0857ac9c52f",
+    "13990eb4-bef2-58ce-bf3e-0e3bc294caab",
+    "b92ede07-74a7-524a-8d2c-54b2559e8425",
+    "6d3bfe47-f26e-50dc-8d77-19f3797e53a0",
+    "396708f1-aa0a-571e-a8d3-7cb8404e9502",
+    "e57aa746-20f1-50b3-b8ab-3139a9a910fc",
+    "3b0cb0ab-421d-54d7-9816-c6a2e6f1ac68",
+    "eb8d8e40-a484-57cb-8125-3fd5eb3f6389",
+    "32528f9c-b6bb-593e-94c5-1ed12d0ac4ad"
+  ],
+  "document_id": [
+    "62b635c3-040e-512a-b016-6ef295308a1e",
+    "dcaf7b09-2d54-5cbf-b061-e3c4e6c6c518",
+    "61d9c326-d36e-55c1-a891-335dc943e70f",
+    "d8bc729b-7513-58b7-b12e-0db1fb6d3b7d",
+    "09c78a17-4a1f-52c1-be4d-994fd9fd71d0",
+    "8cfb5529-7f0c-58fc-b6e4-b3ee800fb72f",
+    "61d9c326-d36e-55c1-a891-335dc943e70f",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "09c78a17-4a1f-52c1-be4d-994fd9fd71d0",
+    "ec3e4f66-1619-5f71-9860-c1ad048d1841"
+  ],
+  "id": [
+    "chatcmpl-ADZJpRmTN4COm0TDjwpOtSCKK6Mex",
+    "28e98b7e-f273-5bdd-9979-185133f311af",
+    "bb069c10-45f1-5a83-95e3-4b7655874ba7",
+    "5f940245-af1d-5eee-84dc-942017c523d0",
+    "7fad29bd-12bf-53d0-af89-aadd38b974ff",
+    "607cbd31-d430-5517-8212-208b25af32bf",
+    "53508a9e-d064-58a3-a4f9-0785470a1462",
+    "36de43a5-e151-5300-8c34-ed15ec66ea52",
+    "f181e6da-58b6-5f26-87a2-355e25388673",
+    "64ef9964-1831-5a7a-8a69-5e8d0c332d37",
+    "dd9a3905-0225-5345-891b-4469af6336ee"
+  ],
+  "contexts": [
+    "Telomeres are arrays of linked nucleotide hexamer repeats that are found at the  ends of chromosomes in a vast clade of organisms [14]. While the sequence of  these telomeric repeats can vary between organisms, their biological function is  highly conserved, which is to limit damage inflicted on genes during the replica- tion of chromosomes. Telomere length is progressively shortened with each round  of genomic replication, unless it is restored through the action of a ribonucleo-",
+    "repetitive nucleotide sequences at the end of each eukaryotic chromosome, which protects them from attrition and damage. Although the relationship between leukocyte telomere length (LTL) and diabetes is still questioned 8,  different studies have shown that T2D individuals have shorter leukocyte telomeres than non-T2D individuals9, 10   that may be associated with disease progression11. Indeed, the decreased antioxidant capacity described in patients",
+    "telomere length,a phenomenon attributed to higher levels of oxidativestress at the cellular level (70). More recent studies havelinked telomere length in smooth muscle cells with senes-cence and disease severity in patients with atherosclero-sis (141, 150). Leukocyte telomere length was also short ina cohort of similar patients and associated with a higherrisk of developing occult cardiovascular disease (71).More data are needed to understand and validate the useof leukocyte telomere length as a biomarker",
+    "TTAGGG sequence that cap the ends of chromosomes, protect-ing them from degradation and fusion. The length of telomererepeats is primarily maintained by active telomerase, which iscomposed of Telomerase RNA (TR) and a catalytic subunitTelomerase Reverse Transcriptase (TERT) (Blackburn, 2001).Extensive evidence has shown that telomere shortening anderosion lead to chromosome end-to-end fusions and genomicinstability (Blasco   et al  ., 1997; Hande   et al  ., 1999), causing",
+    "age telomere length through accumulation of several short telo- meres (Londono-Vallejo et al., 2001; Martens et al., 2000) is  responsible for senescence or whether a speci  c chromosome  arm limits the replication potential of human cells (Hemann et al., 2001). Individual chromosome arms were shown to have  large variations in their length (Lansdorp et al., 1996; Benn,  1997; Londono-Vallejo et al., 2001), and chromosome 17p seemed to be equipped with especially short telomeres in hu-",
+    "Telomeres are specialized structures that protect the ends of linear chromosomes. They shorten during aging due to the unidirectional activity of DNA polymerase, which leaves a section of DNA unrepli-cated on the lagging strand. Telomeres also are subject to shortening by genotoxic stress, such as oxidative damage (33). Among many eukaryotes, the enzyme telomerase maintains telomere length; but telomerase activity varies over the lifespan and between cell types, tissues, and species (34). In most human",
+    "ends. For example, chromosome 17p typi-cally has shorter telomeres than most other chromosomeends (26, 137). In human nucleated blood cells, the aver-age telomere length shows a highly signicant declinewith age that is most pronounced for the cells of theimmune system (Fig. 2). Telomeres prevent the ends oflinear chromosomes from appearing as DNA double-strand (ds) breaks and protect chromosome ends fromdegradation and fusion. It has been proposed that telo-meres can switch between an open state (in",
+    "telomeres, the repetitive sequence at the end of linear chromosomes, has garnered much attention for its relation to aging. Telomere repeats serve as an internal clock for cycling cells because each round of replication results in the loss of telomeric DNA in the absence of active telomerase (reviewed in [66]). Eventually, this loss over cellular generations culminates in telomere crisis and a permanent state of",
+    "a pivotal role in maintenance of genomic integrity and func-tion (Moyzis et al., 1988; McElligott and Wellinger, 1997; van  Steensel and de Lange, 1997).   It is generally accepted that telomeres shorten during DNA  replication both in vitro and in vivo. In individuals, short telomeres are considered to be a sign of advanced age. Cawthon  and coworkers (2003) showed that telomere shortening in hu-",
+    "Telomeres are nucleoprotein complexes situated at the ends of thelinear chromosomes that prevent chromosome termini from beingrecognized as broken DNA ends ( i.e., DSBs). In most of the organisms studied, telomeres consist of long repetitive G-rich and C-rich DNAstrands, the ribonucleoprotein telomerase, and telomere bindingand associated proteins [179] . Loss of telomeric repeats or loss of"
+  ],
+  "answer": "The length of telomeres is significant for several reasons:  1. **Protection of Chromosomes**: Telomeres are repetitive nucleotide sequences at the ends of chromosomes that protect them from attrition and damage during replication [2], [4], [7]. They prevent chromosome ends from being recognized as broken DNA ends, which could otherwise lead to degradation and fusion [4], [10].  2. **Aging and Cellular Senescence**: Telomere length decreases with each round of genomic replication due to the unidirectional activity of DNA polymerase, which leaves a section of DNA unreplated on the lagging strand [1], [6]. This progressive shortening acts as an internal clock for cellular aging, eventually leading to cellular senescence when telomeres become critically short [8].  3. **Disease Association**: Shorter telomeres have been linked to various diseases. For example, individuals with Type 2 Diabetes (T2D) have been shown to have shorter leukocyte telomeres, which may be associated with disease progression [2]. Similarly, shorter telomeres in smooth muscle cells have been linked to senescence and disease severity in patients with atherosclerosis [3]. Leukocyte telomere length has also been associated with a higher risk of developing cardiovascular disease [3].  4. **Genomic Stability**: Telomere shortening and erosion can lead to chromosome end-to-end fusions and genomic instability, which are detrimental to cellular function and can lead to various genetic disorders [4].  5. **Biomarker Potential**: Due to their association with aging and disease, telomere length is being studied as a potential biomarker for various conditions, including cardiovascular diseases and diabetes [3].  In summary, telomere length is crucial for maintaining chromosome integrity, regulating cellular aging, and potentially serving as a biomarker for disease progression and risk assessment.",
+  "question": "What is the significance of the length of telomeres?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_11 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_11
new file mode 100644
index 0000000..2ea8436
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_11
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - Opening Up the Conversation on Genetics.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2007 - Promoting_Student_Scientific_Literacy_of_Molecular Genetics and Genomics.pdf",
+    "2019 - Sexual Dimorphism in the Age of Genomics How, When, Where.pdf",
+    "2008 -  Study Design and Statistical Issues.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2008 - Genotype-phenotype relationships and the patterning of complex traits as exemplified in the mammalian dentition.pdf",
+    "2019 - The influence of paternal diet on sncRNA-mediated epigenetic.pdf",
+    "2019 - Mother or Father who is in the front line.pdf"
+  ],
+  "extraction_id": [
+    "51dbd5e2-fde6-5097-aa05-fcf57d3ca6b1",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "67369433-749b-5d6a-b5ef-3f0afe78b767",
+    "e22bb6fb-bec4-5c4c-8690-c96d0b8d13d4",
+    "06bf0605-388a-592c-96ad-3a53bb36362c",
+    "8a1ce8fa-b5f4-5942-b7b1-14a8a7887710",
+    "5aab3e60-b8b0-52ad-b4d3-817cf012cfa5",
+    "84335575-34d7-56b6-aa06-5a8ac13d637a",
+    "297793b1-93f9-5626-ac63-6d8675c02d27"
+  ],
+  "document_id": [
+    "b62a8f54-c2f5-5bbb-9324-af80f7537167",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "755f34c4-cc06-5275-a744-16d48162b012",
+    "3f8c03b0-4235-5774-9d26-e43d55c1001b",
+    "c3bd9cf0-f768-55c4-be94-96590d7acc21",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "f6e866b8-b233-5862-bfb8-9949d0dabb97",
+    "dfcbd6e6-f60d-5eb7-867b-34ec78415e82",
+    "8011b04b-2199-5913-b8da-42c83334d4b7"
+  ],
+  "id": [
+    "chatcmpl-ADZK3Fpc5jWofKuSsq6lJRZP4Zmhy",
+    "60ad1512-b0c0-59cd-ace4-c146e2c04b52",
+    "a66b8b00-d51c-575b-b6ac-fa445c4ca715",
+    "df4c6108-740d-5bcf-99e6-dbda74f7e41a",
+    "8e3fdc2c-0962-5854-83e7-a60ab05cf6de",
+    "0158f264-120f-5942-ad55-ef5fde1f188a",
+    "1e151ad5-59d9-598d-97ba-90ba0e64c4cb",
+    "4472740a-d22d-5bb1-98e3-e91332cbb303",
+    "47b9142f-98a3-5a45-8eaa-d327c9cc055d",
+    "be93ee68-72ae-5015-a3f0-19e7bf24827a",
+    "53364cbf-8069-50ec-b008-5d7f7a8ea1b8"
+  ],
+  "contexts": [
+    "the egg and the sperm. Such a process would result in genetic changes that will be copied into every cell of the future adult, including reproductive cells (Stock & Campbell, 2000), opening the door to irreversibly alter the human species. Inevitably, signifi  cant self-disclosure and discussion challenges await families",
+    "a fertilized egg is a complicated process that relies on controlling: which genes are active; whenthese genes activate; and for how long they are active. In broad terms, there are four ways that thiscontrol can be achieved: First, inside the sperm or egg, genes can be marked with small chemical tags that flag these genes",
+    "to be activated (or remain inactive) after fertilization, depending on whether the modification wasmade by the father (in the sperm) or the mother (in the egg); this process is known as imprinting. Second, the mother can alter the gene activity in her offspring via the placenta; this process is known as maternal effect. Third, instructions encoded within the embryos DNA can directly control if, andwhen, a nearby gene becomes activated; this is known as  cis-regulation. Finally, similar instructions",
+    "genes.  An altered gene may be passed on to every cell that develops from it.  The resulting features my help, harm, or have little or no effect on the offsprings success in its environment. (AAAS, pg. 109, 5B:9-12#4 ) 6. Heritable material: The information passed from parents to offspring is coded in DNA molecules (AAAS, pg 108, 5B:9-12#3) 7. Mutagens: Gene mutations can be caused by such things as radiation and chemicals.  When they occur in sex cells, the mutations can be passed onto offspring; if they",
+    "sex chromosome effects. (B)Soon after fertilization, male and female cells have sex-specic transcriptomes, epigenomes, and phenotypes (for example, male embryos grow faster than female embryos). At implantation, lineage determination begins and gene expression differences are reduced. Epigenetic marks, however, are less constrained and some are maintained, affecting gene expression, and phenotype later in development. Once specic lineages are established, differences in",
+    "phenomena such as mutations and gene conversion events) occur in relevant meioses  leading up to the formation of the gametes (i.e., egg and sperm) which are combined  during fertilization and the formation of zygotes. Thus, individuals inherit a patch- work of chromosomal segments from maternal and paternal chromosomes.",
+    "(Figures 8 and 9). Two gametes (egg and sperm) ultimately  join into a single cell, the zygote, which has the full comple-ment of 23 chromosome pairs restored. If all goes well, the zygote gives rise to a live offspring. The Mendel Laws: Segregation and Independent  Assortment Both of the Mendel laws pertain directly to the process of  meiosis. The first Mendel law, the law of segregation, states  that each parent passes a randomly selected allele for a given",
+    "the subset of that genetic information that is active.  But how does the differentiation process  begin?  The key insight in resolving this conundrum came from fly genetics and was the  realization that the egg is not a homogenous sack of protoplasm.  The maternally-derived genes  active in the fertilized egg are asymmetrically distributed such that at the first cell division each  daughter cell receives a different complement of factors.  Development continues as a",
+    "spermatozoa: more than the sum of its parts? DNA, histones, pro - tamines and epigenetics. Reproduction 139:287301 Nilsson EE, Sadler-Riggleman I, Skinner MK (2018) Environmentally  induced epigenetic transgenerational inheritance of disease. Envi-ron Epigenet 4:dvy016Pembrey M, Saffery R, Bygren LO, Network in Epigenetic Epide-",
+    "mediated through the transmission of epigenetic  information through the paternal sperm cells  [6,80,81].  4.1. Persistence of Maternal Exposure to A dverse Environmental Conditions along Generations  In some cases, developmentally programmed traits may simply be the result of persistent or  replicated exposure during critical periods of deve lopment, generation after generation. It has been  suggested that the history of seve re socio-political disruptions and economic disadvantage suffered"
+  ],
+  "answer": "Once a sperm combines with an egg, the traits passed on to the resulting lifeform are determined by several factors:  1. **Genetic Information**: The fertilized egg, or zygote, contains a full complement of 23 chromosome pairs, which includes a mix of genetic material from both the maternal and paternal chromosomes [7]. This genetic information is coded in DNA molecules and is heritable [4].  2. **Gene Activation and Regulation**: The process of trait determination involves controlling which genes are active, when they activate, and for how long they remain active. This control can be achieved through several mechanisms:    - **Imprinting**: Genes can be marked with small chemical tags that flag them to be activated or remain inactive, depending on whether the modification was made by the father (in the sperm) or the mother (in the egg) [2], [3].    - **Maternal Effect**: The mother can alter gene activity in her offspring via the placenta [3].    - **Cis-Regulation**: Instructions encoded within the embryo's DNA can directly control if and when a nearby gene becomes activated [3].  3. **Epigenetic Marks**: Soon after fertilization, male and female cells have sex-specific transcriptomes, epigenomes, and phenotypes. Epigenetic marks, which are less constrained, can be maintained and affect gene expression and phenotype later in development [5].  4. **Mutations and Environmental Factors**: Gene mutations caused by factors such as radiation and chemicals can also be passed on to offspring if they occur in sex cells [4]. Additionally, environmental conditions experienced by the parents can influence the traits of the offspring through epigenetic mechanisms [10].  These combined factors ensure that the resulting lifeform inherits a unique set of traits from both parents, influenced by genetic, epigenetic, and environmental factors.",
+  "question": "Once a sperm combines with an egg, what determines how traits are passed on to the resulting lifeform?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_12 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_12
new file mode 100644
index 0000000..6d71c04
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_12
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - When Family Means More (or Less) Than Genetics.pdf",
+    "2012 - Mitochondrial Genomic Analysis of Late Onset.pdf",
+    "2017 - Parental influence on human germline de novo.pdf",
+    "2020 - Mitonuclear genomics and aging.pdf",
+    "2015 - Self-reported race or ethnicity in the age of genomic.pdf",
+    "2009 - When Family Means More (or Less) Than Genetics.pdf",
+    "2017 - Parental influence on human germline de novo.pdf",
+    "2016 - A genetic method for dating ancient genomes provides.pdf",
+    "1996 - IDDM2-VNTR-encoded Susceptibility to Type 1 Diabetes.pdf",
+    "2012 - Mitochondrial Genomic Analysis of Late Onset.pdf"
+  ],
+  "extraction_id": [
+    "baf15552-4198-5701-9175-c3fd31b2068e",
+    "ed29f84f-f2c9-5cbe-bab1-f5d5d2a334b6",
+    "a3b7edd7-f50f-53f1-b875-6d6733ddfde9",
+    "472c8adc-54e7-5c27-a7b8-882b7e49cd2b",
+    "6d68e979-ad62-5f85-ab03-5e898ce1c73b",
+    "baf15552-4198-5701-9175-c3fd31b2068e",
+    "163ce027-26ce-5625-8b63-5b7a910b4462",
+    "fcf5296e-6be4-5789-b1e1-ac57fef15119",
+    "a324397e-1525-55ff-a9e8-92dc2aafa237",
+    "ed29f84f-f2c9-5cbe-bab1-f5d5d2a334b6"
+  ],
+  "document_id": [
+    "7ba44399-3765-5ef5-9fdd-119b62117f66",
+    "5404a17c-34a9-5881-8b1a-2acacdc996a8",
+    "7c8bee23-b142-5fce-be77-6910277a136f",
+    "e05fdc09-c8d8-5134-a1fd-bf07a1564981",
+    "51ff0b84-193b-525a-b686-f29a423fcef9",
+    "7ba44399-3765-5ef5-9fdd-119b62117f66",
+    "7c8bee23-b142-5fce-be77-6910277a136f",
+    "5a5e67ea-4830-5fe8-95c3-ccfcc8324036",
+    "bbaa99aa-3ae9-558d-bc97-7f85b6d0cf81",
+    "5404a17c-34a9-5881-8b1a-2acacdc996a8"
+  ],
+  "id": [
+    "chatcmpl-ADZKEWDIh0sRr2bP9lVmh6YIK1qiT",
+    "f19ba98e-963f-5ecf-ad88-47215a3096e1",
+    "0e3b3480-c288-53cb-ac18-1d57478f9d34",
+    "06d4d82e-6eb9-59aa-a762-64de13149041",
+    "99a2cfc1-5a54-53af-b2a4-4c274e1d5ef1",
+    "612366c9-fcdc-5081-bc6d-47cd39922eeb",
+    "2ca2ab07-78b5-5268-93f1-297d83447163",
+    "db1fe67a-3d0c-549f-a54a-74ea0fa44d11",
+    "74484e0c-c862-5091-9fb5-957453a069af",
+    "74ef6cdc-ea40-5d10-9ee8-b4288b3a70b4",
+    "27f40683-de33-5ec1-852d-6905f2dc389c"
+  ],
+  "contexts": [
+    "variation with cultural practices around lineage. In certain societies, individuals place greater importance on (and have greater knowledge about) one side of the family than another (unilineal descent). Thus, individuals in patrilineal groups trace relationships through males only so that your fathers brothers children are members of your family, but not your fathers sisters (Kottak, 2007 ). They are members of their husbands group or family. Efforts to create",
+    "maternal lineage membership with those who weredirectly genotyped. Based on these pedigree (matrilineal) relation-",
+    "in three-generation families, and read pair tracing DNMs with phased variants. In the former approach, we determined the parent of origin as in our previous  analysis4. For example, if an offspring of the proband was a carrier of the DNM  allele and had haplotype sharing to paternal chromosome of the proband, we  assigned the mutation to the father. Meanwhile, if the offspring was not a DNM  allele carrier, we would assign it to the maternal germline. We restricted the haplo -",
+    "Unlike the nuclear genome, which requires both paternal and maternal contributions, mtDNA is inherited solely from the maternal lineage. It is unclear what advantage a uniparental mtDNA transmission confers, but one possibil-ity is to minimize the number of distinct genomes to maxi-mize the efficiency of a multi-genomic system (Hill etal. 2019). In fact, humans have developed complex, redundant mechanisms to ensure uniparental inheritance of mtDNA (DeLuca and OFarrell 2012; Rojansky etal. 2016). Paternal",
+    "c) Mitochondrial DNA (maternal line testing) markers: mitochondrial DNA or mtDNA haploid is the maternally inherited mitochondrial genome (mtDNA) [ 44]. All children inherit mtDNA from their mother, with no admixture from the father. Like Y-line DNA, mtDNA is passed intact from one generation to the next but through maternal line. Mitochondrial DNA does not follow any surname. In fact, the surname changes in every generation when women marry. Polymorphisms of mtDNA",
+    "a family pedigree may be hampered if the participant is not familiar with her mothers relatives, but her mothers brothers children (her cousins) may be able to supplement her overall family history. Knowledge about the cultural system of unilineal descent avoids assuming the universality of bilateral descent. Cultural beliefs such as these also have implications in the conduct of genetic research in terms of confidentiality and autonomy (Benkendorf et al.,",
+    "225 three-generation families using haplotype sharing (Fig. 1c and  Methods), 80.4% were found to be of paternal origin (Extended Data  Fig. 1). Figure 1e shows a strong relationship between the number of  paternal DNMs and the fathers age at conception (1.47 per year, 95%  CI 1.341.59) and a weaker impact of the mothers age on the number  of maternal DNMs (0.37 per year, 95% CI 0.300.45). The parental origin of all DNMs was also assessed by read pair",
+    "genetics-based population divergence studies. Am J Phys Anthropol 128(2):415 423.22. Helgason A, Hrafnkelsson B, Gulcher JR, War d R, Stefnsson K (2003) A populationwide coalescent analysis of Icelandic matrilineal and patrilineal genealogies: Evidence for a faster evolutionary rate of mtDNA lineages than Y chromosomes. Am J Hum Genet 72(6): 1370 1388. 23. Amster G, Sella G (2015) Life history effects on the molecular clock of autosomes and sex chromosomes. Proc Natl Acad Sci USA 113(6):1588 1593.",
+    "sistent with a maternal imprinting effect in familiesfrom France [18], the USA[10, 18, 21] (Figure 2; Table3) and Canada [27]. However, in a large family dataset from the UK, and in smaller data sets fromDenmark and Sardinia, the transmission of VNTRsusceptibility alleles is more pronounced frommothersthanfromfathers,andnowsignicantlysoinUK families (Figure 2; Table 3). Comparison of theresults from the USAwith those from the UK suggestthat unexplained inter-population differences in thisparent-of-origin",
+    "started with the largest matrilineage and worked down the list. Theparticipants selected for mtDNA sequencing were selected inde-pendent of their cognitive or dementia status. 274 matrilineageswere represented by this dataset. As a result, the sequencedmitochondrial genomes also represent as many different majormitochondrial haplogroups and clusters as possible (Table 1).Selection was made blind to case-control status. 287 samples weresent to Family Tree DNA (www.familytreedna.com) for Sangersequencing of"
+  ],
+  "answer": "Genetic tracing is often matrilineal rather than patrilineal primarily because mitochondrial DNA (mtDNA) is inherited solely from the maternal lineage. Unlike the nuclear genome, which requires contributions from both parents, mtDNA is passed down from mother to offspring without any admixture from the father [4], [5]. This uniparental inheritance of mtDNA simplifies the tracing of maternal lineage as it remains relatively unchanged through generations, making it a reliable marker for genetic studies [4], [5]. Additionally, humans have developed complex mechanisms to ensure this uniparental inheritance of mtDNA, which further supports its use in genetic tracing [4].",
+  "question": "Why is genetic tracing matrilineal rather than patrilineal?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_13 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_13
new file mode 100644
index 0000000..4263d86
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_13
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - eQTL analysis in mice and rats.pdf",
+    "2015 - Genetic Control of Survival and Weight Loss during Pneumonic Burk.pdf",
+    "2015 -Emery- Genetic Control of Survival and Weight Loss during Pneumonic Burk.pdf",
+    "2005 - quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2005 -Broadkin- quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2006 - From_gene_to_behavior_and_back_again_new.pdf",
+    "2009 - Experimental_Evolution.pdf",
+    "2009 - Garland_and_Rose_Experimental_Evolution.pdf",
+    "2005 - quantitative-trait-analysis-in-the-investigation-of-function-and.pdf",
+    "2012 - Systems genetic analysis of the effects of iron deficiency in mouse brain.pdf"
+  ],
+  "extraction_id": [
+    "71981bfb-284e-50ad-854e-2055c07f77a7",
+    "615ee0cd-5960-57e5-b4e6-56e4b8020a1b",
+    "268a23e8-f528-5b59-89f2-188331e0a03c",
+    "9de93371-6239-53c2-b42c-71f615a0614b",
+    "0a5c759e-8dab-55f1-ac59-e8211ec683b8",
+    "64c0287d-aeea-52eb-a074-e9591c5593ae",
+    "8ee78018-b998-590c-99ab-788a447ede81",
+    "cbce50ea-be78-5d54-beb1-849222c5bfdd",
+    "0a895880-91c0-5079-b258-73926b38430f",
+    "6ab990b0-4f9c-5be3-ab79-9ca6835271fa"
+  ],
+  "document_id": [
+    "8d67ea90-f7b1-5bb8-937c-4a9eceddff43",
+    "ae1025b0-1410-51ae-9be2-26fa2e9d5808",
+    "a9aceace-bf48-5472-b54c-59a458a84c62",
+    "0dc730ba-4ff4-52aa-a988-71075113c416",
+    "e6027e7f-aec0-5e76-8aff-96b36389e701",
+    "7a088b36-11b7-5379-bfe5-ce571e11de07",
+    "34821353-1b74-5ee2-ac39-66dd46f145bf",
+    "496faa7f-9623-5ab7-9816-7c3755abb3aa",
+    "dac1c73c-0b5f-5a54-bb12-7e8b654009c0",
+    "99fc80f0-f3c3-5766-a604-921552bb3298"
+  ],
+  "id": [
+    "chatcmpl-ADZKK0KDNVDBm3vRB6dGuJYB5JlVa",
+    "73540700-b5cf-5838-852b-b281ca086140",
+    "374c456a-d1db-5b4a-8713-97abe4162d77",
+    "b9d52798-0235-5018-bccd-560565d16cc3",
+    "c8f17022-aeae-5242-9082-d6d1eee4c4bf",
+    "1b2de424-be9f-572d-bd62-dc2ecd92192b",
+    "fef212bc-631b-591d-b8e3-d1523da0507d",
+    "f72795a1-66c3-5a98-84bc-b085e8008073",
+    "31a32dc5-81ac-52ba-a463-c61e293f21e5",
+    "b660d882-1cb0-5150-ae76-8eb3ccb88a58",
+    "7ef9df1d-b21a-597a-9e74-6eace5d0c33c"
+  ],
+  "contexts": [
+    "While most of the Y chromosome does not undergo recombination, the recombination rate of the X chromosomeis slower than that of the autosomes. This has important consequences on the detection of significant QTLs. For a comprehensive view of these issues, see(43). 9.Probe hybridization artifacts When several probes are available for the same gene, it is not uncommon to observe a difference in the mapping results",
+    "8 QTL Mapping     Allelic variation exists among natural populations and inbred strains, and this is  reflective of the segregation of quantitative tr ait loci (QTLs) [96]. QTLs are stretches of  DNA that are closely linked to genes that underlie a phenotype of interest. QTL analysis has been proven to be an invaluable tool to  help unravel heritable traits, by enabling  researchers to map different quantitative traits back to the genomic location involved in the regulation of these phenotypes.",
+    "8 QTL Mapping     Allelic variation exists among natural populations and inbred strains, and this is  reflective of the segregation of quantitative tr ait loci (QTLs) [96]. QTLs are stretches of  DNA that are closely linked to genes that underlie a phenotype of interest. QTL analysis has been proven to be an invaluable tool to  help unravel heritable traits, by enabling  researchers to map different quantitative traits back to the genomic location involved in the regulation of these phenotypes.",
+    "genes underlying QTLs in animals and plants (see for example Shirley et al 2004,Korstanje & Paigen 2002, Fridman et al 2004). I should also point out, though, that even in a single QTL region isolated in a congenic strain, it is possible that there is more than one allele that aects the phenotype. So, you have a fair pointabout the challenges and complexities of QTL analysis. Koolhaas: There are dierent questions underlying both approaches. The QTL",
+    "genes underlying QTLs in animals and plants (see for example Shirley et al 2004,Korstanje & Paigen 2002, Fridman et al 2004). I should also point out, though, that even in a single QTL region isolated in a congenic strain, it is possible that there is more than one allele that aects the phenotype. So, you have a fair pointabout the challenges and complexities of QTL analysis. Koolhaas: There are dierent questions underlying both approaches. The QTL",
+    "The basic  pr emise  of QTL  an alysis  is simple  (Ph illips  and Belknap, 2002 ) . First,  one must  meas  ure a speci  c phen  otype  within  a popul  ation. Next, the population must be genotyped at a hundred or more marker loci186 Boehm II et al.",
+    "through analysis of line crosses, quantitative trait loci (QTL) mapping, and verification of candidate genes with quantitative complementation tests or genetic engineering (e.g.,McGuire and Tully 1987; Chandra et al. 2001; Dierick and Greenspan 2006; Edwardset al. 2006). They can also be used to study the underlying physiological, neural, andmolecular mechanisms of the differences in behavior between selected and controllines, or between divergently selected lines.",
+    "through analysis of line crosses, quantitative trait loci (QTL) mapping, and verification of candidate genes with quantitative complementation tests or genetic engineering (e.g.,McGuire and Tully 1987; Chandra et al. 2001; Dierick and Greenspan 2006; Edwardset al. 2006). They can also be used to study the underlying physiological, neural, andmolecular mechanisms of the differences in behavior between selected and controllines, or between divergently selected lines.",
+    "genetic background. Gene identification of QTL should be distinguished from identification of the quanti- tative trait nucleotide (QTN). The latter is a daunting task, since SNPs are so frequent. Final proof for a QTN in mice would require placing a genomic segment containing theputative QTN from a donor mouse strain on the background of another strain using homologous recombination and reproducing the phenotype of the donor strain.",
+    "because these strains have been genotyped at more than 14,000 markers, including single nucleotide polymorphisms (SNP). Hundreds of genes may lie within a QTL interval, so identifying the underlying genes requires complementary methods. One method is to use BXD gene expression data (a public resource at www.genenetwork.org) to screen for genes within the QTL interval whose expression correlates with the trait of interest [23]."
+  ],
+  "answer": "To differentiate between X and Y DNA with respect to DNA tracing and determining quantitative trait loci (QTLs), one must consider the unique characteristics of each chromosome:  1. **Recombination Rates**:    - The Y chromosome largely does not undergo recombination, which means it is passed down relatively unchanged from father to son. This makes it easier to trace paternal lineage but limits its utility in QTL mapping due to the lack of recombination [1].    - The X chromosome, on the other hand, does undergo recombination, although at a slower rate compared to autosomes. This recombination can complicate tracing but is crucial for detecting significant QTLs [1].  2. **QTL Mapping**:    - QTLs are stretches of DNA closely linked to genes that influence a phenotype of interest. The detection and mapping of QTLs on the X chromosome can be more complex due to its recombination, but this recombination is necessary for identifying the genomic locations involved in regulating phenotypes [2].    - The Y chromosome's lack of recombination means it is less useful for QTL mapping, as the identification of QTLs relies on the recombination events that help pinpoint the genomic regions associated with specific traits [1].  In summary, the primary difference lies in the recombination behavior of the X and Y chromosomes, which affects their utility in DNA tracing and QTL determination. The X chromosome's recombination, although slower, is essential for QTL mapping, whereas the Y chromosome's lack of recombination makes it more suitable for tracing paternal lineage but less useful for QTL analysis.",
+  "question": "How does one tell the difference between X and Y DNA, with respect to DNA tracing and determining QTLs?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_14 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_14
new file mode 100644
index 0000000..eb9eb37
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_14
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2016 - A genetic screen identifies hypothalamic Fgf15 as a regulator of glucagon secretion.pdf",
+    "2015 - Systems genetic analysis of hippocampal neuroanatomy and spatial learning in mice.pdf",
+    "2007 - Integration of mouse phenome data resources.pdf",
+    "2016 - Genetic Regulation of Gelsolin in Lung in Mouse Model and its Potential.pdf",
+    "2005 -Integrated gene expression profiling and linkage analysis in the rat.pdf",
+    "2019 - The expanded BXD family of mice A cohort for experimental systems genetics and precision medicine.pdf",
+    "2018 - Molecular Brain Adaptations to Ethanol_ Role of Glycogen Synthase (2).pdf",
+    "2008 -Han- Comparing Quantitative Trait Loci.pdf",
+    "2008 - Comparing Quantitative Trait Loci.pdf",
+    "2008 - Towards systems genetic analyses in barley Integration of phenotypic, expression and genotype data into GeneNetwork.pdf"
+  ],
+  "extraction_id": [
+    "7eae53fa-ac5e-5cf4-807c-5d13dffdcf83",
+    "69504f91-c34d-5555-a05a-ac485356cec6",
+    "6ba5dba3-6135-5545-bec9-eee2e1465e7b",
+    "311be2a2-4428-5887-8ed2-35875eac9fcb",
+    "80a6f32f-a473-58ba-98ce-30100f5cc913",
+    "22772f7f-a42d-5438-a910-9e26c2916be2",
+    "1047bf10-3878-5b70-8bb2-c0249f2a9c53",
+    "e0bc4e49-6d6f-5b60-b7bc-18fd622629a8",
+    "476c90a3-1613-5e45-81b4-358519368bda",
+    "a6c480d1-b384-5c6f-b21b-94fe0b3b0f4d"
+  ],
+  "document_id": [
+    "288adb9b-a547-5e61-8593-1b2ab36271d3",
+    "8708ead5-20bc-5d41-82db-61a807eb3f90",
+    "08a3ce6e-947b-5ee9-b723-946807cf7d23",
+    "ec8452c0-1c16-54e6-9b9f-3e741a8c7340",
+    "7b3a7517-2967-5693-b4e8-8423a9fa432b",
+    "8df14e3b-644f-5a18-94a6-5ff5a1eae053",
+    "cc2690a9-5a87-5f09-87d5-115a6a6b8349",
+    "e6904cbd-8265-5e40-8978-d461ee6e151a",
+    "bfbddb84-c0e5-5d74-8e2d-9e54e75e8c49",
+    "8513abbe-65ed-5f35-9f86-ba93cfc5a194"
+  ],
+  "id": [
+    "chatcmpl-ADZKSZUCeTbC5g92NfqE6Fmp3TXXx",
+    "a2ffc857-6d79-5889-8344-cae8f1ca5e32",
+    "1e23f2e3-f4b1-5195-9061-5e525a13fb32",
+    "6c1e5cb1-ab19-5246-859d-a2f58d48232a",
+    "51757b6b-0492-5077-ba69-90a2ddf3da9d",
+    "dae9312b-c464-5fb7-bbc1-06ba2998e462",
+    "0b3d48d1-f253-508c-9a9e-5060e02d54a6",
+    "d261c68c-c253-52c9-8e27-f76fb8d0b4f8",
+    "9fbea8b6-25ad-5da9-bc9a-988784e33f0b",
+    "bd69b879-f1fe-57ee-8b36-b621708bdcc3",
+    "969d6ade-dc87-5f19-bd57-3f58882f11e8"
+  ],
+  "contexts": [
+    "QTL Mapping GeneNetwork ( www.genenetwork.org ) variants data set comprising about",
+    "Bioinformatics All of the genetic analyses were carried out in GeneNetwork, whichis an open source bioinformatics resource for systems genetics thatexists as both a repository for genetic, genomic and phenotypicdata together with a suite of statistical programs for data analy-sis that includes mapping and evaluating QTLs, examining pheno-type/genotype correlations and building interaction networks. QTL mapping The QTL mapping module of GeneNetwork was used to identify",
+    "the database is that each data collection is associated with a protocol which describes how  the data were generated. The project also provides online analysis tools to allow  identification of correlations within its data set. GeneNetwork ( http://www.genenetwork.org ), encompassing WebQTL, is a database of  genotypes and complex phenotypes ranging from gene expression to behaviour in standard  inbred strains, and six panels of mouse recombinant inbred strains including the two largest",
+    "QTL/interval analysis  QTL  mapping  was conducted  using  publically  available  software   on GeneNetwork  (http://www .genenetwork .org/webqtl /main .py).  One  important  feature  of the GeneNetwork  is WebQTL , which  is the  leading  GeneNetwork  module , and has been  optimized  for on-line  analysis  of traits  that are controlled  by combinations  of allelic  variants   and environmental  factors  [15]. A simple  graphical  user interface",
+    "WebQTL is the primary module in the Gene- Network online resource (www.genenetwork.org),and provides a powerful environment to analyzetraits controlled by genetic variants (Chesler et al.2004; Wang et al. 2003). It includes data from manypermanent genetic reference populations, including the HXB rat strains, and allows for phenotypic traits,",
+    "67. As described above, loci are identified in GeneNet work by the computation of a  likelihood ratio statistic score and significance  was determined using at least 5,000  permutations of the phenotype data.   Updated QTL mapping methods , such as  R/qtl 2 66,146, Multiple QTL mapping  64,  GEMMA  156 and pyLMM  63, have been implimented on t he GeneNetwork2 site 46.",
+    "genetic mapping, and correlation of quantitative traits such as gene expression data and behavioral parameters (Wang  et al, 2003) . GeneNetwork employs  genotype data from 3809 markers, selected based on their being informative (i.e., different between progenitor strains). GeneNetwork outputs peak likelihood ratio statistic (LRS) locations for each trait, whic h can be directly converted to",
+    "tool for combined visualization and exploration of geneexpression data and QTL. The methodology developedin this work is complementary to the analyses that canbe performed on the GeneNetwork website (WebQTL,http://www.genenetwork.org/ ), which allows assessment of the relationship between gene expressions and QTL inrecombinant in bred mice [ 3]. Comparing QTL and microarray data is not completely",
+    "tool for combined visualization and exploration of geneexpression data and QTL. The methodology developedin this work is complementary to the analyses that canbe performed on the GeneNetwork website (WebQTL,http://www.genenetwork.org/ ), which allows assessment of the relationship between gene expressions and QTL inrecombinant in bred mice [ 3]. Comparing QTL and microarray data is not completely",
+    "the database entries. Once the resulting record set of the query is returned, it can be further restricted by selecting relevant records based on attached annotations before for- warding it for further analysis. To map genetic loci associated with mRNA abundance or trait phenotypes, any one of the three QTL mapping func- tions currently employed by GeneNetwork's WebQTL module can be used. These are 1. interval mapping, 2. sin- gle-marker regression, or 3. composite mapping [29,30]."
+  ],
+  "answer": "For QTL mapping analysis in GeneNetwork2, datasets that are useful include genotypes and complex phenotypes ranging from gene expression to behavior in standard inbred strains and panels of mouse recombinant inbred strains [3]. Additionally, data from permanent genetic reference populations, such as the HXB rat strains, are also valuable [5]. These datasets allow for the identification of correlations within the data and the mapping of genetic loci associated with mRNA abundance or trait phenotypes [10].",
+  "question": "what type of dataset is useful for qtl mapping analysis in genenetwork2?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_15 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_15
new file mode 100644
index 0000000..79f613c
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_15
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2005 -Dipetrillo- Bioinformatics toolbox QTL.pdf",
+    "2005 - Bioinformatics toolbox for narrowing rodent quantitative trait loci .pdf",
+    "2020 - A Multi-Omics Perspective of Quantitative Trait Loci in Precision Medicine.pdf",
+    "2016 - Genotyping by sequencing for identification and mapping of QTLs for bioenergy-related traits in sweet sorghum.pdf",
+    "2005 -Dipetrillo- Bioinformatics toolbox QTL.pdf",
+    "2005 - Bioinformatics toolbox for narrowing rodent quantitative trait loci .pdf",
+    "2016 - Genetic Regulation of Gelsolin in Lung in Mouse Model and its Potential.pdf",
+    "2009 - Detection and interpretation of expression quantitative trait loci (eQTL).pdf",
+    "2020 - A Multi-Omics Perspective of Quantitative Trait Loci in Precision Medicine.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf"
+  ],
+  "extraction_id": [
+    "63fcced2-fd9b-5b8c-917e-8a5502f89624",
+    "ede4bc5e-f495-5c65-b2e6-a5dc0625b0d0",
+    "03e2ebd6-ce89-551c-ba81-59a4ded02515",
+    "ea640aeb-71cc-578d-8ad3-6940f2b892da",
+    "294efef3-6516-5c74-8cc5-bc8401f6602b",
+    "4cf47fab-c25f-52a4-953b-3c3508a26274",
+    "311be2a2-4428-5887-8ed2-35875eac9fcb",
+    "2b670f5c-5b0c-5d8f-b236-2cbff81eff5a",
+    "6f44c216-c9a1-582e-8081-d6ad912369db",
+    "52bb366e-161f-51fd-a5a2-bef21f1b4c01"
+  ],
+  "document_id": [
+    "9b089457-5804-594a-99ea-e716b65c216c",
+    "5d87aefe-dee5-5f25-8b46-d87b24907dcc",
+    "8503b166-b917-5efb-a356-5ba371504cc1",
+    "d6da662e-cb6e-5628-8a42-5aca1b978447",
+    "9b089457-5804-594a-99ea-e716b65c216c",
+    "5d87aefe-dee5-5f25-8b46-d87b24907dcc",
+    "ec8452c0-1c16-54e6-9b9f-3e741a8c7340",
+    "ef974b09-4ea2-5382-85e5-c2169f440fda",
+    "8503b166-b917-5efb-a356-5ba371504cc1",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101"
+  ],
+  "id": [
+    "chatcmpl-ADZKWNKiSvqnmJLiG5DiGZqUhCrfq",
+    "7a9f6af0-22c9-5bd7-a443-f0b0111551fa",
+    "b89fda54-1dd8-5033-9caa-c8e6079d4e28",
+    "db4d7722-ff83-54a4-9fb6-23d331ead769",
+    "5604e763-06b5-5528-be49-9003bf547ae2",
+    "7019c554-cbae-528e-8207-b8575d99daf4",
+    "3fe2119e-e576-5608-91e1-2a010b91515c",
+    "51757b6b-0492-5077-ba69-90a2ddf3da9d",
+    "7cd326b3-1669-55f1-b4ce-376b5159a6fb",
+    "ae35202f-70ed-5fb8-a075-ce1e63616fb2",
+    "0049fb65-142a-54a1-8ab5-2d747bc521a0"
+  ],
+  "contexts": [
+    "rodent QTLs. Here we discuss each tool, illustrate itsapplication and generate a bioinformatics strategy fornarrowing QTLs. Combining these bioinformatics toolswith classical experimental methods should accelerateQTL gene identication. Introduction Quantitative trait locus (QTL) analysis is a method to localize chromosomal regions harboring genetic variants that affect a continuously distributed, polygenic phenotype(including many common diseases) [1]. It is particularly",
+    "rodent QTLs. Here we discuss each tool, illustrate itsapplication and generate a bioinformatics strategy fornarrowing QTLs. Combining these bioinformatics toolswith classical experimental methods should accelerateQTL gene identication. Introduction Quantitative trait locus (QTL) analysis is a method to localize chromosomal regions harboring genetic variants that affect a continuously distributed, polygenic phenotype(including many common diseases) [1]. It is particularly",
+    "Table 2. Computational Approaches for Identi cation of QTLs Tools Link Programming languageRefs Linear models CPMAtranseqtl https://github.com/cotsapaslab/CPMAtranseqtl R/Python [ 176] eMap www.gnu.org/software/gsl/ R FastMap https://sourceforge.net/projects/fastmapunix/ JAVA [ 134] lme4qtl https://github.com/variani/lme4qtl R[ 175] Matrix eQTL www.bios.unc.edu/research/genomic_software/ Matrix_eQTLR/Matlab [ 133] Meta-eQTL https://haok01.u.hpc.mssm.edu/meta_eQTL/ R/C [ 177]",
+    "2012). Tools for QTL analysis have been de veloped and released for researchers such as  R/qtl, QTL cartographer, M apQTL, and WebQTL. Recently, Wang et al. (2012)  developed a free software for QTL mapping called QTL IciMapping which constructs genetic linkage maps and QTL analysis  by simple interval mapping and inclusive  composite interval mapping.  QTL IciMapping is available for segregating and inbred                PREVIEW",
+    "incorrect, the analysis can separate the QTL peak into twoTable 1. Summary of bioinformatics tools for dissecting rodent QTLs Bioinformatics tool Summary Resolution Comparative genomics Identies regions of chromosomal synteny in QTLs that are concordant across species1020 Mb Combined cross analysis Recodes genotype information from multiple crosses detecting a shared QTL into one susceptibility and one resistance genotype to combine the crosses in a singleQTL analysis1020 Mb Interval-specic haplotype",
+    "incorrect, the analysis can separate the QTL peak into twoTable 1. Summary of bioinformatics tools for dissecting rodent QTLs Bioinformatics tool Summary Resolution Comparative genomics Identies regions of chromosomal synteny in QTLs that are concordant across species1020 Mb Combined cross analysis Recodes genotype information from multiple crosses detecting a shared QTL into one susceptibility and one resistance genotype to combine the crosses in a singleQTL analysis1020 Mb Interval-specic haplotype",
+    "QTL/interval analysis  QTL  mapping  was conducted  using  publically  available  software   on GeneNetwork  (http://www .genenetwork .org/webqtl /main .py).  One  important  feature  of the GeneNetwork  is WebQTL , which  is the  leading  GeneNetwork  module , and has been  optimized  for on-line  analysis  of traits  that are controlled  by combinations  of allelic  variants   and environmental  factors  [15]. A simple  graphical  user interface",
+    "model selection approach for mapping multiple interacting QTL [376] and Plink, a library for association QTL mapping on single nu cleotide polymorphisms (SNP) in natural populations [277]. 3.2.3 Add new analysis tools xQTL workbench supports exible adding of more QTL analysis s oft- ware: any R-based, or command-line tool, can be plugged in. A ll anal- ysis results are uploaded, stored and tracked in the xQTL workbench database through an R-API. When new tools are added, they can b uild",
+    "717 730 14. Delaneau, O. et al. (2017) A complete tool set for molecular QTL discovery and analysis. Nat. Commun. 8, 1545215. Liu, B.H. (2017) Statistical Genomics: Linkage, Mapping, and QTL Analysis , CRC Press 16. Gibson, G. et al. (2015) Expression quantitative trait locus anal- ysis for translational medicine. Genome Med. 7, 1 14 17. Ritchie, M.D. et al. (2015) Methods of integrating data to uncover genotype-phenotype interactions. Nat. Rev. Genet. 16, 185 197",
+    "236 CH 10 TOOLS FOR STATISTICAL GENETICS Lastly, Bayesian methods allow the consideration of multiple QTLs, QTL positions and QTL strengths (Jansen, 1996; Satagopan et al. , 1996; Uimari et al. , 1996; Sillanpaa and Arjas, 1998, Borevitz et al. , 2002). Multimapper (Sillanpaa, 1998), for example, allows the automatic building of models of multiple QTLs within the same linkage group. It is designed to work as a companion program to QTL Cartographer (Basten"
+  ],
+  "answer": "The bioinformatics tools for QTLs analysis include:  1. **R/qtl** - A comprehensive tool for QTL mapping [4]. 2. **QTL Cartographer** - Used for QTL mapping and analysis [4]. 3. **MapQTL** - Another tool for QTL mapping [4]. 4. **WebQTL** - An online tool optimized for the analysis of traits controlled by combinations of allelic variants and environmental factors [4], [7]. 5. **QTL IciMapping** - A free software for QTL mapping that constructs genetic linkage maps and performs QTL analysis by simple interval mapping and inclusive composite interval mapping [4]. 6. **CPMAtranseqtl** - Available on GitHub, supports R/Python [3]. 7. **eMap** - Available on GNU, supports R [3]. 8. **FastMap** - Available on SourceForge, supports JAVA [3]. 9. **lme4qtl** - Available on GitHub, supports R [3]. 10. **Matrix eQTL** - Available on the UNC website, supports R/Matlab [3]. 11. **Meta-eQTL** - Available on HPC MSSM, supports R/C [3]. 12. **Plink** - A library for association QTL mapping on single nucleotide polymorphisms (SNP) in natural populations [8]. 13. **Multimapper** - Allows the automatic building of models of multiple QTLs within the same linkage group and works as a companion program to QTL Cartographer [10].  These tools can be used in combination with classical experimental methods to accelerate QTL gene identification [1], [2].",
+  "question": "what are the bioinformatics tools for QTLs analysis?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_16 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_16
new file mode 100644
index 0000000..b345237
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_16
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - Identification of Quantitative Trait Loci in Alcoholism.pdf",
+    "2018 - Reduced complexity cross design for behavioral genetics.pdf",
+    "2005 - Genetics of body weight in the LXS recombinant inbred mouse strains.pdf",
+    "2006 - From_gene_to_behavior_and_back_again_new.pdf",
+    "2012 - Bioinformatics tools and database resources for systems genetics analysis in mice\u2014a short review and an evaluation of future needs.pdf",
+    "2012 - Bioinformatics tools and database resources for systems genetics analysis in mice\u2014a short review and an evaluation of future needs.pdf",
+    "2012 - Genetic regulation of adult hippocampal neurogenesis A systems genetics approach using BXD recombinant inbred mouse strains.pdf",
+    "2007 - Metabolic and genomic dissection of diabetes in the Cohen rat.pdf",
+    "2007 - Metabolic and genomic dissection of diabetes.pdf",
+    "2011 - Genetical genomics approaches for systems genetics.pdf"
+  ],
+  "extraction_id": [
+    "59e1cde3-dd67-55c0-aceb-0d4dbf22ed4d",
+    "d18c973d-30ee-5069-a101-b4d3000333eb",
+    "def0e506-3ca4-5a7f-8a4d-5968e2a36f1e",
+    "64c0287d-aeea-52eb-a074-e9591c5593ae",
+    "88873c88-94cd-5caf-b675-a99f0ae6235f",
+    "88873c88-94cd-5caf-b675-a99f0ae6235f",
+    "17184903-e412-5545-8dfc-c17e31f5201b",
+    "a20d5dd5-6dd1-54ab-8c52-647fdf644ae7",
+    "1aa37aaa-5635-57a5-b8d4-2dd9fa17d028",
+    "fb1b1f9d-81a6-59b2-b31c-80a5940d8b3f"
+  ],
+  "document_id": [
+    "11c67421-d1e1-5bde-bf97-3e313232fec7",
+    "b6797de4-6bdf-52ae-a848-d8fc4f048587",
+    "1a5be6d7-d1b8-5405-a0cb-696a5eb6a0f1",
+    "7a088b36-11b7-5379-bfe5-ce571e11de07",
+    "4bb4798b-3969-5448-ac4b-13c1b8506268",
+    "4bb4798b-3969-5448-ac4b-13c1b8506268",
+    "c54da858-9620-588e-8e41-76a960af2ff6",
+    "ce608956-7efb-5ce8-ab42-400075d012bb",
+    "5503f978-238f-59bc-ad3f-f500eb712aef",
+    "de78a01d-8d03-5afb-af5b-ce2ed2167766"
+  ],
+  "id": [
+    "chatcmpl-ADZKiurNCvLvQlfZEPvqlUva8Sekv",
+    "5db68dae-9dc1-5065-b61f-067ba20b6e19",
+    "e5fcabd8-0d42-5aa4-bebb-a355493e8ced",
+    "8efc851d-4fd4-5355-946a-4e183083eadd",
+    "fef212bc-631b-591d-b8e3-d1523da0507d",
+    "9dc3af1c-27a0-5527-b788-719c3ff01cd4",
+    "4940ec57-f3dc-55f7-9cfa-71f1e5b66287",
+    "280734af-e950-5339-b984-8718e98448ad",
+    "9ee9d05e-d3fb-5dd7-b1b5-9862c1894099",
+    "7e038f11-0794-5424-9465-eb0034442369",
+    "9a2b996d-7480-57e8-9c6a-da084c4be200"
+  ],
+  "contexts": [
+    "Methods   31  statistical language/software R (R DEVELOPMENT CORE TEAM 2008) . The core of R/qtl is a set  of functions that make use of the hidden Markov model (HMM) technology to calculate QTL  genotype probabilities, to simulate from the  joint genotype distribution and to calculate the  most likely sequence of underlying genotypes (all conditional on the observed marker data)  (BROMAN  et al.  2003) . R/qtl also calculates several functio ns that are useful for a quality",
+    "A variety of analytical methodologies are available in the R/qtl package, including, e.g., composite interval mapping or Haley-Knott regression (see Ref. 42for discussion). The scanone function in R/qtl is used to calculate log of the odds (LOD) scores. Per- mutation analysis (perm 1000) is used to establish the signi cance threshold for each phenotype ( P<.05). Additive and/or interactive covariates can be added to the model",
+    "WebQTL (Chesler et al. 2003; http://www.web- qtl.org/home.html), because each has some uniquecapabilities. R/qtl is an interactive environment for mapping QTLs in experimental crosses, implemented as anadd-on package for the freely available statisticallanguage/software R. Empirical significance valuesare calculated by permutation tests by comparing the peak likelihood ratio statistic (LRS) obtained from 1000 permutations (Churchill and Doerge1994). The permutation test results of highly sig-",
+    "The basic  pr emise  of QTL  an alysis  is simple  (Ph illips  and Belknap, 2002 ) . First,  one must  meas  ure a speci  c phen  otype  within  a popul  ation. Next, the population must be genotyped at a hundred or more marker loci186 Boehm II et al.",
+    "analyses on whole assays of (molecular) phenotypesas a batch. This enables genetical genomics studieswithout waiting times. TIQS is particularly strong inusing a cloud for large scale computing while xQTL uses pbs based traditional clusters and is more developed for data management and definitionof new analyses, so the desire is to work together.Both systems use R as the back-end language for dataanalysis in all platforms, which will enable transfer of analysis protocols between experiments and insti-",
+    "tional protocols to analyse all expression, proteomicsand metabolomics QTLs on marker maps of everincreasing density. These should include web accesstools for both experts and non-experts in sophisti-cated statistics analysis and high performance computing. The interactive QTL System (TIQS) (http://eqtl .berlios.de) is a web application that guides its usersthrough the analysis steps needed. It maximizes the distribution of computational effort (supporting trad-",
+    "four commonly used methods for doing a linkage analysis, namely; regression method, likelihood method, variance component method and Bayesian method. For statistical purpose, to check significant thresholds, either permutation test or Bayesian factors are used and for confidence interval check, bootstrapping is the preferred method.  For our study, we use WebQTL for QTL mapping. WebQTL (http://webqtl.org) uses interval mapping, to estimate the position of QTLs across a chromosome (Wang et al., 2003,",
+    "MultiQTL software package, version 2.5 (www.multiqtl.com), aspreviously described in detail (37). In brief, for initial analysis, weused by default an unrestricted model. When the results suggested thepresence of a QTL, we attempted to t the simplest and statisticallyjustied model (dominant, recessive, or additive effect) by comparingit with the nonrestricted model and replacing it if the difference wasnonsignicant. When applicable, we utilized the single-trait, multi-trait, and multienvironment analyses",
+    "MultiQTL software package, version 2.5 (www.multiqtl.com), aspreviously described in detail (37). In brief, for initial analysis, weused by default an unrestricted model. When the results suggested thepresence of a QTL, we attempted to t the simplest and statisticallyjustied model (dominant, recessive, or additive effect) by comparingit with the nonrestricted model and replacing it if the difference wasnonsignicant. When applicable, we utilized the single-trait, multi-trait, and multienvironment analyses",
+    "R/QTL [35] is an R package which includes many func tions for mapping, including  an algorithm to infer missing genotype data using H idden Markov Models. Gene-  Network (www.genenetwork.org [11]) also offers eQTL  analysis for user uploaded  data, one trait at a time, and genome-wide analysis  tools for a number of published  datasets.    4. Alternative Illumina data pre-processing  Compared with Affymetrix for example, Illumina is a  relatively new technology and"
+  ],
+  "answer": "The statistical approaches for QTLs (Quantitative Trait Loci) analysis include several methodologies and techniques. Here are some of the key approaches mentioned in the provided context:  1. **Hidden Markov Model (HMM)**: This approach is used to calculate QTL genotype probabilities, simulate from the joint genotype distribution, and calculate the most likely sequence of underlying genotypes, all conditional on the observed marker data [1].  2. **Composite Interval Mapping (CIM)**: This method is available in the R/qtl package and is used for QTL mapping [2].  3. **Haley-Knott Regression**: Another method available in the R/qtl package, used for QTL analysis [2].  4. **Log of the Odds (LOD) Scores Calculation**: The scanone function in R/qtl is used to calculate LOD scores, which are essential for identifying QTLs [2].  5. **Permutation Analysis**: This technique is used to establish the significance threshold for each phenotype, ensuring that the identified QTLs are statistically significant [2].  6. **Additive and Interactive Covariates**: These can be added to the model to account for additional sources of variation in the data [2].  7. **Regression Method**: One of the commonly used methods for linkage analysis in QTL studies [7].  8. **Likelihood Method**: Another method used for linkage analysis in QTL studies [7].  9. **Variance Component Method**: This method is also used for linkage analysis in QTL studies [7].  10. **Bayesian Method**: A method used for linkage analysis, which can also be used to check significant thresholds using Bayesian factors [7].  11. **Bootstrapping**: This method is preferred for checking confidence intervals in QTL analysis [7].  These approaches provide a comprehensive toolkit for conducting QTL analysis, allowing researchers to identify and validate QTLs effectively.",
+  "question": "what are the statistical approaches for qtls analysis?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_17 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_17
new file mode 100644
index 0000000..b31471c
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_17
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2017 - INTEGRATIVE ANALYSIS OF GENETIC, GENOMIC AND PHENOTYPIC DATA FOR ETHANOL BEHAVIORS A NETWORK-BASED PIPELINE FOR IDENTIFYING MECHANISMS AND POTENTIAL DRUG TARGETS.pdf",
+    "2008 - The Environmental Genome Project Reference Polymorphisms for Drug Metabolism Genes and Genome Wide Association Studies.pdf",
+    "2022 - Using Recurrent Neural Networks for Predicting Type-2 Diabetes from Genomic and Tabular Data.pdf",
+    "2022 - Using Recurrent Neural Networks for Predicting Type-2 Diabetes from Genomic and Tabular Data.pdf",
+    "2013 - Genome-Wide Contribution of Genotype by Environment Interaction.pdf",
+    "2009 - Processing Large-Scale, High-Dimension Genetic and Gene Expression Data.pdf",
+    "2019 - Beyond Genome-wide Significance Integrative Approaches to the Interpretation and Extension of GWAS Findings for Alcohol Use Disorder.pdf",
+    "2016- Gene-Based Genome-Wide Association.pdf",
+    "2008 - The Environmental Genome Project Reference Polymorphisms for Drug Metabolism Genes and Genome Wide Association Studies.pdf",
+    "2015 - Genetic associations at 53 loci highlight cell types and biological pathways relevant for kidney function.pdf"
+  ],
+  "extraction_id": [
+    "cc02b251-60c5-571f-9ff8-ef64c61eee5a",
+    "0f19f50f-ee04-5e99-8547-8a7e71a1dd9c",
+    "200d489e-301f-50bc-9870-260894c8fc41",
+    "6b4157fa-dcf0-5b70-b508-38ffb5fcda8d",
+    "5ade83ec-421a-58be-ac06-c9076076483c",
+    "1d401588-b6dc-532f-8194-4667a7d31153",
+    "bca29f20-2764-5d16-888e-3af671c9d8b0",
+    "db605926-64e1-5fc5-ac90-22f0f33b2a50",
+    "1b1aabee-8555-5ba8-b147-7f250fdcbc6b",
+    "0127b2c2-37b8-580d-b974-a2e3c69015ab"
+  ],
+  "document_id": [
+    "0e2a1075-1e04-5097-b87f-3ca41d55e025",
+    "15e4c746-42a2-598b-992f-dfbf468865ed",
+    "be0e50e0-3de8-53c5-8126-a0b618647f80",
+    "be0e50e0-3de8-53c5-8126-a0b618647f80",
+    "8c310d76-0a3b-574c-9859-859258870ee5",
+    "17264155-b665-59db-94cb-f4d67eac20fc",
+    "f59b3e10-a887-5708-b520-c5e8adb48dcd",
+    "8cb14287-762d-5366-8ad9-3d638f02d0d6",
+    "15e4c746-42a2-598b-992f-dfbf468865ed",
+    "ea82333b-b64c-5416-9843-2e3ffeb1902a"
+  ],
+  "id": [
+    "chatcmpl-ADZKtYz4STZ5YGDkrchFPqAthSpVB",
+    "1b947a05-d204-5524-b7a6-4ddce62449f8",
+    "47097a55-da1c-5802-8ee7-549e16db2927",
+    "1dbbef8d-ece1-534d-a3f0-0cc46024cae6",
+    "0b7e9c6d-60e3-5d66-b23f-8222b327d91e",
+    "43aa64fe-556a-5938-a489-fff5aac6829d",
+    "6e7cd04d-d23a-5a7d-a0cd-7958608010f2",
+    "3a9e43ef-294d-5b1b-b4f9-62fa70064045",
+    "b4a50b95-3a61-5495-b8b2-c18f8edcaa8f",
+    "5e4b2bf5-f842-5c20-8031-48a29fd3d25a",
+    "619bcf7e-2724-571a-ba3c-4214ff014f21"
+  ],
+  "contexts": [
+    "1. Formatting genome wide association study (GWAS) data . For this step, a  human GWAS results file is needed that contains SNP names and raw p- values for the association of each SNP with a trait of interest. Because the  nodes of the dmGWAS network will represent genes, as opposed to SNPs,  gene-wise p-values need to be calculated from the raw SNP p-values. This  can be accomplished by using programs like VEGAS2 (Versatile Gene- Based Association Study) [ 10] or KGG (Knowledge-based mining system",
+    "A general outline for GWAS is provided in Figure 2. These studies usually begin with thousands of individuals who are charact erized for the phenotype of interest using continuous measurements, or dichotomous classi fication as a case (affected) or control (unaffected). Statistical analysis, typically us ing linear or logistic regression, tests the association of each SNP against the phenotype (including relevant covariate variables) to",
+    "GWAS has also provided polygenic characteristics of diseases. Figure 1 presents a block  of GWAS in disease prediction. There are many steps  during a gene-set analysis. They are  shown below as Steps 1 through Step 6:  Step 1:  Preliminary genome-wide analysis and data preproces sing;  Step 2:  Identifying gene-set definitions whose patterns have  to be recognized;   Step 3:  Processing genomic data such as filtering and ident ifying gene patterns;",
+    "GWAS in disease prediction. There are many steps during a gene-set analysis. They are shown below as Steps 1 through Step 6: Step 1: Preliminary genome-wide analysis and data preprocessing; Step 2: Identifying gene-set denitions whose patterns have to be recognized; Step 3: Processing genomic data such as ltering and identifying gene patterns; Step 4: Identify gene set analysis models, such as identifying the statistical hypothesis; Step 5: Assessing the statistical magnitude;",
+    "include: 1) generate bed, bimand fam files for GWAS genotype data using PLINK; 2) generategrm.gz and grm.id files using make-grm; 3) prepare a",
+    "7 Constructing Gene Networks to Enhance GWAS and GOGE Results As discussed, generating a GOGE data set and performing a rst-pass analysis on this scale of data is a major undertaking. The identication of or other DNA markersthat associate with the expression of one or more genes is a primary goal of a GOGE study. However, if analysis of GOGE data stopped at the identication of SNPs that associate with expression, the true v alue of these data would not be realized.",
+    "Aggregating GWAS data into biological units GWAS data can be further combined into biological units using gene and network-based  approaches. Gene-based approaches There is a high multiple testing burden in the context of a GWAS. Gene-based approaches,  which aggregate across summary statistics derived from association analyses of multiple loci  to derive p-values for association at the level of the gene, developed as one way to reduce",
+    "Steps involved inthegene-based association testwere described asbelow: 1)Generating intermediate datasets which integrate original GWAS Pvalues, rsID, position and chromo- some column foreach SNP. Atotal of6,559,815 European-specific and 5,351,262 Asian-spe- cific autosomal SNPs were used forsubsequent analysis after excluding theSNPs that could notberecognized byKGG and that located insexchromosomes (XorY);2)Defining asetof",
+    "248 M. J. RIEDER ET AL. Figure 2 An overview of GWAS. Samples with  a phenotype(s) or trait(s) of interest are identified; typically, thousands of samples are required to achieve appropri ate statistical power. Large-scale genotyping is carried out using commercially available chips (Affymetrix or  Illumina). P-values are generated from the associa- tion between the phenotype and genotype for each SNP tested. Highly associated SNPs will typically cluster",
+    "2006). 40. Welter, D. et al. The NHGRI GWAS Catalog, a curated resource of SNP-trait associations. Nucleic Acids Res. 42,D1001D1006 (2014).41. Wang, X. et al. Comparing methods for performing trans-ethnic meta-analysis of genome-wide association studies. Hum. Mol. Genet. 22,23032311 (2013). 42. Purcell, S. et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. 81,559575 (2007)."
+  ],
+  "answer": "### How-to Guide for GWAS Analysis  Genome-Wide Association Studies (GWAS) are a powerful tool for identifying genetic variants associated with traits or diseases. Below is a step-by-step guide to performing a GWAS analysis:  #### Step 1: Preliminary Genome-Wide Analysis and Data Preprocessing 1. **Sample Collection**: Collect thousands of samples characterized for the phenotype of interest. This can be done using continuous measurements or dichotomous classification as a case (affected) or control (unaffected) [2]. 2. **Genotyping**: Perform large-scale genotyping using commercially available chips such as Affymetrix or Illumina [9]. 3. **Data Formatting**: Prepare the GWAS genotype data by generating bed, bim, and fam files using tools like PLINK [5].  #### Step 2: Identifying Gene-Set Definitions 1. **Gene-Set Identification**: Identify gene-set definitions whose patterns need to be recognized [4].  #### Step 3: Processing Genomic Data 1. **Filtering and Identifying Gene Patterns**: Process the genomic data by filtering and identifying gene patterns [4]. 2. **Intermediate Datasets**: Generate intermediate datasets that integrate original GWAS p-values, rsID, position, and chromosome column for each SNP [8].  #### Step 4: Statistical Analysis 1. **Association Testing**: Perform statistical analysis, typically using linear or logistic regression, to test the association of each SNP against the phenotype, including relevant covariate variables [2]. 2. **Gene-Wise P-Values**: Calculate gene-wise p-values from the raw SNP p-values using programs like VEGAS2 or KGG [1].  #### Step 5: Aggregating GWAS Data 1. **Gene-Based Approaches**: Aggregate GWAS data into biological units using gene-based approaches to reduce the multiple testing burden. This involves deriving p-values for association at the level of the gene [7].  #### Step 6: Assessing Statistical Magnitude 1. **Statistical Hypothesis**: Identify gene set analysis models and assess the statistical magnitude of the associations [4].  #### Step 7: Constructing Gene Networks 1. **Enhancing GWAS Results**: Construct gene networks to enhance GWAS results. This involves identifying DNA markers that associate with the expression of one or more genes [6].  By following these steps, you can systematically perform a GWAS analysis to identify genetic variants associated with your trait or disease of interest.",
+  "question": "Create a how-to guide for GWAS analysis?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_18 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_18
new file mode 100644
index 0000000..cfe846d
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_18
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2020 - Precision and Personalized Medicine How Genomic.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2015 -Pandey- Functional Analysis of Genomic Variation and Impact on Molecular.pdf"
+  ],
+  "extraction_id": [
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "de09f30d-e9ba-5379-8c7a-85b2cd2ed6c8",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "c91e328e-4a01-5952-85b8-d7b5b47237c5",
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "c14d1c74-a14a-5037-8d3f-f32a60faa9a5",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "f4e989e5-c3d0-5d5c-b8c3-95894a14630b"
+  ],
+  "document_id": [
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "cd11028a-933b-52a0-9534-c173323056ef",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "93381e23-494b-5bc2-9d09-ef315506601f"
+  ],
+  "id": [
+    "chatcmpl-ADZL3IjE1j3CPp4KvVU6JGEtG2NTv",
+    "98010acc-fd11-5d33-bced-626ef29f2896",
+    "13a6b6f9-4a9a-5eb9-ac79-d986d9e613f0",
+    "8477a774-dddb-5541-b8d7-d51a7e56b0af",
+    "d2540614-9397-5e3e-8b5f-ad328ca973b2",
+    "1e324977-2ca5-5062-8a09-7659d516e899",
+    "a2d9c614-903d-513a-ad88-5a40f3534988",
+    "97f2aa12-623b-53ec-9793-5834311a37dd",
+    "3e782f01-a06e-51b6-ac8a-0e0a56939d08",
+    "199e1929-dc7c-58d4-8c8d-1c931e658e9c",
+    "d1158643-3625-5855-a03d-eec4ac96eb4d"
+  ],
+  "contexts": [
+    "FURTHER INFORMATION 10X Genomics: http://www.10xgenomics.com 454 Sequencing: http://www.454.com Advances in Genome Biology and Technology (AGBT):  http://www.agbt.org BGISEQ500: http://seq500.com/en/portal/Sequencer.shtml Illumina: http://www.illumina.com Ion Torrent: https://www.thermofisher.com/us/en/home/ brands/ion-torrent.html Oxford Nanopore Technologies: https://www.nanoporetech. com Pacific Biosciences: http://www.pacb.com Personal Genome Project: http://www.personalgenomes.org",
+    "22. Karow, J. Qiagen launches GeneReader NGS System  atAMP; presents performance evaluation by broad.  GenomeWeb  [online], https:// www.genomeweb.com/ molecular-diagnostics/qiagen-launches-genereader- ngs-system-amp-presents-performance-evaluation   (4Nov 2015). 23. Smith,D.R. & McKernan,K. Methods of producing  and sequencing modified polynucleotides . US Patent  8058030 (2011). 24. Margulies,M. etal.  Genome sequencing in  microfabricated high-density picolitre reactors. Nature  437, 376380 (2005).",
+    "36. Sequencing, H.G. Finishing the euchromatic sequence of the human genome. Nature 2004 ,431, 931945. 37. Heather, J.M.; Chain, B. The sequence of sequencers: The history of sequencing DNA. Genomics 2016 ,107, 18. [CrossRef] 38. Rothberg, J.M.; Leamon, J.H. The development and impact of 454 sequencing. Nat. Biotechnol. 2008 ,26, 11171124. [CrossRef] [PubMed] 39. Shendure, J.; Ji, H. Next-generation DNA sequencing. Nat. Biotechnol. 2008 ,26, 11351145. [CrossRef] [PubMed]",
+    "sequencing. Genome Res. 20, 11651173 (2010). 64. English,A.C. etal.  Assessing structural variation in a  personal genome-towards a human reference diploid  genome. BMC Genomics 16, 286 (2015). 65. Carneiro,M.O. etal.  Pacific Biosciences sequencing  technology for genotyping and variation discovery in  human data. BMC Genomics 13, 375 (2012). 66. Quail,M.A. etal.  A tale of three next generation  sequencing platforms: comparison of Ion T orrent,  Pacific Biosciences and Illumina MiSeq sequencers.",
+    "sequencing. Bioinformatics 31, 20402042 (2015). 46. Qiagen.  Oncology insights enabled by knowledge base- guided panel design and the seamless workflow of the  GeneReader NGS system  Press Release. Qiagen   [online], http://www.genereaderngs.com/PROM-9192- 001_1100403_WP_GeneReader_NGS_0116_NA.pdf   (2016). 47. Forgetta,V. etal.  Sequencing of the Dutch elm disease  fungus genome using the Roche/454 GS-FLX Titanium  System in a comparison of multiple genomics core",
+    "for sequencing on existing short-read instrumentation,  after which data are split by barcode and reassembled  with the knowledge that fragments sharing barcodes Barcodes A series of known bases  addedto a template molecule  either through ligation or  amplification. After  sequencing, these barcodes  can be used to identify which  sample a particular read is  derived from. Figure 5 | Real-time and synthetic long-read sequencing approaches.",
+    "160. Glenn,T .C. Field guide to next-generation DNA  sequencers. Mol. Ecol. Resour. 11, 759769 (2011). 161. Karow,J. At AGBT , 10X Genomics launches GemCode  platform; shipments slated for Q2 as firm battles IP  lawsuits.  GenomeWeb  [online], https://www. genomeweb.com/sample-prep/agbt-10x-genomics- launches-gemcode-platform-shipments-slated-q2-firm- battles-ip-lawsuits  (2Mar 2015). Competing interests statement The authors declare competing interests: see Web version  for  details. FURTHER INFORMATION",
+    "DNA), and provide the means to link sequences containing applications. First,  base- callers like Phred (4,5) extract raw sequences from raw data. There are  also contig assemblers like Phrap (University of Washington, http://bozeman. mbt.washington.edu/phrap.docs/phrap.html ) or CAP3 (6) that assemble frag- ments to contigs and packages like consed (7) or GAP4 (8), which are used to  finish sequencing projects. These programs are not explained in detail here.",
+    "Nat. Biotechnol. 30, 10331036 (2012). 111. Chrystoja,C.C. & Diamandis,E.P . Whole genome  sequencing as a diagnostic test: challenges and  opportunities. Clin. Chem. 60, 724733 (2014). 112. McGuire,A.L. etal.  Point-counterpoint. Ethics and  genomic incidental findings. Science 340, 10471048  (2013). 113. Bowers,J. etal.  Virtual terminator nucleotides for  next-generation DNA sequencing. Nat. Methods 6,  593595 (2009). 114. Heger,M. Chinas Direct Genomics unveils new",
+    "11 BIOINFORMATIC CHALLENGES FOR GENOMIC MEDICINE      Processing and managing of high-throughput sequence data     High throughput sequencing offers severa l advantages relative to array-based  genotyping or expression assays. First, unlike genotyping arrays, whole genome sequencing is not limited to interrogating onl y known sequence variants. Similarly, RNA- sequencing (RNA-seq) enables expression quanti fication of novel transcripts that are not"
+  ],
+  "answer": "### How-to Guide for Genetic Sequencing  Genetic sequencing is a powerful tool used to determine the order of nucleotides in DNA. This guide will walk you through the basic steps involved in genetic sequencing, referencing specific resources for further information.  #### Step 1: Sample Preparation 1. **Collect DNA Sample**: Obtain a DNA sample from the organism or tissue of interest. This can be done using various methods such as blood draws, buccal swabs, or tissue biopsies. 2. **Extract DNA**: Use a DNA extraction kit to isolate DNA from the collected sample. Ensure the DNA is of high quality and free from contaminants.  #### Step 2: Library Preparation 1. **Fragment DNA**: Break the DNA into smaller fragments. This can be done using mechanical shearing or enzymatic digestion. 2. **Add Adapters**: Ligate specific adapters to the ends of the DNA fragments. These adapters are necessary for the sequencing process and often contain barcodes for sample identification [6].  #### Step 3: Amplification 1. **PCR Amplification**: Amplify the DNA fragments using Polymerase Chain Reaction (PCR). This step increases the quantity of DNA to ensure there is enough material for sequencing.  #### Step 4: Sequencing 1. **Choose a Sequencing Platform**: Select an appropriate sequencing platform based on your needs. Popular platforms include Illumina, Ion Torrent, Pacific Biosciences, and Oxford Nanopore Technologies [1], [4], [7]. 2. **Load the Sample**: Load the prepared DNA library onto the sequencing machine. Follow the manufacturer's instructions for your specific platform.  #### Step 5: Data Acquisition 1. **Run the Sequencer**: Start the sequencing run. The machine will read the nucleotide sequences of the DNA fragments and generate raw data. 2. **Monitor the Run**: Keep an eye on the sequencing run to ensure it is proceeding correctly. Some platforms provide real-time monitoring capabilities [6].  #### Step 6: Data Analysis 1. **Base Calling**: Use base-calling software to convert raw sequencing data into nucleotide sequences. Tools like Phred are commonly used for this purpose [8]. 2. **Sequence Assembly**: Assemble the short reads into longer contiguous sequences (contigs) using assemblers like Phrap or CAP3 [8]. 3. **Data Interpretation**: Analyze the assembled sequences to identify genetic variants, structural variations, and other genomic features. This step may involve bioinformatics tools and databases [10].  #### Step 7: Validation and Reporting 1. **Validate Results**: Confirm the accuracy of your sequencing results using additional methods such as Sanger sequencing or quantitative PCR. 2. **Report Findings**: Compile your findings into a comprehensive report. Include details about the methods used, the quality of the data, and any significant genetic variations identified.  #### Further Information For more detailed information on specific sequencing technologies and platforms, refer to the following resources: - [10X Genomics](http://www.10xgenomics.com) - [Illumina](http://www.illumina.com) - [Ion Torrent](https://www.thermofisher.com/us/en/home/brands/ion-torrent.html) - [Pacific Biosciences](http://www.pacb.com) - [Oxford Nanopore Technologies](https://www.nanoporetech.com) [1]  By following these steps and utilizing the referenced resources, you can successfully perform genetic sequencing and analyze the resulting data.",
+  "question": "Create a how-to guide for genetic sequencing"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_19 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_19
new file mode 100644
index 0000000..e480449
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_19
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2020 - Precision and Personalized Medicine How Genomic.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2015 -Pandey- Functional Analysis of Genomic Variation and Impact on Molecular.pdf",
+    "2015 - Functional Analysis of Genomic Variation and Impact on Molecular and Higher Order Phenotypes.pdf",
+    "2017 - Infection control in the new age of genomic epidemiology.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2008 - Gene Expression Profiling.pdf"
+  ],
+  "extraction_id": [
+    "de09f30d-e9ba-5379-8c7a-85b2cd2ed6c8",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "f4e989e5-c3d0-5d5c-b8c3-95894a14630b",
+    "fa426831-7c04-56c1-a191-1ebbc35342ed",
+    "04f06fb6-b2ff-57d4-bac0-de5cf4782ff3",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "c14d1c74-a14a-5037-8d3f-f32a60faa9a5"
+  ],
+  "document_id": [
+    "cd11028a-933b-52a0-9534-c173323056ef",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "93381e23-494b-5bc2-9d09-ef315506601f",
+    "263d327b-f5db-54e4-a215-b3f8a51cd7d6",
+    "8f028916-b990-5e95-b2a6-e69f451cc291",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c"
+  ],
+  "id": [
+    "chatcmpl-ADZLI3pesAhKfrcor9tw6wwv6rY5l",
+    "8477a774-dddb-5541-b8d7-d51a7e56b0af",
+    "13a6b6f9-4a9a-5eb9-ac79-d986d9e613f0",
+    "d1158643-3625-5855-a03d-eec4ac96eb4d",
+    "cdf2b80f-1509-50a2-9cb2-a36dd6f3f2cc",
+    "f8ae01ae-cea8-5b8b-95c0-7147055de596",
+    "d2540614-9397-5e3e-8b5f-ad328ca973b2",
+    "199e1929-dc7c-58d4-8c8d-1c931e658e9c",
+    "1e324977-2ca5-5062-8a09-7659d516e899",
+    "98010acc-fd11-5d33-bced-626ef29f2896",
+    "3e782f01-a06e-51b6-ac8a-0e0a56939d08"
+  ],
+  "contexts": [
+    "36. Sequencing, H.G. Finishing the euchromatic sequence of the human genome. Nature 2004 ,431, 931945. 37. Heather, J.M.; Chain, B. The sequence of sequencers: The history of sequencing DNA. Genomics 2016 ,107, 18. [CrossRef] 38. Rothberg, J.M.; Leamon, J.H. The development and impact of 454 sequencing. Nat. Biotechnol. 2008 ,26, 11171124. [CrossRef] [PubMed] 39. Shendure, J.; Ji, H. Next-generation DNA sequencing. Nat. Biotechnol. 2008 ,26, 11351145. [CrossRef] [PubMed]",
+    "22. Karow, J. Qiagen launches GeneReader NGS System  atAMP; presents performance evaluation by broad.  GenomeWeb  [online], https:// www.genomeweb.com/ molecular-diagnostics/qiagen-launches-genereader- ngs-system-amp-presents-performance-evaluation   (4Nov 2015). 23. Smith,D.R. & McKernan,K. Methods of producing  and sequencing modified polynucleotides . US Patent  8058030 (2011). 24. Margulies,M. etal.  Genome sequencing in  microfabricated high-density picolitre reactors. Nature  437, 376380 (2005).",
+    "11 BIOINFORMATIC CHALLENGES FOR GENOMIC MEDICINE      Processing and managing of high-throughput sequence data     High throughput sequencing offers severa l advantages relative to array-based  genotyping or expression assays. First, unlike genotyping arrays, whole genome sequencing is not limited to interrogating onl y known sequence variants. Similarly, RNA- sequencing (RNA-seq) enables expression quanti fication of novel transcripts that are not",
+    "11 BIOINFORMATIC CHALLENGES FOR GENOMIC MEDICINE      Processing and managing of high-throughput sequence data     High throughput sequencing offers severa l advantages relative to array-based  genotyping or expression assays. First, unlike genotyping arrays, whole genome sequencing is not limited to interrogating onl y known sequence variants. Similarly, RNA- sequencing (RNA-seq) enables expression quanti fication of novel transcripts that are not",
+    "High-throughput bacterial genome sequencing: an embarrassment of choice, aworldof opportunity.NatRevMicrobiol2012;10:599-606. 11.CroucherNJ,DidelotX.Theapplicationof genomicstotracingbacterialpathogen transmission.CurrOpinMicrobiol2015;23:62-7. 12.ShendureJ,JiH.Next-generationDNAsequencing.NatBiotechnol2008;26:1135- 45. 13.MillerJR,KorenS,SuttonG.Assemblyalgorithmsfornext-generationsequencing data.Genomics2010;95:315-27. 14.OlsonND,LundSP,ColmanRE,FosterJT,SahlJW,SchuppJM,etal.Bestpractices",
+    "sequencing. Genome Res. 20, 11651173 (2010). 64. English,A.C. etal.  Assessing structural variation in a  personal genome-towards a human reference diploid  genome. BMC Genomics 16, 286 (2015). 65. Carneiro,M.O. etal.  Pacific Biosciences sequencing  technology for genotyping and variation discovery in  human data. BMC Genomics 13, 375 (2012). 66. Quail,M.A. etal.  A tale of three next generation  sequencing platforms: comparison of Ion T orrent,  Pacific Biosciences and Illumina MiSeq sequencers.",
+    "Nat. Biotechnol. 30, 10331036 (2012). 111. Chrystoja,C.C. & Diamandis,E.P . Whole genome  sequencing as a diagnostic test: challenges and  opportunities. Clin. Chem. 60, 724733 (2014). 112. McGuire,A.L. etal.  Point-counterpoint. Ethics and  genomic incidental findings. Science 340, 10471048  (2013). 113. Bowers,J. etal.  Virtual terminator nucleotides for  next-generation DNA sequencing. Nat. Methods 6,  593595 (2009). 114. Heger,M. Chinas Direct Genomics unveils new",
+    "sequencing. Bioinformatics 31, 20402042 (2015). 46. Qiagen.  Oncology insights enabled by knowledge base- guided panel design and the seamless workflow of the  GeneReader NGS system  Press Release. Qiagen   [online], http://www.genereaderngs.com/PROM-9192- 001_1100403_WP_GeneReader_NGS_0116_NA.pdf   (2016). 47. Forgetta,V. etal.  Sequencing of the Dutch elm disease  fungus genome using the Roche/454 GS-FLX Titanium  System in a comparison of multiple genomics core",
+    "FURTHER INFORMATION 10X Genomics: http://www.10xgenomics.com 454 Sequencing: http://www.454.com Advances in Genome Biology and Technology (AGBT):  http://www.agbt.org BGISEQ500: http://seq500.com/en/portal/Sequencer.shtml Illumina: http://www.illumina.com Ion Torrent: https://www.thermofisher.com/us/en/home/ brands/ion-torrent.html Oxford Nanopore Technologies: https://www.nanoporetech. com Pacific Biosciences: http://www.pacb.com Personal Genome Project: http://www.personalgenomes.org",
+    "DNA), and provide the means to link sequences containing applications. First,  base- callers like Phred (4,5) extract raw sequences from raw data. There are  also contig assemblers like Phrap (University of Washington, http://bozeman. mbt.washington.edu/phrap.docs/phrap.html ) or CAP3 (6) that assemble frag- ments to contigs and packages like consed (7) or GAP4 (8), which are used to  finish sequencing projects. These programs are not explained in detail here."
+  ],
+  "answer": "### How-to Guide for Genetic Sequencing  Genetic sequencing is a powerful tool used to determine the order of nucleotides in DNA. This guide will walk you through the basic steps involved in genetic sequencing, referencing specific sources for more detailed information.  #### Step 1: Sample Preparation 1. **DNA Extraction**: Begin by extracting DNA from your sample. This can be done using various commercial kits or standard laboratory protocols. 2. **DNA Quantification and Quality Check**: Measure the concentration and purity of the extracted DNA using spectrophotometry or fluorometry. Ensure the DNA is of high quality and free from contaminants.  #### Step 2: Library Preparation 1. **Fragmentation**: Fragment the DNA into smaller pieces. This can be achieved through mechanical shearing, enzymatic digestion, or sonication. 2. **End Repair and A-tailing**: Repair the fragmented DNA ends and add an adenine (A) base to the 3' ends to prepare them for adapter ligation. 3. **Adapter Ligation**: Ligate sequencing adapters to the ends of the DNA fragments. These adapters are necessary for the fragments to bind to the sequencing platform.  #### Step 3: Amplification and Enrichment 1. **PCR Amplification**: Amplify the adapter-ligated DNA fragments using polymerase chain reaction (PCR). This step increases the quantity of DNA available for sequencing. 2. **Size Selection**: Select DNA fragments of the desired size range using gel electrophoresis or magnetic beads.  #### Step 4: Sequencing 1. **Loading the Sequencer**: Load the prepared DNA library onto the sequencing platform. Popular platforms include Illumina, Ion Torrent, and Pacific Biosciences [6], [9]. 2. **Sequencing Run**: Initiate the sequencing run. The platform will read the nucleotide sequences of the DNA fragments and generate raw sequence data [1], [5].  #### Step 5: Data Processing 1. **Base Calling**: Convert raw data into nucleotide sequences using base-calling software such as Phred [10]. 2. **Sequence Assembly**: Assemble the short DNA sequences into longer contiguous sequences (contigs) using software like Phrap or CAP3 [10]. 3. **Quality Control**: Assess the quality of the assembled sequences and remove any errors or low-quality reads.  #### Step 6: Data Analysis 1. **Variant Calling**: Identify genetic variants such as single nucleotide polymorphisms (SNPs) and insertions/deletions (indels) using variant calling software. 2. **Annotation**: Annotate the identified variants to understand their potential impact on gene function and association with diseases.  #### Step 7: Interpretation and Reporting 1. **Bioinformatics Analysis**: Use bioinformatics tools to interpret the sequencing data in the context of the research question or clinical application [3], [4]. 2. **Report Generation**: Generate a comprehensive report summarizing the findings, including identified variants and their potential implications.  #### Additional Resources - For a detailed history and development of sequencing technologies, refer to Heather and Chain's review [1]. - For information on specific sequencing platforms and their comparisons, see Quail et al. [6]. - For best practices in sequencing and data management, consult Olson et al. [5].  By following these steps, you can successfully perform genetic sequencing and analyze the resulting data to gain valuable insights into the genetic makeup of your samples.",
+  "question": "Create a how-to guide for genetic sequencing."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_2 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_2
new file mode 100644
index 0000000..9143c59
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_2
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2008 - Gene Expression Profiling.pdf",
+    "2017 - Metformin alters the gut microbiome of individuals with treatment-naive type 2 diabetes, contributing to the therapeutic effects of the drug.pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2016 - Sequence Capture and Phylogenetic Utility.pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2016 - Sequence Capture and Phylogenetic Utility.pdf",
+    "2016 - Integrated multi-omics of the human gut microbiome in a case study of familial type 1 diabetes.pdf",
+    "2016 - Sequence Capture and Phylogenetic Utility.pdf",
+    "2004 - Linking nutrition to genomics.pdf"
+  ],
+  "extraction_id": [
+    "3f898a5b-0b72-59b9-b923-a5bca2db11c6",
+    "7595d721-9b06-5442-a876-e389ca4a66be",
+    "5a11860d-c422-5e6d-8a31-be81de4e1c8d",
+    "c5beca95-6108-5a67-8f74-fb39b9a36d3c",
+    "3aa1db4d-6c18-53ab-8859-676d34d2b2ae",
+    "99821df5-c257-5c1f-9fe8-18d5865d5c1e",
+    "f9e001fe-b0b0-5cd5-be1b-9377ac52b079",
+    "1c7453d1-119d-5575-b950-7b400de2b3a4",
+    "c9f26c8e-b56c-5a1a-95f4-5824f05ba3d0",
+    "b7d8dfc5-094a-5d4e-969a-97e287939187"
+  ],
+  "document_id": [
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "448d68d1-19a8-5f4c-a48b-8d33597bd03b",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "6232f392-169a-50c5-b8c9-a250f3d840cc",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "6232f392-169a-50c5-b8c9-a250f3d840cc",
+    "f0405966-38bf-5a04-aa2c-1474b11362bb",
+    "6232f392-169a-50c5-b8c9-a250f3d840cc",
+    "99891ef7-0589-5c41-a61f-1ab1fe1c8939"
+  ],
+  "id": [
+    "chatcmpl-ADZIljdVVoktIlIQ3BBIkNiAq5m4n",
+    "4067a893-52a9-5e8e-9221-c32be3241c2a",
+    "045c27b0-dad8-56f1-8772-ae9d0da11c8a",
+    "61393b99-58f3-5f1d-899d-809166e88442",
+    "3a090421-e3e5-5f38-8acf-b8053b43287b",
+    "29a51de9-1da1-5a4b-9de6-19a88c8593a3",
+    "559fdf4f-5d14-5277-ba7b-a367d4795ed2",
+    "3252d040-7281-54ca-a478-46a30b6d84f6",
+    "f2d72429-c697-5c58-aee0-6cf90b0387e5",
+    "4498331b-aea3-5c0c-9f0b-77a45cc400a2",
+    "dbae2fad-ec06-52a8-9dc0-7bc154faecc8"
+  ],
+  "contexts": [
+    "by shearing. A flow diagram summarizing the extraction of DNA is given in  Fig. 1.2. The above-described procedure is suitable for total cellular DNA.  If the DNA from a specific organelle or viral particle is needed, it is best to  isolate the organelle or virus before extracting its DNA, because the recovery  of a particular type of DNA from a mixture is usually rather difficult. Where  a high degree of purity is required, DNA may be subjected to density gradient",
+    "2017 Nature America, Inc., part of Springer Nature. All rights reserved. nature medicine doi:10.1038/nm.434564. Salonen, A. et al.  Comparative analysis of fecal DNA extraction methods with  phylogenetic microarray: effective recovery of bacterial and archaeal DNA using  mechanical cell lysis. J. Microbiol. Methods  81, 127134 (2010). 65. Murphy, N.R. & Hellwig, R.J. Improved nucleic acid organic extraction through use  of a unique gel barrier material. Biotechniques  21, 934936, 938939 (1996).",
+    "is the suitable preparation of the DNA template with a high level of purity  and free from contaminating DNA (14). Different procedures are used for  DNA extraction with specific protocol for mammals, plants, fungi, bacteria,  protozoan, helminthes, insects, and others. In specific cases, such as insects,  contamination can be reduced by hypochlorite treatment before extraction  to avoid contact with foreign DNA (15). DNA preparation includes the",
+    "this method is well suited for larger scale investigations of museum insect phylogenomics. We did extract DNA from relatively large insects, where one leg yields more tissue than is availablefrom crushing the entire body of most ants, for example. Thus, it remains now to be tested whether sufficient input DNA can also be obtained from smaller dried insect specimens. None-",
+    "usually requires that it be isolated and purified to a certain degree. DNA is  usually recovered from cells by methods that include cell rupture but that  prevent the DNA from fragmenting by mechanical shearing. This is gener- ally undertaken in the presence of EDTA, which chelates the magnesium ions  needed as cofactors for enzymes that degrade DNA, termed DNase. Ideally,  cell walls, if present, should be digested enzymatically (e.g., lysozyme in the",
+    "DNA and then using a gene probe representing a protein or enzyme from  one of the organisms. In this way, it is possible to search for related genes in  different species. This technique is generally termed Zoo blotting. A similar  process of nucleic acid blotting can be used to transfer RNA separated by gel  electrophoresis onto membranes similar to that used in Southern blotting. This  process, termed Northern blotting , allows the identification of specific mRNA",
+    "6. Staats M, Erkens RH, van de Vossenberg B, Wieringa JJ, Kraaijeveld K, Stielow B, et al. Genomic trea- sure troves: complete genome sequencing of herbarium and insect museum specimens. PLOS ONE. 2013; 8:e69189. doi: 10.1371/journal.pone.0069189 PMID: 23922691 7. Burrell AS, Disotell TR, Bergey CM. The use of museum specimens with high-throughput DNA sequencers. J Hum Evol. 2015; 79:35 44. doi: 10.1016/j.jhevol.2014.10.015 PMID: 25532801",
+    "were extracted from unthawed, frozen faecal subsamples (150 mg) after pretreatment of the weighed subsamples with 1.5 ml RNAlater ICE (LifeTechnologies) overnight.The faeces-RNAlater ICE mixture was homogenized by bead-beating, as previously described 53. Differential centrifugation and extraction using the All-In-One kit (Norgen Biotek) to recover DNA and proteins were carried out as previously described53. DNA fractions were supplemented with DNA extracted from 200 mg",
+    "DNA was then extracted destructively by grinding the frozen tissue with a sterile pestle, using aDNeasy Blood and TissueKit (Qiagen, Valencia, CA, USA) and following the manufacturer s protocol, except the DNA was eluted in 130 L ddH 2O instead of the supplied buffer. We ran 10L of each extract for 60 min at 100 volt on 1.5% agarose SB (sodium borate) gels, to estimate size of the genomic DNA. From a pool of 60 successful extractions (12 extractions produced no quantifiable DNA), we",
+    "Extracting biological information"
+  ],
+  "answer": "To extract DNA from flora or fauna, one can follow these general steps:  1. **Cell Rupture**: The cells need to be ruptured to release the DNA. This can be done using mechanical methods such as grinding the tissue with a sterile pestle [9] or using bead-beating for homogenization [8]. Enzymatic digestion of cell walls, if present, can also be used (e.g., lysozyme) [5].  2. **Preventing DNA Degradation**: During the extraction process, it is crucial to prevent DNA from fragmenting. This can be achieved by using EDTA, which chelates magnesium ions necessary for DNase enzymes that degrade DNA [5].  3. **Purification**: The DNA needs to be purified to remove contaminants. This can involve differential centrifugation and using specific kits designed for DNA extraction, such as the All-In-One kit [8] or the DNeasy Blood and Tissue Kit [9].  4. **Specific Protocols for Different Organisms**: Different organisms may require specific protocols. For example, insects may need hypochlorite treatment before extraction to reduce contamination [3].  5. **High Purity Requirements**: If a high degree of purity is required, DNA can be subjected to density gradient centrifugation [1].  By following these steps, one can effectively extract DNA from various types of flora and fauna.",
+  "question": "How would one extract the DNA, from say, flora or fauna?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_20 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_20
new file mode 100644
index 0000000..0a09b86
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_20
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2017 - Regular exercise participation improves genomic stability in diabetic patients an exploratory study to analyse telomere length and DNA damage.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf",
+    "2008 - Telomeres and Aging.pdf",
+    "2006 - Sex-specific telomere length profiles.pdf",
+    "2018 - Sex Differences in Aging Genomic Instability.pdf",
+    "2002 - Mitochondrial dysfunction leads to telomere attrition.pdf",
+    "2006 - Sex-specific telomere length profiles.pdf",
+    "2017 - The Aging Cardiovascular System.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf"
+  ],
+  "extraction_id": [
+    "0e53122e-a308-55f7-8ee8-a0857ac9c52f",
+    "efd18101-9cf2-56b5-8f86-c2aba6caa0bc",
+    "13990eb4-bef2-58ce-bf3e-0e3bc294caab",
+    "6d3bfe47-f26e-50dc-8d77-19f3797e53a0",
+    "396708f1-aa0a-571e-a8d3-7cb8404e9502",
+    "b92ede07-74a7-524a-8d2c-54b2559e8425",
+    "eb8d8e40-a484-57cb-8125-3fd5eb3f6389",
+    "6949970f-7bc7-5585-a57a-96de1b5ba6ec",
+    "d4afa45a-5efa-577b-822e-7a82c2f6508d",
+    "3b0cb0ab-421d-54d7-9816-c6a2e6f1ac68"
+  ],
+  "document_id": [
+    "dcaf7b09-2d54-5cbf-b061-e3c4e6c6c518",
+    "62b635c3-040e-512a-b016-6ef295308a1e",
+    "61d9c326-d36e-55c1-a891-335dc943e70f",
+    "09c78a17-4a1f-52c1-be4d-994fd9fd71d0",
+    "8cfb5529-7f0c-58fc-b6e4-b3ee800fb72f",
+    "d8bc729b-7513-58b7-b12e-0db1fb6d3b7d",
+    "09c78a17-4a1f-52c1-be4d-994fd9fd71d0",
+    "d3ff8471-986b-5fa0-b9c4-96eaaa8fce7c",
+    "62b635c3-040e-512a-b016-6ef295308a1e",
+    "262df0d6-ad68-544a-88ed-b4568f305858"
+  ],
+  "id": [
+    "chatcmpl-ADZLSMnXSYde3yxfC3WAn4RccN6wO",
+    "bb069c10-45f1-5a83-95e3-4b7655874ba7",
+    "28e98b7e-f273-5bdd-9979-185133f311af",
+    "5f940245-af1d-5eee-84dc-942017c523d0",
+    "607cbd31-d430-5517-8212-208b25af32bf",
+    "53508a9e-d064-58a3-a4f9-0785470a1462",
+    "7fad29bd-12bf-53d0-af89-aadd38b974ff",
+    "64ef9964-1831-5a7a-8a69-5e8d0c332d37",
+    "1b453e12-a0c4-59db-a978-bbebd689e7dc",
+    "65fb74aa-f3c3-5c80-919f-329169db982f",
+    "f181e6da-58b6-5f26-87a2-355e25388673"
+  ],
+  "contexts": [
+    "repetitive nucleotide sequences at the end of each eukaryotic chromosome, which protects them from attrition and damage. Although the relationship between leukocyte telomere length (LTL) and diabetes is still questioned 8,  different studies have shown that T2D individuals have shorter leukocyte telomeres than non-T2D individuals9, 10   that may be associated with disease progression11. Indeed, the decreased antioxidant capacity described in patients",
+    "Telomeres are arrays of linked nucleotide hexamer repeats that are found at the  ends of chromosomes in a vast clade of organisms [14]. While the sequence of  these telomeric repeats can vary between organisms, their biological function is  highly conserved, which is to limit damage inflicted on genes during the replica- tion of chromosomes. Telomere length is progressively shortened with each round  of genomic replication, unless it is restored through the action of a ribonucleo-",
+    "telomere length,a phenomenon attributed to higher levels of oxidativestress at the cellular level (70). More recent studies havelinked telomere length in smooth muscle cells with senes-cence and disease severity in patients with atherosclero-sis (141, 150). Leukocyte telomere length was also short ina cohort of similar patients and associated with a higherrisk of developing occult cardiovascular disease (71).More data are needed to understand and validate the useof leukocyte telomere length as a biomarker",
+    "age telomere length through accumulation of several short telo- meres (Londono-Vallejo et al., 2001; Martens et al., 2000) is  responsible for senescence or whether a speci  c chromosome  arm limits the replication potential of human cells (Hemann et al., 2001). Individual chromosome arms were shown to have  large variations in their length (Lansdorp et al., 1996; Benn,  1997; Londono-Vallejo et al., 2001), and chromosome 17p seemed to be equipped with especially short telomeres in hu-",
+    "Telomeres are specialized structures that protect the ends of linear chromosomes. They shorten during aging due to the unidirectional activity of DNA polymerase, which leaves a section of DNA unrepli-cated on the lagging strand. Telomeres also are subject to shortening by genotoxic stress, such as oxidative damage (33). Among many eukaryotes, the enzyme telomerase maintains telomere length; but telomerase activity varies over the lifespan and between cell types, tissues, and species (34). In most human",
+    "TTAGGG sequence that cap the ends of chromosomes, protect-ing them from degradation and fusion. The length of telomererepeats is primarily maintained by active telomerase, which iscomposed of Telomerase RNA (TR) and a catalytic subunitTelomerase Reverse Transcriptase (TERT) (Blackburn, 2001).Extensive evidence has shown that telomere shortening anderosion lead to chromosome end-to-end fusions and genomicinstability (Blasco   et al  ., 1997; Hande   et al  ., 1999), causing",
+    "a pivotal role in maintenance of genomic integrity and func-tion (Moyzis et al., 1988; McElligott and Wellinger, 1997; van  Steensel and de Lange, 1997).   It is generally accepted that telomeres shorten during DNA  replication both in vitro and in vivo. In individuals, short telomeres are considered to be a sign of advanced age. Cawthon  and coworkers (2003) showed that telomere shortening in hu-",
+    "Each cell division shortens telomeric DNA until, at a critical length, the cells lose capping function at thechromosomal ends, activating DNA damage check-points, cell senescence, and eventually apoptosis.Telomere shortening has particular relevance in thesetting of CVD. Leukocyte telomere length (LTL) associates signi cantly with vascular cell senescence,",
+    "nization may directly affect telomere attrition, resulting in accelerated replicative  senescence and progeroid phenotypes [180]. Telomeres are regions constituted by tandem repeats of non-coding DNA  sequences 5-(TTAGGG)n-3 and a protein complex called shelterin, bound to them.  This structure ensures the stability of the genome and protects the chromosomes  from a wrong action of the DNA repair machinery [184] by allowing the formation  of a chromatin loop called T-Loop [185].",
+    "telomeres, the repetitive sequence at the end of linear chromosomes, has garnered much attention for its relation to aging. Telomere repeats serve as an internal clock for cycling cells because each round of replication results in the loss of telomeric DNA in the absence of active telomerase (reviewed in [66]). Eventually, this loss over cellular generations culminates in telomere crisis and a permanent state of"
+  ],
+  "answer": "The length of telomeres is significant for several reasons:  1. **Protection of Chromosomes**: Telomeres are repetitive nucleotide sequences at the ends of eukaryotic chromosomes that protect them from attrition and damage during replication [1], [2], [6]. They prevent chromosome ends from being recognized as DNA breaks, thereby avoiding unnecessary repair mechanisms that could lead to genomic instability [9].  2. **Aging and Cellular Senescence**: Telomere length is progressively shortened with each round of genomic replication unless restored by the enzyme telomerase [2], [5]. Short telomeres are considered a sign of advanced age and are associated with cellular senescence, where cells lose their ability to divide and function properly [4], [7], [8].  3. **Disease Association**: Shorter telomeres have been linked to various diseases. For instance, individuals with Type 2 Diabetes (T2D) have been shown to have shorter leukocyte telomeres, which may be associated with disease progression [1]. Similarly, telomere length in smooth muscle cells has been linked to disease severity in atherosclerosis, and shorter leukocyte telomeres are associated with a higher risk of developing cardiovascular disease [3], [8].  4. **Genomic Stability**: Telomere shortening and erosion can lead to chromosome end-to-end fusions and genomic instability, which are detrimental to cellular function and can lead to cancer and other genetic disorders [6].  5. **Biomarker Potential**: Due to their association with aging and disease, telomere length is being studied as a potential biomarker for various conditions, including cardiovascular diseases and diabetes [3].  In summary, the length of telomeres is crucial for maintaining chromosomal integrity, regulating cellular aging, and potentially serving as a biomarker for certain diseases.",
+  "question": "What is the significance of the length of telomeres?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_21 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_21
new file mode 100644
index 0000000..1cd69f5
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_21
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2020 - Precision and Personalized Medicine How Genomic.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2015 -Pandey- Functional Analysis of Genomic Variation and Impact on Molecular.pdf"
+  ],
+  "extraction_id": [
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "de09f30d-e9ba-5379-8c7a-85b2cd2ed6c8",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "c91e328e-4a01-5952-85b8-d7b5b47237c5",
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "c14d1c74-a14a-5037-8d3f-f32a60faa9a5",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "f4e989e5-c3d0-5d5c-b8c3-95894a14630b"
+  ],
+  "document_id": [
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "cd11028a-933b-52a0-9534-c173323056ef",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "93381e23-494b-5bc2-9d09-ef315506601f"
+  ],
+  "id": [
+    "chatcmpl-ADZLdc2V8rjlDwihUfGh20lox5Tad",
+    "98010acc-fd11-5d33-bced-626ef29f2896",
+    "13a6b6f9-4a9a-5eb9-ac79-d986d9e613f0",
+    "8477a774-dddb-5541-b8d7-d51a7e56b0af",
+    "d2540614-9397-5e3e-8b5f-ad328ca973b2",
+    "1e324977-2ca5-5062-8a09-7659d516e899",
+    "a2d9c614-903d-513a-ad88-5a40f3534988",
+    "97f2aa12-623b-53ec-9793-5834311a37dd",
+    "3e782f01-a06e-51b6-ac8a-0e0a56939d08",
+    "199e1929-dc7c-58d4-8c8d-1c931e658e9c",
+    "d1158643-3625-5855-a03d-eec4ac96eb4d"
+  ],
+  "contexts": [
+    "FURTHER INFORMATION 10X Genomics: http://www.10xgenomics.com 454 Sequencing: http://www.454.com Advances in Genome Biology and Technology (AGBT):  http://www.agbt.org BGISEQ500: http://seq500.com/en/portal/Sequencer.shtml Illumina: http://www.illumina.com Ion Torrent: https://www.thermofisher.com/us/en/home/ brands/ion-torrent.html Oxford Nanopore Technologies: https://www.nanoporetech. com Pacific Biosciences: http://www.pacb.com Personal Genome Project: http://www.personalgenomes.org",
+    "22. Karow, J. Qiagen launches GeneReader NGS System  atAMP; presents performance evaluation by broad.  GenomeWeb  [online], https:// www.genomeweb.com/ molecular-diagnostics/qiagen-launches-genereader- ngs-system-amp-presents-performance-evaluation   (4Nov 2015). 23. Smith,D.R. & McKernan,K. Methods of producing  and sequencing modified polynucleotides . US Patent  8058030 (2011). 24. Margulies,M. etal.  Genome sequencing in  microfabricated high-density picolitre reactors. Nature  437, 376380 (2005).",
+    "36. Sequencing, H.G. Finishing the euchromatic sequence of the human genome. Nature 2004 ,431, 931945. 37. Heather, J.M.; Chain, B. The sequence of sequencers: The history of sequencing DNA. Genomics 2016 ,107, 18. [CrossRef] 38. Rothberg, J.M.; Leamon, J.H. The development and impact of 454 sequencing. Nat. Biotechnol. 2008 ,26, 11171124. [CrossRef] [PubMed] 39. Shendure, J.; Ji, H. Next-generation DNA sequencing. Nat. Biotechnol. 2008 ,26, 11351145. [CrossRef] [PubMed]",
+    "sequencing. Genome Res. 20, 11651173 (2010). 64. English,A.C. etal.  Assessing structural variation in a  personal genome-towards a human reference diploid  genome. BMC Genomics 16, 286 (2015). 65. Carneiro,M.O. etal.  Pacific Biosciences sequencing  technology for genotyping and variation discovery in  human data. BMC Genomics 13, 375 (2012). 66. Quail,M.A. etal.  A tale of three next generation  sequencing platforms: comparison of Ion T orrent,  Pacific Biosciences and Illumina MiSeq sequencers.",
+    "sequencing. Bioinformatics 31, 20402042 (2015). 46. Qiagen.  Oncology insights enabled by knowledge base- guided panel design and the seamless workflow of the  GeneReader NGS system  Press Release. Qiagen   [online], http://www.genereaderngs.com/PROM-9192- 001_1100403_WP_GeneReader_NGS_0116_NA.pdf   (2016). 47. Forgetta,V. etal.  Sequencing of the Dutch elm disease  fungus genome using the Roche/454 GS-FLX Titanium  System in a comparison of multiple genomics core",
+    "for sequencing on existing short-read instrumentation,  after which data are split by barcode and reassembled  with the knowledge that fragments sharing barcodes Barcodes A series of known bases  addedto a template molecule  either through ligation or  amplification. After  sequencing, these barcodes  can be used to identify which  sample a particular read is  derived from. Figure 5 | Real-time and synthetic long-read sequencing approaches.",
+    "160. Glenn,T .C. Field guide to next-generation DNA  sequencers. Mol. Ecol. Resour. 11, 759769 (2011). 161. Karow,J. At AGBT , 10X Genomics launches GemCode  platform; shipments slated for Q2 as firm battles IP  lawsuits.  GenomeWeb  [online], https://www. genomeweb.com/sample-prep/agbt-10x-genomics- launches-gemcode-platform-shipments-slated-q2-firm- battles-ip-lawsuits  (2Mar 2015). Competing interests statement The authors declare competing interests: see Web version  for  details. FURTHER INFORMATION",
+    "DNA), and provide the means to link sequences containing applications. First,  base- callers like Phred (4,5) extract raw sequences from raw data. There are  also contig assemblers like Phrap (University of Washington, http://bozeman. mbt.washington.edu/phrap.docs/phrap.html ) or CAP3 (6) that assemble frag- ments to contigs and packages like consed (7) or GAP4 (8), which are used to  finish sequencing projects. These programs are not explained in detail here.",
+    "Nat. Biotechnol. 30, 10331036 (2012). 111. Chrystoja,C.C. & Diamandis,E.P . Whole genome  sequencing as a diagnostic test: challenges and  opportunities. Clin. Chem. 60, 724733 (2014). 112. McGuire,A.L. etal.  Point-counterpoint. Ethics and  genomic incidental findings. Science 340, 10471048  (2013). 113. Bowers,J. etal.  Virtual terminator nucleotides for  next-generation DNA sequencing. Nat. Methods 6,  593595 (2009). 114. Heger,M. Chinas Direct Genomics unveils new",
+    "11 BIOINFORMATIC CHALLENGES FOR GENOMIC MEDICINE      Processing and managing of high-throughput sequence data     High throughput sequencing offers severa l advantages relative to array-based  genotyping or expression assays. First, unlike genotyping arrays, whole genome sequencing is not limited to interrogating onl y known sequence variants. Similarly, RNA- sequencing (RNA-seq) enables expression quanti fication of novel transcripts that are not"
+  ],
+  "answer": "### How-to Guide for Genetic Sequencing  Genetic sequencing is a powerful tool used to determine the order of nucleotides in DNA. This guide will walk you through the basic steps involved in genetic sequencing, referencing specific resources for further information.  #### Step 1: Sample Preparation 1. **Collect DNA Sample**: Obtain a sample containing DNA. This could be from blood, saliva, or tissue. 2. **Extract DNA**: Use a DNA extraction kit to isolate DNA from the sample. Ensure the DNA is pure and of high quality.  #### Step 2: Library Preparation 1. **Fragment DNA**: Break the DNA into smaller fragments. This can be done using mechanical shearing or enzymatic methods. 2. **Add Adapters**: Ligate specific sequences (adapters) to the ends of the DNA fragments. These adapters are necessary for the sequencing process and allow the fragments to bind to the sequencing platform [6].  #### Step 3: Amplification 1. **PCR Amplification**: Amplify the DNA fragments using Polymerase Chain Reaction (PCR). This step increases the quantity of DNA to ensure there is enough material for sequencing.  #### Step 4: Sequencing 1. **Choose a Sequencing Platform**: Select a sequencing technology based on your needs. Popular platforms include Illumina, Ion Torrent, Pacific Biosciences, and Oxford Nanopore Technologies [1], [4], [7]. 2. **Load DNA onto Sequencer**: Load the prepared DNA library onto the sequencing machine. Each platform has specific protocols for loading and running the sequencing process.  #### Step 5: Data Analysis 1. **Base Calling**: Use software to convert raw sequencing data into nucleotide sequences. Tools like Phred are commonly used for this purpose [8]. 2. **Sequence Assembly**: Assemble the short DNA reads into longer sequences (contigs). Software like Phrap or CAP3 can be used for this step [8]. 3. **Variant Calling**: Identify genetic variants by comparing the assembled sequences to a reference genome. This step is crucial for understanding genetic differences and mutations.  #### Step 6: Interpretation 1. **Bioinformatics Analysis**: Use bioinformatics tools to analyze the sequencing data. This includes identifying genes, predicting functions, and understanding genetic variations [10]. 2. **Report Findings**: Summarize the results in a report. This should include information on identified variants, potential implications, and any recommendations for further analysis or validation.  #### Additional Resources - **10X Genomics**: Offers advanced solutions for single-cell sequencing and spatial genomics [1]. - **Illumina**: Provides a wide range of sequencing platforms and reagents [1]. - **Pacific Biosciences**: Known for long-read sequencing technology [1], [4]. - **Oxford Nanopore Technologies**: Offers portable and real-time sequencing solutions [1].  By following these steps and utilizing the referenced resources, you can successfully perform genetic sequencing and analyze the resulting data.",
+  "question": "Create a how-to guide for genetic sequencing"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_22 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_22
new file mode 100644
index 0000000..36c1153
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_22
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2019 - Machine-learning to stratify diabetic patients using novel cardiac biomarkers and integrative genomics.pdf",
+    "2014 - Computational tools to aid the design and development of a genetic reference population.pdf",
+    "2020 - Precision and Personalized Medicine How Genomic.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2012 - Next Steps in Cardiovascular Disease Genomic Research.pdf",
+    "2016 - Coming of age ten years of next.pdf"
+  ],
+  "extraction_id": [
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "8c9e74de-fe33-53c9-a26a-c4e4be6ab217",
+    "a744f8ce-7920-5fb8-acce-912f70112924",
+    "de09f30d-e9ba-5379-8c7a-85b2cd2ed6c8",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "c2635fbd-ed4f-574a-be56-076a770af2b4",
+    "abff2242-b300-56f4-9974-2eefc93ae1aa"
+  ],
+  "document_id": [
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "332ac2ec-accc-5370-a4d2-6fec9ce7e072",
+    "70cbde25-6406-5a31-91ae-57f430e8f267",
+    "cd11028a-933b-52a0-9534-c173323056ef",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "5cde24e3-2463-5751-8ef3-97cda391449b",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0"
+  ],
+  "id": [
+    "chatcmpl-ADZLof1DVn6jbHUiHhxz6hRe0WZVb",
+    "97f2aa12-623b-53ec-9793-5834311a37dd",
+    "98010acc-fd11-5d33-bced-626ef29f2896",
+    "d2540614-9397-5e3e-8b5f-ad328ca973b2",
+    "13a6b6f9-4a9a-5eb9-ac79-d986d9e613f0",
+    "822f10c2-37f6-5543-a1d4-6f640c464fb7",
+    "da667832-cd2f-5af6-a0a8-a17542b0a2e2",
+    "8477a774-dddb-5541-b8d7-d51a7e56b0af",
+    "1e324977-2ca5-5062-8a09-7659d516e899",
+    "943d9de4-1181-5811-aa37-e8d560c39562",
+    "571b0089-beff-5726-a831-5b5c1f95c53a"
+  ],
+  "contexts": [
+    "160. Glenn,T .C. Field guide to next-generation DNA  sequencers. Mol. Ecol. Resour. 11, 759769 (2011). 161. Karow,J. At AGBT , 10X Genomics launches GemCode  platform; shipments slated for Q2 as firm battles IP  lawsuits.  GenomeWeb  [online], https://www. genomeweb.com/sample-prep/agbt-10x-genomics- launches-gemcode-platform-shipments-slated-q2-firm- battles-ip-lawsuits  (2Mar 2015). Competing interests statement The authors declare competing interests: see Web version  for  details. FURTHER INFORMATION",
+    "FURTHER INFORMATION 10X Genomics: http://www.10xgenomics.com 454 Sequencing: http://www.454.com Advances in Genome Biology and Technology (AGBT):  http://www.agbt.org BGISEQ500: http://seq500.com/en/portal/Sequencer.shtml Illumina: http://www.illumina.com Ion Torrent: https://www.thermofisher.com/us/en/home/ brands/ion-torrent.html Oxford Nanopore Technologies: https://www.nanoporetech. com Pacific Biosciences: http://www.pacb.com Personal Genome Project: http://www.personalgenomes.org",
+    "sequencing. Genome Res. 20, 11651173 (2010). 64. English,A.C. etal.  Assessing structural variation in a  personal genome-towards a human reference diploid  genome. BMC Genomics 16, 286 (2015). 65. Carneiro,M.O. etal.  Pacific Biosciences sequencing  technology for genotyping and variation discovery in  human data. BMC Genomics 13, 375 (2012). 66. Quail,M.A. etal.  A tale of three next generation  sequencing platforms: comparison of Ion T orrent,  Pacific Biosciences and Illumina MiSeq sequencers.",
+    "22. Karow, J. Qiagen launches GeneReader NGS System  atAMP; presents performance evaluation by broad.  GenomeWeb  [online], https:// www.genomeweb.com/ molecular-diagnostics/qiagen-launches-genereader- ngs-system-amp-presents-performance-evaluation   (4Nov 2015). 23. Smith,D.R. & McKernan,K. Methods of producing  and sequencing modified polynucleotides . US Patent  8058030 (2011). 24. Margulies,M. etal.  Genome sequencing in  microfabricated high-density picolitre reactors. Nature  437, 376380 (2005).",
+    "mina barcoded adapters and prepared using a 300-cycle MiSeq Reagent Micro Kit v2 (Illumina, San Diego, CA). PCR amplicons were sequenced on the MiSeq with paired-end (PE) 250 base pair reads. Files were aligned to the bisulfite converted reference genome GRCh38 release 94 implementing Bismark [35, 36]. Alignment was  obtained through Bismark using the Bowtie2 [37] engine using non-directional and paired-end.  Complete sequencing code is provided (https  ://githu b.com/qahat",
+    "sequencing data to solutions from the genotyping array data. iv               PREVIEW",
+    "36. Sequencing, H.G. Finishing the euchromatic sequence of the human genome. Nature 2004 ,431, 931945. 37. Heather, J.M.; Chain, B. The sequence of sequencers: The history of sequencing DNA. Genomics 2016 ,107, 18. [CrossRef] 38. Rothberg, J.M.; Leamon, J.H. The development and impact of 454 sequencing. Nat. Biotechnol. 2008 ,26, 11171124. [CrossRef] [PubMed] 39. Shendure, J.; Ji, H. Next-generation DNA sequencing. Nat. Biotechnol. 2008 ,26, 11351145. [CrossRef] [PubMed]",
+    "sequencing. Bioinformatics 31, 20402042 (2015). 46. Qiagen.  Oncology insights enabled by knowledge base- guided panel design and the seamless workflow of the  GeneReader NGS system  Press Release. Qiagen   [online], http://www.genereaderngs.com/PROM-9192- 001_1100403_WP_GeneReader_NGS_0116_NA.pdf   (2016). 47. Forgetta,V. etal.  Sequencing of the Dutch elm disease  fungus genome using the Roche/454 GS-FLX Titanium  System in a comparison of multiple genomics core",
+    "Conventional sequencing Next-generation sequencing Sequencing  Subcloning in vectors, amplification in hosts for every single DNA fragment Direct DNA fragment sequencing  Sequencing of 100 fragments in parallel  Optional PCR amplification  Parallel sequencing of millions of small fragments Yield 1 /H11003105bp/sequencing run /H110221/H110031011bp/sequencing run Computational requirements Moderate High Cost per megabase High LowAccuracy High HighFuture directions   Direct sequencing of DNA molecules",
+    "Nature Reviews | GeneticsCleavage agent Single-base-encoded probes A probe with a single known base and degenerate bases hybridizes to a template and is imagedResetAfter each imaging step, both the probe and anchor are removed Probe with known base at n+1a  SOLiD       (Thermo Fisher) b  Complete Genomics      (BGI) Paired-end sequencingSequencing is performed for both the left and right sides of the adapterTTG AG TC CC GA CT TATA A"
+  ],
+  "answer": "### Guide to Genetic Sequencing  Genetic sequencing is a powerful tool used to determine the order of nucleotides in DNA. This guide will walk you through the basics of genetic sequencing, the different technologies available, and the steps involved in the sequencing process.  #### 1. Introduction to Genetic Sequencing Genetic sequencing involves determining the precise order of nucleotides within a DNA molecule. This information is crucial for understanding genetic information, diagnosing genetic disorders, and conducting research in genomics.  #### 2. Types of Sequencing Technologies There are several sequencing technologies available, each with its own advantages and applications:  - **Sanger Sequencing**: The first-generation sequencing method, known for its high accuracy but lower throughput. - **Next-Generation Sequencing (NGS)**: Includes various platforms that allow for high-throughput sequencing, enabling the sequencing of millions of fragments simultaneously [9].  #### 3. Next-Generation Sequencing Platforms Several NGS platforms are widely used in the field:  - **Illumina**: Known for its high accuracy and throughput, suitable for a wide range of applications [2]. - **Ion Torrent**: Uses semiconductor technology for sequencing, offering a cost-effective solution [2]. - **Pacific Biosciences (PacBio)**: Provides long-read sequencing, which is useful for identifying structural variations and complex regions in the genome [3], [2]. - **Oxford Nanopore Technologies**: Offers portable sequencing devices with real-time data analysis capabilities [2]. - **10X Genomics**: Known for its GemCode platform, which provides linked-read sequencing for better genome assembly [1], [2].  #### 4. Steps in Genetic Sequencing The sequencing process generally involves the following steps:  1. **Sample Preparation**: DNA is extracted from the sample and prepared for sequencing. This may involve fragmentation, adapter ligation, and amplification [5]. 2. **Library Construction**: The prepared DNA fragments are converted into a library suitable for the sequencing platform being used [5]. 3. **Sequencing**: The library is loaded onto the sequencing platform, where the DNA fragments are sequenced in parallel [9]. 4. **Data Analysis**: The raw sequencing data is processed and aligned to a reference genome. Tools like Bismark and Bowtie2 are commonly used for alignment and analysis [5]. 5. **Interpretation**: The processed data is interpreted to identify genetic variants, structural variations, and other genomic features [3].  #### 5. Applications of Genetic Sequencing Genetic sequencing has a wide range of applications, including:  - **Medical Diagnostics**: Identifying genetic mutations associated with diseases. - **Genomic Research**: Studying genetic variations and their implications. - **Personal Genomics**: Providing individuals with information about their genetic makeup [2].  #### 6. Future Directions The field of genetic sequencing is rapidly evolving, with ongoing advancements aimed at increasing accuracy, reducing costs, and enabling direct sequencing of DNA molecules without amplification [9].  #### 7. Further Information For more detailed information on specific sequencing platforms and technologies, you can visit the following resources:  - **Illumina**: [http://www.illumina.com](http://www.illumina.com) - **Ion Torrent**: [https://www.thermofisher.com/us/en/home/brands/ion-torrent.html](https://www.thermofisher.com/us/en/home/brands/ion-torrent.html) - **Pacific Biosciences**: [http://www.pacb.com](http://www.pacb.com) - **Oxford Nanopore Technologies**: [https://www.nanoporetech.com](https://www.nanoporetech.com) - **10X Genomics**: [http://www.10xgenomics.com](http://www.10xgenomics.com)  By understanding the basics of genetic sequencing and the available technologies, you can better appreciate the power and potential of this transformative field in genomics.",
+  "question": "Create a guide for genetic sequencing"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_23 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_23
new file mode 100644
index 0000000..91915f8
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_23
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2011 - Genetics and Genomics of Cholesterol and Polyunsaturated Fatty Acid Metabolism in Relation to Coronary Heart Disease Risk.pdf",
+    "2011 - Analysis of cognitive functions in recombinant inbred strains of rats produced by crossbreeding of SHR and BN Lx. lines.pdf",
+    "2018 - Multivariate analysis of genomics data to identify potential pleiotropic genes.pdf",
+    "2008 - The Common P446L Polymorphism in GCKR Inversely.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "2018 - Genomic 5-mC contents in peripheral.pdf",
+    "2021- Development of genome-wide polygenic risk scores for lipid traits and clinical applications for dyslipidemia, subclinical atherosclerosis, and diabetes cardiovascular complications among East Asians.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf"
+  ],
+  "extraction_id": [
+    "1745eb7d-e39e-5304-96a5-c351809d4795",
+    "b3d1c55f-bcdc-59b2-8191-623e8e79b87b",
+    "4bee64c1-92ce-5b8c-925d-f30c4acab84b",
+    "e54089b3-5559-55f8-b482-ceae887ce6ca",
+    "9738a79c-f506-5134-87c7-0ef5020c0077",
+    "3fc1141e-011e-5606-952c-5d7d9201459e",
+    "a95613b6-a2e8-5d84-841f-ae8879611a9e",
+    "e860a438-567e-50e4-99a9-759ff52ffdda",
+    "c194ef31-2e93-5de6-9c35-6365056b1e54",
+    "e464416a-2dc9-53c0-988c-b0131883aa79"
+  ],
+  "document_id": [
+    "111e0e1e-d336-55ee-87a8-2f03b02473c2",
+    "6f628ea8-1286-5d74-80e5-55439f21805d",
+    "2f7bad8a-28aa-5add-b9c3-8c2d445719f5",
+    "1d74871a-be20-5ca3-ab8f-0a68e885dcf4",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "f720cb59-3a8f-58e0-9cb8-e34b7d0bb74f",
+    "ce8040c7-157f-54c5-b28b-3224e8871415",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae"
+  ],
+  "id": [
+    "chatcmpl-ADZM6xG6YQyyKS0yjhUsqz3mB8jmi",
+    "53aa581f-06d1-52b3-b847-08ea3d95a980",
+    "799c27b2-d017-5ded-bb75-76b3d65b0bf6",
+    "142eead0-6648-5c97-a2da-770aff4986f6",
+    "0cbbec43-43bb-502d-a26d-fbc669ff29ee",
+    "60c771fb-a2fa-5f19-a13c-e4086864bcd5",
+    "bd69128b-7357-5e87-ab9a-af6f4f3fc733",
+    "3fd58cb6-d19a-5337-9a84-a8e4e4e0b97c",
+    "134d285e-3f83-5ed6-ab9d-774b81068a3d",
+    "7a2c163e-e4ef-58ee-86dc-399d15d20eb7",
+    "cba6153e-0a7f-540c-897b-40cbf9284ea9"
+  ],
+  "contexts": [
+    "Deregulated lipid metabolism (dyslipidemia) that manifests as hypercholesterolemia,  hypertriglyceridemia, low high -density -lipoprotein (HDL) cholesterol levels or a  combination of those is an established risk factor for CHD among other established risk  factors. The liver is of major importance in maintaining whole- body lipid metabolic",
+    "23   Atherogenic dyslipidemia, manifested by raised triglycerides and low  concentrations of HDL cholesterol. There could be p resent other lipoprotein abnormalities  as well, e.g., increased lipoproteins, elevated apo lipoprotein B, small LDL and HDL  particles. All of these abnormalities have been imp licated as being atherogenic (Kolovou et  al., 2005; Ginsberget al., 2000).   Elevated blood pressure  strongly associates with obesity and commonly occu rs in  insulin-resistant persons.",
+    "plasma TGisdetermined bythelevel ofVLDL-TG (the balance between synthesis and clear- ance ofVLDL-TG), and thesynthesis ofVLDL-TG isassociated with total fatmass and liver fat[59]. Thus, thelarge amount offatmass inobese patients leads toincreasing synthesis of VLDL-TG, buttheclearance ofVLDL-TG remains unchanged. Hypertriglyceridemia isaprin- cipal characteristic ofdyslipidemia and islinked tomany other types ofdyslipidemia such as",
+    "Dyslipidemia status Normolipidemia 2,731 898 (0.33) 1,319 (0.48) 514 (0.19) 42.97End-of-study cases 2,102 611 (0.29) 1,057 (0.50) 434 (0.21) 45.79 0.01, 1.12 (1.021.22)Incident cases 959 293 (0.31) 472 (0.49) 194 (0.20) 44.84 0.9, 0.99 (0.911.09) Overall risk data are P, OR (95% CI) and incident risk data are P, HR (95% CI). Hyperglycemia and type 2 diabetes were dened according to 1997 American Diabetes Association criteria",
+    "The most characteristic lipoprotein abnormality in patients with  diabetes, especially type 2, is elevated triglyceride, i.e. VLDL, reduced HDL,  and smaller dense LDL. This lipoprotein profile is sometimes referred to as  diabetic dyslipidemia. Moreover, in conjunction with obesity, and insulin  resistance this lipoprotein profile constitutes part of the \"polymetabolic  syndrome\". The primary lipoprotein abnormality is hypertriglyceridemia .",
+    "Hyperlipidemia 63 (23%) 100 (38%) < 0.001c Diabetes 66 (24%) 106 (40%) < 0.001c TC (mmol/L) 4.36  0.55 4.37  1.07 0.832b,d TG (mmol/L) 1.01 (0.77~1.28) 1.35 (1.00~1.92) < 0.001d,e HDL-C (mmol/L) 1.26 (1.13~1.42) 1.10 (0.94~1.34) < 0.001d,e LDL-C (mmol/L) 2.57  0.36 2.43  0.88 0.017b,d FBG (mmol/L) 4.71 (4.35~5.15) 5.84 (5.31~6.87) < 0.001e PBLs counts (109/L) 5.30 (4.60~6.29) 6.58 (5.33~7.92) < 0.001e PBLs classifications (PBMCs %)40.31  8.11 34.48  10.16 < 0.001b",
+    "lipid traits as (lipid follow-up lipid baseline ) / lipid baseline . Dyslipidemia/abnormal lipid levels were defined according to the thresholds used in clinical practice guidelines [ 19]: (1) TC 5.1 mmol/l; TG 1.1 mmol/l; and LDL-C 3.4 mmol/l in children; (2) TC 5.1 mmol/l; TG1.4 mmol/l; and LDL-C 3.4 mmol/l in adolescents; (3) TC 5.2 mmol/l; TG 1.7 or 1.97 mmol/l; and LDL- C1.8 or 2.6 mmol/l in adults or patients with T2D. In the two cohorts of adult women, cIMT was mea-",
+    "dyslipidemia. It also lowered in ammatory biomarkers (CRP and PAI - 1) associated",
+    "usually associated with reduced HDL cholesterol and small dense LDL.  Biliary  cholesterol  +  Bile acids  Blood vessel  Figure 3. HDL metabolism: HDL production requires addition of lipid  to small, nascent particles. This lipid arrives via hydrolysis of VLDL and  chylomicrons with transfer of surface lipids (phospholipid PL, and free  cholesterol, FC) via the actions of phospholipid transfer protein (PL TP). A  second pathway is via effiux of cellular free cholesterol (FC), a process",
+    "shift in the composition of the lipoprotein particle from one de  ned as VLDL to"
+  ],
+  "answer": "Dyslipidemia is defined as deregulated lipid metabolism that manifests as hypercholesterolemia (high cholesterol levels), hypertriglyceridemia (high triglyceride levels), low high-density lipoprotein (HDL) cholesterol levels, or a combination of these conditions [1]. It is an established risk factor for coronary heart disease (CHD) and can involve various lipoprotein abnormalities, such as increased lipoproteins, elevated apolipoprotein B, and small LDL and HDL particles [2].",
+  "question": "Define dyslipidemia."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_24 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_24
new file mode 100644
index 0000000..50f3dd9
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_24
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2007 - Functional genomic approach to identify novel genes.pdf",
+    "2017 - Gene-based genome-wide association study identified 19p13.3 for lean body mass.pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2007 - Functional genomic approach to identify novel genes.pdf",
+    "2012 - Quantitative proteomic analysis reveals novel mitochondrial targets.pdf",
+    "2017 - Systems Genetics Analysis to Identify the Genetic Modulation of a Glaucoma-Associated Gene.pdf",
+    "2007 - Functional genomic approach to identify novel genes.pdf",
+    "2007 - Functional genomic approach to identify novel genes.pdf",
+    "2014 - Evidence for the presence of somatic mitochondrial DNA.pdf",
+    "2007 - Functional genomic approach to identify novel genes.pdf"
+  ],
+  "extraction_id": [
+    "3aebacd5-b198-5144-8fe3-34ac09f6e1e8",
+    "088aee9f-e751-5b68-9142-235a84d4297d",
+    "d69da2f6-dced-546c-b3d7-bd57253a15a6",
+    "3aebacd5-b198-5144-8fe3-34ac09f6e1e8",
+    "ad33e7d7-8861-52b8-92ac-b307a6c42a0d",
+    "d83136ee-cf42-5167-902b-470a6e0b2d3c",
+    "47e612a2-c181-5c19-8b1c-c6aaa107e88a",
+    "90107b5e-bd2c-56ae-a7b9-ac4ca506e3e5",
+    "655a0cc4-b432-5b84-9eac-43b932700af5",
+    "3aebacd5-b198-5144-8fe3-34ac09f6e1e8"
+  ],
+  "document_id": [
+    "6d78d3b0-1754-5f4c-8dbc-c972c68d21cd",
+    "970711f0-abe2-5129-8374-b7bb24c58b9e",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "6d78d3b0-1754-5f4c-8dbc-c972c68d21cd",
+    "a4ca3799-0273-5765-98fd-4902eeacf894",
+    "67e804db-8127-5938-8d7f-a5918cdf4f86",
+    "6d78d3b0-1754-5f4c-8dbc-c972c68d21cd",
+    "6d78d3b0-1754-5f4c-8dbc-c972c68d21cd",
+    "a835ef5d-9a88-58ac-a0bc-5411c9f9031a",
+    "6d78d3b0-1754-5f4c-8dbc-c972c68d21cd"
+  ],
+  "id": [
+    "chatcmpl-ADZMBjMCSl2sOAftMGQRmV7FIDy31",
+    "7bab87cd-7bd1-525e-98fe-14e9c51f617a",
+    "a6bd961a-ed80-5266-8519-4399a1c7a6ba",
+    "84d74cdd-eccb-55a7-a407-a14f90e30d78",
+    "cf86caa8-56c6-58f5-8233-e1de07f1ef37",
+    "53fee260-155f-58c5-810e-ff890292a46b",
+    "7ff5d5ed-f6a9-52ab-8994-4bdb61161f4f",
+    "cf251057-96a1-581c-b60a-63a04d845511",
+    "3b3c603d-bb36-5995-b4c5-fd06e733dfda",
+    "2e0e68ba-1804-5040-81aa-8746d263083a",
+    "a2b4ddd9-1f98-5271-85a0-5d79c529253e"
+  ],
+  "contexts": [
+    "oxidoreductase MitochondriaF29C4.2 IV Cytochrome",
+    "complex III. It functions to form a part of the mitochondrial respiratory chain. It may also act as a binding fac-tor for the iron-sulfur protein. Mitochondrial Complex III is composed of one mitochondrial-encoded subunit (MT-CYB) and ten nuclear-encoded subunits. The complex is located within the mitochondrial inner mem- brane and plays an important role in biochemical synthesis of ATP . It functions to catalyze electrons to trans-",
+    "Chapter 36 Directed Protein Evolution  653 3.1.9. SHIPREC Cytochromes are proteins that contain heme groups and are responsible for  the transport of electrons. P450 is a family of membrane-bound cytochromes  with an absorption maximum of 450 nm when complexed with CO. One of the  major roles of the cytochrome P450 system is the detoxification of harmful  substances. Sieber et al. (23) produced hybrids of two cytochromes, which share only",
+    "F42A9.5 cyp-33E2 IV Cytochrome P450 MitochondriaF21D5.8 IV Mitochondrial 28S ribosomal protein S33 MitochondriaC33A12.1 IV NADH: ubiquinone oxidoreductase, ETS complex I subunit MitochondriaZK809.3 IV NADH: ubiquinone oxidoreductase MitochondriaC47E12.2 IV Mitochondrial ADP/ATP carrier protein MitochondriaY57G11C.12 IV NADH: ubiquinone oxidoreductase MitochondriaY41E3.4 ers-1 IV Glutaminyl tRNA synthetase, predicted to be mitochondrial MitochondriaY55F3B_743.b IV Mitochondrial ribosomal protein",
+    "Process 2.9 2.9 25.4 gi 149058974 rCG44669 (cytochrome c oxidase, subunit VIIc;Cox7c)1.19 0.2121 1.35 1.42 0.05 1.30 1.26 0.0480 1.26 unclassied 29.6 29.7 56.0 gi 149016520 rCG50966 (3-oxoacid-CoA transferase 1(OXCT1/SCOT)1.12 0.3615 1.27 1.08 0.46 1.23 1.33 <0.0001 1.12 metabolism: ketone metabolism 60.9 60.9 67.6 gi 116242506 stress-70 protein, mitochondrial precursor(75 kDa glucose-regulatedprotein) (Heat shock 70kDa protein 9)1.07 0.1432 1.12 1.02 0.39 1.10 1.13 0.0300 1.09 protein folding; protein",
+    "413 Table 2 Gene ontology Database: molecular function name: Cytochrome  c oxidase activity ID:GO:0004129 C = 16 O = 2 E = 0.12 R = 17.06  rawP  = 0.0060  adjP  = 0.0590 Index User IDGene   symbol Gene namesEntrez   gene Ensemble 1 ILMN_2657141 Surf1 Surfeit gene 1 20930 ENSMUSG00000015790 2 ILMN_1254971 Cox6b1 Cytochrome c oxidase,  subunit VIb polypeptide110323 ENSMUSG00000036751 Database: molecular function Name: NADH dehydrogenase activity ID:GO:0003954",
+    "F42A9.5 cyp-33E2, cytochrome P450 family 13.81 (  0.49) 118 0.0010 C47E12.2 Mitochondrial ADP/ATP carrier protein 16.00 (  0.78) 136 < 0.0001 F21D5.8 Mitochondrial 28S ribosomal protein S33 15.95 (  0.99) 136 < 0.0001 C33A12.1 NADH: ubiquinone oxidoreductase 16.28 (  1.05) 139 0.0003 ZK809.3 NADH: ubiquinone oxidoreductase 23.46 (  1.14) 200 < 0.0001 Y57G11C.12 nuo-3, NADH: ubiquinone oxidoreductase 20.71 (  1.18) 177 < 0.0001",
+    "Y66A7A1 100 52 33 4 0 9.00 (  0.29) 0.0572 210 Y71H2_388.c PP2A regulatory subunit (cytochrome C oxidase subunit) 100 82 48 2 0 5.57 (  0.20) < 0.0001 130 F54D8.2 Cytochrome c oxidase subunit Vla 100 70 41 22 3 5.62 (  0.27) < 0.0001 131 F56D2.1 Mitochondrial processing peptidase 100 55 17 3 0 4.46 (  0.20) 0.4303 104 K04G7.4 Nuo-4, NADH: ubiquinone oxidoreductase 100 78 55 4 0 5.06 (  0.23) < 0.0001 118 T20H4.5 Ubiquinone Fe-S protein 100 99 89 45 2 7.58 (  0.18) < 0.0001 177",
+    "and (Iso211Ser) 1.1383 . (ii) Overview of MT-CYB mutation on electron transport chain. From the complex II the reduced form of ubiquinone move through the hydrophobic region of the membrane by diffusion. When the ubiquinone comes in contact with the next carrier in the electron-transport chain, the electron is transferred to cytochrome reductase, or the cytochrome b-c1 complex (Complex  III). The mutated cytochrome b  loses the ability to accept incoming",
+    "c   oxidase polypeptide Mitochondria K08F11.4 year-1 IV Tyrosyl-tRNA synthetase, predicted to be mitochondrial MitochondriaE04A4.7 IV Cytochrome   c  Mitochondria"
+  ],
+  "answer": "Cytochromes are proteins that contain heme groups and are responsible for the transport of electrons [3]. They play a crucial role in the mitochondrial respiratory chain, which is essential for the biochemical synthesis of ATP [2]. One specific family of cytochromes, known as cytochrome P450, is involved in the detoxification of harmful substances [3].",
+  "question": "What is cytochrome?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_25 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_25
new file mode 100644
index 0000000..c14d3d9
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_25
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - eQTL analysis in mice and rats.pdf",
+    "2015 - Genetic Control of Survival and Weight Loss during Pneumonic Burk.pdf",
+    "2015 -Emery- Genetic Control of Survival and Weight Loss during Pneumonic Burk.pdf",
+    "2006 - From_gene_to_behavior_and_back_again_new.pdf",
+    "2005 - quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2005 -Broadkin- quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2009 - Experimental_Evolution.pdf",
+    "2009 - Garland_and_Rose_Experimental_Evolution.pdf",
+    "2005 - quantitative-trait-analysis-in-the-investigation-of-function-and.pdf",
+    "2016 - Social interactions and indirect genetic effects on complex juvenile and adult traits.pdf"
+  ],
+  "extraction_id": [
+    "71981bfb-284e-50ad-854e-2055c07f77a7",
+    "615ee0cd-5960-57e5-b4e6-56e4b8020a1b",
+    "268a23e8-f528-5b59-89f2-188331e0a03c",
+    "64c0287d-aeea-52eb-a074-e9591c5593ae",
+    "9de93371-6239-53c2-b42c-71f615a0614b",
+    "0a5c759e-8dab-55f1-ac59-e8211ec683b8",
+    "8ee78018-b998-590c-99ab-788a447ede81",
+    "cbce50ea-be78-5d54-beb1-849222c5bfdd",
+    "0a895880-91c0-5079-b258-73926b38430f",
+    "0b91ce42-1ba4-530c-8d77-6ddbdc0e759d"
+  ],
+  "document_id": [
+    "8d67ea90-f7b1-5bb8-937c-4a9eceddff43",
+    "ae1025b0-1410-51ae-9be2-26fa2e9d5808",
+    "a9aceace-bf48-5472-b54c-59a458a84c62",
+    "7a088b36-11b7-5379-bfe5-ce571e11de07",
+    "0dc730ba-4ff4-52aa-a988-71075113c416",
+    "e6027e7f-aec0-5e76-8aff-96b36389e701",
+    "34821353-1b74-5ee2-ac39-66dd46f145bf",
+    "496faa7f-9623-5ab7-9816-7c3755abb3aa",
+    "dac1c73c-0b5f-5a54-bb12-7e8b654009c0",
+    "06e126d3-b75d-57db-8edb-09de6ae13b24"
+  ],
+  "id": [
+    "chatcmpl-ADZMFYjDRlNaYIo2GAk3sVCPv4DGN",
+    "73540700-b5cf-5838-852b-b281ca086140",
+    "374c456a-d1db-5b4a-8713-97abe4162d77",
+    "b9d52798-0235-5018-bccd-560565d16cc3",
+    "fef212bc-631b-591d-b8e3-d1523da0507d",
+    "c8f17022-aeae-5242-9082-d6d1eee4c4bf",
+    "1b2de424-be9f-572d-bd62-dc2ecd92192b",
+    "f72795a1-66c3-5a98-84bc-b085e8008073",
+    "31a32dc5-81ac-52ba-a463-c61e293f21e5",
+    "b660d882-1cb0-5150-ae76-8eb3ccb88a58",
+    "985378d7-e164-581b-ac1c-97bbcda9c06f"
+  ],
+  "contexts": [
+    "While most of the Y chromosome does not undergo recombination, the recombination rate of the X chromosomeis slower than that of the autosomes. This has important consequences on the detection of significant QTLs. For a comprehensive view of these issues, see(43). 9.Probe hybridization artifacts When several probes are available for the same gene, it is not uncommon to observe a difference in the mapping results",
+    "8 QTL Mapping     Allelic variation exists among natural populations and inbred strains, and this is  reflective of the segregation of quantitative tr ait loci (QTLs) [96]. QTLs are stretches of  DNA that are closely linked to genes that underlie a phenotype of interest. QTL analysis has been proven to be an invaluable tool to  help unravel heritable traits, by enabling  researchers to map different quantitative traits back to the genomic location involved in the regulation of these phenotypes.",
+    "8 QTL Mapping     Allelic variation exists among natural populations and inbred strains, and this is  reflective of the segregation of quantitative tr ait loci (QTLs) [96]. QTLs are stretches of  DNA that are closely linked to genes that underlie a phenotype of interest. QTL analysis has been proven to be an invaluable tool to  help unravel heritable traits, by enabling  researchers to map different quantitative traits back to the genomic location involved in the regulation of these phenotypes.",
+    "The basic  pr emise  of QTL  an alysis  is simple  (Ph illips  and Belknap, 2002 ) . First,  one must  meas  ure a speci  c phen  otype  within  a popul  ation. Next, the population must be genotyped at a hundred or more marker loci186 Boehm II et al.",
+    "genes underlying QTLs in animals and plants (see for example Shirley et al 2004,Korstanje & Paigen 2002, Fridman et al 2004). I should also point out, though, that even in a single QTL region isolated in a congenic strain, it is possible that there is more than one allele that aects the phenotype. So, you have a fair pointabout the challenges and complexities of QTL analysis. Koolhaas: There are dierent questions underlying both approaches. The QTL",
+    "genes underlying QTLs in animals and plants (see for example Shirley et al 2004,Korstanje & Paigen 2002, Fridman et al 2004). I should also point out, though, that even in a single QTL region isolated in a congenic strain, it is possible that there is more than one allele that aects the phenotype. So, you have a fair pointabout the challenges and complexities of QTL analysis. Koolhaas: There are dierent questions underlying both approaches. The QTL",
+    "through analysis of line crosses, quantitative trait loci (QTL) mapping, and verification of candidate genes with quantitative complementation tests or genetic engineering (e.g.,McGuire and Tully 1987; Chandra et al. 2001; Dierick and Greenspan 2006; Edwardset al. 2006). They can also be used to study the underlying physiological, neural, andmolecular mechanisms of the differences in behavior between selected and controllines, or between divergently selected lines.",
+    "through analysis of line crosses, quantitative trait loci (QTL) mapping, and verification of candidate genes with quantitative complementation tests or genetic engineering (e.g.,McGuire and Tully 1987; Chandra et al. 2001; Dierick and Greenspan 2006; Edwardset al. 2006). They can also be used to study the underlying physiological, neural, andmolecular mechanisms of the differences in behavior between selected and controllines, or between divergently selected lines.",
+    "genetic background. Gene identification of QTL should be distinguished from identification of the quanti- tative trait nucleotide (QTN). The latter is a daunting task, since SNPs are so frequent. Final proof for a QTN in mice would require placing a genomic segment containing theputative QTN from a donor mouse strain on the background of another strain using homologous recombination and reproducing the phenotype of the donor strain.",
+    "measuring correlations between genetic markers and phenotypic  traits in a population. Individuals are scored for their  phenotype      for  a particular trait, and their genotype at a marker. If there is a differ- ence in mean phenotype between those individuals with one geno- type at a particular locus compared with the other, than we can infer  that there is a QTL linked to that marker [ 40 ,  153 ]. 2.3  Analysis and QTL  MappingDavid G. Ashbrook and Reinmar Hager"
+  ],
+  "answer": "To differentiate between X and Y DNA with respect to DNA tracing and determining quantitative trait loci (QTLs), consider the following points:  1. **Recombination Rates**:    - The Y chromosome largely does not undergo recombination, which means it is passed down relatively unchanged from father to son. This makes it easier to trace paternal lineage but limits its utility in QTL mapping due to the lack of recombination [1].    - The X chromosome, on the other hand, does undergo recombination, albeit at a slower rate compared to autosomes. This slower recombination rate can affect the detection of significant QTLs, making it a unique consideration in QTL analysis [1].  2. **QTL Mapping**:    - QTLs are stretches of DNA closely linked to genes that underlie specific phenotypes. The process of QTL mapping involves measuring phenotypes within a population and genotyping at numerous marker loci to identify genomic locations associated with these traits [2], [4].    - The differences in recombination rates between the X and Y chromosomes mean that QTL mapping on the X chromosome can be more complex due to its unique inheritance pattern and recombination characteristics, whereas the Y chromosome's lack of recombination simplifies paternal lineage tracing but complicates QTL mapping [1], [2].  In summary, the primary differences between X and Y DNA in the context of DNA tracing and QTL determination lie in their recombination rates and inheritance patterns. The Y chromosome's lack of recombination makes it straightforward for paternal lineage tracing but less useful for QTL mapping. In contrast, the X chromosome's slower recombination rate compared to autosomes presents unique challenges and considerations in QTL analysis [1], [2].",
+  "question": "How does one tell the difference between X and Y DNA, with respect  to DNA tracing and determining QTLs?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_26 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_26
new file mode 100644
index 0000000..0e7afb7
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_26
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Functional genomics research in aquaculture principles and general approaches.pdf",
+    "2017 - Identification of quantitative trait loci associated with the susceptibility of mouse spermatozoa to cryopreservation.pdf",
+    "2009 - Garland_and_Rose_Experimental_Evolution.pdf",
+    "2009 - Experimental_Evolution.pdf",
+    "2019 - Discovery of early life stress interacting and sex-specific quantitative trait loci impacting cocaine responsiveness.pdf",
+    "2011 - Evidence for widespread changes in promoter.pdf",
+    "2017 - Identification of quantitative trait loci associated with the susceptibility of mouse spermatozoa to cryopreservation.pdf",
+    "2017 - Identification of quantitative trait loci associated with the susceptibility of mouse spermatozoa to cryopreservation.pdf",
+    "2011 - Using animal models to disentangle the role of genetic, epigenetic, and environmental influences on behavioral outcomes associated with maternal anxiety and depression.pdf",
+    "2012 - Functional genomics research in aquaculture principles and general approaches.pdf"
+  ],
+  "extraction_id": [
+    "c3a2c07f-e216-5dc0-92ea-f7c210e90974",
+    "002f921f-e651-538b-aec0-b357d2c08ee9",
+    "4a07567a-57db-5110-aa52-cc76b8df0d32",
+    "5e459c02-b084-5d1a-80fd-90643c6045f5",
+    "1b0fe3c9-4c86-5e80-9dde-faadbcdd44a1",
+    "1f77a329-36c6-5cef-884f-221358ef64ec",
+    "cf45ca6a-fbcd-52bc-b936-4f992c8e5537",
+    "3a518d3b-30c3-5cee-8344-c21cf4b32941",
+    "c570b41d-fa75-588a-bce6-30c7a64b6997",
+    "a59b4a8a-4f9c-59d3-b184-267db613e8cc"
+  ],
+  "document_id": [
+    "a39b4cc1-8661-578b-a61b-b9962e45fc33",
+    "e549cb7c-db24-540f-aded-e67442470525",
+    "496faa7f-9623-5ab7-9816-7c3755abb3aa",
+    "34821353-1b74-5ee2-ac39-66dd46f145bf",
+    "92c8a239-31ad-5ef3-bc37-a571afcdd187",
+    "1c35d1ec-a7c2-56f0-8ccf-e554814e94ef",
+    "e549cb7c-db24-540f-aded-e67442470525",
+    "e549cb7c-db24-540f-aded-e67442470525",
+    "9c2f76c6-b991-58ac-a332-50e1a3c016d8",
+    "a39b4cc1-8661-578b-a61b-b9962e45fc33"
+  ],
+  "id": [
+    "chatcmpl-ADZMMx7Aorq6SFjdy1f98MjZd0bQz",
+    "5e6a7f2c-1232-5c7e-bba5-9779ec280ed0",
+    "3c7fb887-5dbb-54d9-9664-75ceabd4ebf3",
+    "1fa77703-e337-518f-bdd8-6c3ebad0754e",
+    "833bb510-694a-5646-960e-66b5e38ef609",
+    "30b42710-3270-5e84-9227-266db106f470",
+    "5db85ce4-25a9-5aba-b5ca-ab394fb0394f",
+    "80bc495d-85ab-5b77-8984-1f4e80ace849",
+    "0e360074-4b6b-5b81-a96d-509266b7b637",
+    "084d7eaf-290a-525b-b01b-f0537e46f56e",
+    "2a5ee720-370b-5c4b-b7ea-fe4c3b2a9ea4"
+  ],
+  "contexts": [
+    "ferentiation in animals reared at male- and female-producing temperatures (Fernandino et al., 2011). From a pure experimental point of view, there are several potential sources of environ- mental inuences that need to be under con- trol in order to avoid confounding results when studying gene expression levels (Hodgins-Davis and Townsend, 2009; Table 8.3). One of them is effect of the developmental environment, typi- cally in the range of weeks to years. Size is pos-",
+    "the fertilization rate (Table 1). There was an interaction between the two factors (strain and",
+    "subtle, and often uncontrollable, environmentalfactors. Behaviors are often influenced by multiple genes with complex gene-by-gene,gene-by-environment, and environment-by-environment interactions. This is one reason,for example, that single-gene mutants are relatively uninformative (see also Rauser et al.this volume), though we described a case in which such mutants were useful for explor-ing mechanisms underlying the evolution of mating systems in voles.",
+    "subtle, and often uncontrollable, environmentalfactors. Behaviors are often influenced by multiple genes with complex gene-by-gene,gene-by-environment, and environment-by-environment interactions. This is one reason,for example, that single-gene mutants are relatively uninformative (see also Rauser et al.this volume), though we described a case in which such mutants were useful for explor-ing mechanisms underlying the evolution of mating systems in voles.",
+    "environment interactions, particularly the contribution of environmen- tal factors in utero (Burmeister, McInnis, & Zllner, 2008; Henriksen, Nordgaard, & Jansson, 2017), and these limitations in turn hinder the development of a mechanistic understanding of aetiology. Here, we dissect the impact of gene prenatal environmental interactions on cocaine responsiveness of adult male and female mice from the BXD recombinant inbred panel. Early life stressors, including prenatal stress (PNS), are important",
+    "onmental factors, some of which have been shown toalter placental gene expression, as well as epigeneticmarks [10]. These include diet [11,12], smoking [13],and assisted reproductive techniques [14,15]. Mountingevidence implicates epigenetic marks, such as DNA methylation, in mediating environmentally-induced reg- ulation of genome function. More studies into theeffects of the environment on the placental epigenomeare warranted due the importance of this organ in regu-lating pregnancy development.",
+    "as well as the intrinsic fertilizing ability of the strain. Therefore, the results of the QTL analysis  based on the fertilization rates of frozen thawed spermatozoa might have reflected the 220  cumulative effect of these two factors. T o exclude the possible background strain effects, we  calculated the ratio of the fertilization rate of frozen thawed spermatozoa per that of fresh  spermatozoa in individual male mice (designated here as relative fertilization rate ). As shown",
+    "male ; Relative fertilization  rate (%) = (Fertilization rate with f rozen spermatozoa   (%)/Fertilization rate with f resh spermatozoa  (%))  100  (n = 6 for each strain) .     Fig. 2. Genome -wide interval mapping for suggestive QTLs affecting the fertilization rate 515  using frozen thawed spermatozoa.  (A) Mapping based on the actual fertilization rates. (B)  Mapping based on the relative fertilization rates. Critical intervals were selected based on peak",
+    "duce the behavioral differences observed in these inbred strains.The interaction of genes and the environment to produce phe-notypic outcomes has been acknowledged and accepted for quitesome time in the scientic community. However, the exact mech-anism by which the environment can act on genetic materialhas only recently begun to be investigated in a more systematicmanner. A ROLE FOR EPIGENETICS IN THE LINK BETWEEN MATERNAL CARE AND BEHAVIORAL OUTCOMES IN ANIMAL MODELS",
+    "I na d d i t i o n ,i ts h o u l db en o t e dt h a tt h ee f - fect of temperature on sex determination has a genetic basis itself and an interaction be- tween families and temperature effect has been reported in several species (Schultz, 1993; Van- deputte et al., 2007). Finally, other environ- mental effects such as pH, hypoxia, and so- cial factors have claimed to be involved on sex determination (reviewed by Guerrero-Est evez and Moreno-Mendoza, 2010). All the informa-"
+  ],
+  "answer": "The environment influences fertilization through various factors that can affect both the biological and genetic mechanisms involved. Here are some specific ways in which the environment can impact fertilization:  1. **Developmental Environment**: The developmental environment, which can span from weeks to years, is a significant source of environmental influence. This includes factors such as temperature, which can affect gene expression levels and potentially influence fertilization outcomes [1].  2. **Strain and Environmental Interactions**: There is an interaction between genetic strain and environmental factors that can influence fertilization rates. For example, the fertilization rate of frozen-thawed spermatozoa compared to fresh spermatozoa can be affected by these interactions, indicating that both intrinsic genetic factors and environmental conditions play a role [2], [7].  3. **Gene-by-Environment Interactions**: Behaviors and biological processes, including fertilization, are influenced by complex interactions between multiple genes and environmental factors. These interactions can be subtle and often uncontrollable, making it challenging to isolate the specific environmental factors at play [3], [4].  4. **Epigenetic Modifications**: Environmental factors such as diet, smoking, and assisted reproductive techniques can alter gene expression through epigenetic modifications like DNA methylation. These changes can impact the placental epigenome and, consequently, pregnancy development and fertilization success [6].  5. **Temperature and Other Environmental Factors**: Temperature has a genetic basis in its effect on sex determination, and interactions between temperature and genetic factors have been observed in several species. Other environmental factors such as pH, hypoxia, and social factors can also influence sex determination and potentially affect fertilization processes [10].  In summary, the environment influences fertilization through a combination of developmental conditions, genetic and environmental interactions, epigenetic modifications, and specific environmental factors like temperature and pH [1], [2], [3], [4], [6], [7], [10].",
+  "question": "how does environment influence fertilisation"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_27 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_27
new file mode 100644
index 0000000..81e08d6
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_27
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2015_GN_Diabets_notheses.pdf",
+    "2015 -Bikai- Osteoporosis and Hypertension.pdf",
+    "2008 - Rutter_s child and adolescent psychiatry-Blackwell Pub (2008).pdf",
+    "2019 - Novel Genetic Loci Control L5 Vertebral Trabecular Bone and the Response to Low Calcium Intake in Growing BXD Recombinant Inbred Mice.pdf",
+    "1998 - Type II Diabetes, Essential Hypertension, and Obesity as Syndromes of Impaired Genetic Homeostasis The Thrifty Genotype Hypothesis Enters the 21st Century.pdf",
+    "2018 - Animal models of obesity.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2021 - Estimating genetic and environmental contributions to complex traits and diseases..pdf",
+    "2015_GN_Diabets_notheses.pdf",
+    "2015 -Bikai- Osteoporosis and Hypertension.pdf"
+  ],
+  "extraction_id": [
+    "5c6504ad-cec3-5054-b72e-7c8663b93020",
+    "2580c29f-1bd9-5a0b-bc39-36d2ac780bcd",
+    "55640275-345b-5ed5-bc71-5d2ffe9b2fd1",
+    "f540990d-044f-57aa-855b-ef89677321cb",
+    "cb4c8b17-644c-58a0-b63f-b7e839419dbf",
+    "12eaf8f0-a516-575b-af2f-54c390ad052a",
+    "1d378974-7dbd-54d3-ab94-c2306c450bd0",
+    "c6f13122-c145-50a6-8cb8-2cd0d8d40f4a",
+    "c3ea8f4a-36b2-5c54-b573-f1bb8a9f6126",
+    "b38762d9-8eb9-5290-ad8a-3ff14ffef706"
+  ],
+  "document_id": [
+    "1744d9bf-29f9-52e3-a7c9-62a916999cda",
+    "4d27e689-ce69-566f-8d4c-4101abd064e7",
+    "59daba11-206e-5bbc-8833-9d1b661532b5",
+    "de8dda5e-0e2f-5aa9-bb13-851c526b36a5",
+    "5a32199a-ae34-5829-a6f0-025d55231305",
+    "9e9af9c7-814f-562e-a04d-878528a38002",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "2ac6e31c-a3fd-546a-b18a-7f947faf1f4c",
+    "1744d9bf-29f9-52e3-a7c9-62a916999cda",
+    "4d27e689-ce69-566f-8d4c-4101abd064e7"
+  ],
+  "id": [
+    "chatcmpl-ADZMSefJYQo9NCnH90mHvha0SygR5",
+    "3202ea94-5578-58ff-b5cf-6e856e914c2e",
+    "70ba6e2c-c4a0-5c07-ba63-d91926f401ff",
+    "c775c97a-4e53-50b3-aff0-9f947f1cee70",
+    "6ad9600e-e47a-5a12-8c1d-a1cf84a1342a",
+    "1ecbfb13-5c51-57ac-b23b-09837def6f70",
+    "d49fe981-9f4f-59b2-8d91-c325a30ab87e",
+    "4955053e-da3e-530a-8b72-e8416c962d36",
+    "6c0bb788-256b-56ce-97db-124b60eeed86",
+    "3a585d96-582f-5497-b0b1-ca3a7c79c651",
+    "65d7a65d-a1d7-50f5-923e-f843fc203b21"
+  ],
+  "contexts": [
+    "economic status of a population, for example childhood nutrition status and the disease  environment etc.21 Rare are the stud ies that unveil the relation between height decline  and bone loss. A study performed by Galloway et al. on 1,024 subjects (735 women and  289 men) evaluated the correlation between height decline and bone loss with ageing.  Their findings show that bone mine ral density (BMD) plays the largest role in  determining annual height reduction.22",
+    "economic status of a population, for example childhood nutrition status and the disease  environment etc.21 Rare are the stud ies that unveil the relation between height decline  and bone loss. A study performed by Galloway et al. on 1,024 subjects (735 women and  289 men) evaluated the correlation between height decline and bone loss with ageing.  Their findings show that bone mine ral density (BMD) plays the largest role in  determining annual height reduction.22",
+    "how many eat a high phenylalanine diet.The relationship between gene and disease remains constantacross sites, but diet will act as an effect modier, controllingthe phenotypic consequences of the gene. Another example is the relationship among peak height velocity (PHV: thegrowth spurt of early adolescence), change of school anddepressive symptoms. The period of PHV may be a time whenyoungsters are particularly vulnerable to symptoms of depres-sion (Simmons & Blyth, 1987), particularly when they haveto",
+    "Dietary factor s deserve special attention as an  environmental factor that interacts with  genetics because we are exposed to our diet  every day and we  can modify it to our own benefit.  The findings from several  Ca intervention trials in children and adolescents demonstrated that  there is a large  variability in the acquisition of bone mass , despite the control of  age range and  pubertal maturation  of part icipants.(28) Weaver et al.(102) conducted a 3 -week long, controlled",
+    "rapidly than Paleolithic people andreaching both maximal adult height andsexual maturity earlier. Wehave earlier speculated thatcompression ofthegrowth history predisposes tohigher blood pressure during adoles- cence andincreases theriskofhypertension inadulthood [57] . Arecent interesting series ofstudies byBarker andcolleagues hasfor- warded theargument thatsome fraction ofthepredisposition tohyperten- sionandNIDDM maybeprogrammed inutero bylowbirth weight. Several",
+    "diets are likely to vary in composition by batch, season and  vendor. Variability in non-nutritive dietary components,  such as soluble fibre content and plant- derived phyto- estrogens, affects the progression of DIO and metabolic   disease, even affecting behavioural traits151,152. Another consideration is that humans consume  ~30% of their daily calories from fat. This fat intake is  remarkably consistent across age and BMI153 and lower  than the 40% to 60% calories from fat used in many",
+    "several factors such as age, nutritional status, overall health and geographic location, all of which in  uence the diet of",
+    "4  Hypertension   November 2020 estimated the relative influence of genetic and environmental  factors on height, weight, BMI, SBP, and DBP, as well as the genetic and environmental correlations of BMI with SBP and DBP. Furthermore, the moderating effects of BMI on SBP and DBP heritabilities were tested to explore potential gene-obe-sity interactions on BP. Contributions to the total phenotypic variances of SBP and",
+    "individuals. Augmentation index was in reverse correlation with height, in addition it  was observed that taller participants had less prevalence of hypertension and use of  antihypertensive drugs suggesting th e beneficial role of height in estimating  cardiovascular risks (159). In a study done on patients with end stage renal disease  augmentation index wa s found to negatively correlate with body height, and it was",
+    "individuals. Augmentation index was in reverse correlation with height, in addition it  was observed that taller participants had less prevalence of hypertension and use of  antihypertensive drugs suggesting th e beneficial role of height in estimating  cardiovascular risks (159). In a study done on patients with end stage renal disease  augmentation index wa s found to negatively correlate with body height, and it was"
+  ],
+  "answer": "Diet plays a significant role in determining someone's height, particularly during the growth phases of childhood and adolescence. Several factors related to diet can impact height:  1. **Childhood Nutrition Status**: The nutritional status during childhood is crucial for growth. Proper nutrition ensures that children receive the necessary vitamins and minerals required for bone growth and overall development [1].  2. **Dietary Factors and Genetics**: Diet interacts with genetics, influencing growth outcomes. For example, calcium intake during childhood and adolescence can affect bone mass acquisition, which is essential for achieving optimal height [4].  3. **Peak Height Velocity (PHV)**: The period of rapid growth during early adolescence, known as peak height velocity, can be influenced by diet. Adequate nutrition during this period is essential to support the growth spurt [3].  In summary, a well-balanced diet rich in essential nutrients is vital for achieving optimal height, especially during the critical growth periods of childhood and adolescence.",
+  "question": "how does diet impact someone's height"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_28 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_28
new file mode 100644
index 0000000..9a1b527
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_28
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2003 - Telomeres in the chicken genome stability and chromosome ends.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2010 - Water Buffalo Genome Science Comes of Age.pdf",
+    "2009 - Genetic pathways of Lyst and exfoliation syndrome.pdf",
+    "2003 - Telomeres in the chicken genome stability and chromosome ends.pdf",
+    "2005 - Numerical Algorithms for Mapping of Multiple Quantitative Trait Loci in Experimental Populations.pdf",
+    "2005 -Ljungberg- Numerical algos for Multi QTL.pdf",
+    "2018 - Invited review Genetic and genomic_ xmltexbreak_ mouse models for livestock research.pdf",
+    "2013 - Baboons as a Model to Study Genetics and Epigenetics of Human Disease.pdf"
+  ],
+  "extraction_id": [
+    "34fa36d0-0b64-5c70-8645-ba3576d9262c",
+    "02efe8ed-062d-51d2-9dd6-5a29a178b708",
+    "070b22be-cafb-5fd4-a338-ae3c62939c24",
+    "86b3157e-5b20-5e1f-aeee-f4a6f652694d",
+    "4165230b-bfd7-506c-8cfc-02868fa6bf21",
+    "b5cb2e6d-631c-5dad-bae9-26acf1dd9fb6",
+    "7c86a795-7202-5bfb-8da3-148cd8e66358",
+    "1b359995-cabb-5e75-ba37-7df272c6c232",
+    "fa8c1f01-7655-597d-8718-67ad0bc3b5ee",
+    "4dd22813-9004-571c-a351-80a2ec0f9b92"
+  ],
+  "document_id": [
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "c9124b17-6f3f-50fd-b6fc-d329db6b7cdd",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "fda7e83a-8e8c-5592-8302-687dab622323",
+    "5f35f50f-2f13-5b4c-9cfd-a96926e82f8c",
+    "c9124b17-6f3f-50fd-b6fc-d329db6b7cdd",
+    "dd7d3ea5-b23a-514e-898f-a4259ce6f6f9",
+    "bea0655c-7ef4-5754-ba14-817b72a21be2",
+    "5b167564-85a2-5886-b800-37932c3143a9",
+    "9f0acb79-6236-5add-b27e-1fb81ee4915d"
+  ],
+  "id": [
+    "chatcmpl-ADZMXN8MM8gEy7UyxGzfomf1l430J",
+    "597a0fb1-4a16-5fd3-9bdc-8be977741b82",
+    "06d5d1e7-9474-5389-9f00-5669172e73a7",
+    "65b220a4-b96c-5bcb-a65f-ed6954e44757",
+    "d9101bd9-f565-57c1-98f2-0a43b8a073b1",
+    "9712b652-cddb-522b-a7b6-053cecb6c9d9",
+    "53079eb2-6661-5082-8a3a-e9b577cbcbe9",
+    "b597e6e2-4b16-5955-8b97-972ba3cc7053",
+    "9e3ef47b-6e78-50d9-bc28-01c227f0a2ce",
+    "fbf0608e-28ec-540e-9d18-5acbfaacec5d",
+    "73394dbd-8c20-5c5c-8ac5-ac76d4bab36f"
+  ],
+  "contexts": [
+    "As seen in this karyotypic spread, the typical human cell has 46 chromosomes with 22 pairs of autosomes (numbered 122) and a pair of sex chromosomes, either XX or XY . Downloaded from http://ahajournals.org by on July 10, 2023",
+    "FIGURE 3. Telomere arrays of chicken and human chromosomes: the chicken genome contains more telomere sequence than the human",
+    "In sexually reproducing organisms, body cells contain 2 sets of chromosomes (1 set from each parent). To maintain this state, the egg and sperm that unite during fertilization each contain a single set of chromosomes. During meiosis, diploid cells undergo DNA replication, followed by 2 rounds of  cell division, producing 4 gametes, each of which has 1 set  of chromosomes (for humans, 23 unpaired chromosomes). Recombination occurs during meiosis. Mendelian diseaseSame as monogenic disease. Named",
+    "some set. Therefore, chromosome morphology sup-ports the designation of two separate genera [5].  Sex Chromosomes   Several studies have revealed high degrees of  homology among autosomal chromosomes of bovids with similar banding patterns and gene order among the chromosome arms of ca ttle, river buffalo, sheep,  and goats [14, 15]. Bovid sex chromosomes, unlike the highly similar autosomal chromosomes, share a slightly more complex rearrangement of sequences",
+    "14  Mice share an anatomy, physiology, and genome that is similar, though not  identical, to humans (May a nd Lutjen-Drecoll 2002; Smith 2002; Emes, Goodstadt et al.  2003; Huang, Winter et al. 2004). Mice and hum ans also share a su sceptibility to many  similar diseases. As an experimental genetic platform for vertebrates, tools for studying  and manipulating the mouse genome are near ly, if not completely, unparalleled",
+    "DELANY ET AL. 920 TABLE 1. Cytogenetic and telomere characteristics of vertebrate animal species (in vivo) Organism Terminal reference 2n/no. of telomere Telomere (maximum longevity) Telomeres array sizes shortening Rainbow trout 5860/116120 20 kb Unknown Oncohynchus mykiss Lejnine et al., 1995(20 yr) African clawed toad 36/72 1050 kb No Xenopus laevisBassham et al., 1998(15 yr) Laboratory mouse 40/80 50150 kb No Mus musculusKipling and Cooke, 1990(2 yr) Wild mouse 40/80 525 kb Yes",
+    "A human has 23 pairs of chromosomes, i.e. 46 in total. In each pair one chromosome has been inherited from the mother and the other from the father. The chromosomes in a pair are said to be homologous. They have the same genes at the same loci, but they may have different variants, different so called alleles, of the gene. Recall the eye color example from standard high school texts on genetics. We inherit one eye color allele from each parent, either a",
+    "A human has 23 pairs of chromosomes, i.e. 46 in total. In each pair one chromosome has been inherited from the mother and the other from the father. The chromosomes in a pair are said to be homologous. They have the same genes at the same loci, but they may have different variants, different so called alleles, of the gene. Recall the eye color example from standard high school texts on genetics. We inherit one eye color allele from each parent, either a",
+    "and zebra sh (http://www.alliancegenome.org, last access: 3 January 2018). 3 The mouse as a model animal for livestock research Mice are mammals, sharing 92 to 95 % of protein cod- ing genes with humans and other mammalian livestock species, such as cattle (Elsik et al., 2009), pigs (Humphray et al., 2007), sheep (Iannuzzi et al., 1999), and goats (Schibler et al., 1998). The mouse genome is structured into 19 autosomes and the sex chromosomes. The mouse",
+    "Figure 3: Comparison of human and baboon chromosomes. (A) Conservation of microsatellite marker order for orthologs human 12and baboon 11. (B) C hromosome inversion between orthologs hu- man 4 and baboon 5. The y-axis indicates chromosome length incentimorgans. Microsatellite markers identi ed in human have identi cation numbers that begin with D,and microsatellite markers identi ed in baboon have identi cation numbers that begin swith Pha. Figure 2: Papio hamadryas anubis (Olive baboon)"
+  ],
+  "answer": "Humans have 46 chromosomes, arranged in 23 pairs [1], [7]. The animal that has the same number of chromosomes as humans is the baboon, specifically the Olive baboon (Papio hamadryas anubis), which also has 46 chromosomes [10].",
+  "question": "which animal has the same number of chromosomes as human"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_29 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_29
new file mode 100644
index 0000000..81555eb
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_29
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Genetic architecture supports mosaic brain evolution and independent brain\u2013body size regulation_.pdf",
+    "2012 - Genetic architecture supports mosaic brain evolution and independent brain\u2013body size regulation_(1).pdf",
+    "2018 - Integrative functional genomic.pdf",
+    "2003 - Imaging genomics.pdf",
+    "2008 - The Aging Brain.pdf",
+    "2009 - Age-associated cognitive decline.pdf",
+    "2021 - System genetics in the rat HXBBXH family identifies Tti2 as a pleiotropic quantitative trait gene for adult hippocampal neurogenesis and serum glucose.pdf",
+    "2022 - System genetics in the rat HXBBXH family identifies Tti2 as a pleiotropic quantitative trait gene for adult hippocampal neurogenesis and serum glucose.pdf",
+    "2011 - A genome-wide association study of aging.pdf",
+    "2015 - A Systems-Genetics Analyses of Complex Phenotypes.pdf"
+  ],
+  "extraction_id": [
+    "e4c6a021-c822-5c6e-96ee-bdfcd9e087b6",
+    "cb9a0594-ed63-533f-b872-eea0ab9dd781",
+    "33bb0b60-582f-56b5-87da-66601ba8a482",
+    "76e11f30-b4f4-5fee-ae1f-eaf8daefc962",
+    "64f9170a-04bd-57be-ba0b-cc61edec0f37",
+    "87274deb-c57b-51c7-96f2-17111737c026",
+    "3c4e5025-5c02-522d-81f0-2354118cbf61",
+    "347bc44e-9705-5922-bfcd-22d65eb7cd80",
+    "253a4339-29d4-58c2-8a01-5137d94873b6",
+    "3f7d819b-ed86-50c7-a0c9-1955df2cead9"
+  ],
+  "document_id": [
+    "c2d37851-b1a9-5572-8de1-1cc627e5c89e",
+    "655ce593-3f0f-5065-9ce0-e9c130b6e7e4",
+    "6e37d26b-e45b-5eb8-8d79-339d9c0e05bd",
+    "b4aee92d-491c-5f9d-9c40-adb5c5cceeb6",
+    "874f5d02-35c9-5233-8ded-6e06c7570ca9",
+    "746ed855-8647-558a-9abc-c0e2d4254868",
+    "9ab8b190-fb4f-5bb0-8d04-1cd07a42192a",
+    "4198ec53-60f1-55d1-8759-b9ede1d098c0",
+    "8e9c1150-1047-54a2-bf85-1cc5000a6811",
+    "030d0226-b782-5964-8452-339777dc9658"
+  ],
+  "id": [
+    "chatcmpl-ADZMaWRstSGrYv65Txc4tWg1NphEi",
+    "340e7007-f00f-56b9-b99c-9bbf6591889a",
+    "3e7b4f88-a18f-5cdb-aa31-0eb92d4d226c",
+    "f494980a-326f-5454-8faa-890eed0a343f",
+    "54eeed5e-a1c7-566a-981d-3c40211b3992",
+    "772ad124-6371-5435-ad48-4e8546f766a0",
+    "2e99dbdc-ea40-5e40-864b-4d0ad745bc09",
+    "e5058bc7-2fc5-5a2b-852e-39efb9adc7c0",
+    "e6ce00e3-8a5d-5f20-9d18-fb8b8932dc54",
+    "919fafa2-a013-5549-9f1b-c7ccb2181215",
+    "2cc5e05a-e8fc-57cb-a7dc-c1d3ea8204a9"
+  ],
+  "contexts": [
+    "ARTICLE  nATuRE C ommunICATIons | 3:1079 | DoI: 10.1038/ncomms2086 | www.nature.com/naturecommunications  2012 Macmillan Publishers Limited. All rights reserved.Received 8 may 2012 | Accepted 23 Aug 2012 | Published 25 sep 2012 DOI: 10.1038/ncomms2086 The mammalian brain consists of distinct parts that fulfil different functions. Finlay and   Darlington have argued that evolution of the mammalian brain is constrained by",
+    "ARTICLE  nATuRE C ommunICATIons | 3:1079 | DoI: 10.1038/ncomms2086 | www.nature.com/naturecommunications  2012 Macmillan Publishers Limited. All rights reserved.Received 8 may 2012 | Accepted 23 Aug 2012 | Published 25 sep 2012 DOI: 10.1038/ncomms2086 The mammalian brain consists of distinct parts that fulfil different functions. Finlay and   Darlington have argued that evolution of the mammalian brain is constrained by",
+    "Daniel H. Geschwind, Michael J. Hawrylycz, Matthew W. State, Stephan J. Sanders, Patrick F. Sullivan, Mark B. Gerstein , Ed S. Lein , James A. Knowles , Nenad Sestan  INTRODUCTION: The brain is responsible for cognition, behavior, and much of what makes us uniquely human. The development of the brain is a highly complex process, and this process is reliant on precise regulation of molecular and cellular events grounded in the spatiotemporal regulation of the transcrip-",
+    "addition,each study implemented rigorous controls for non-genetic factors suchas age, gender, IQ and performance on the experimental task. They alsocapitalized on existing functional paradigms designed to explorephysiological aspects of distinct neural systems.",
+    "brain to prevent theapoptosis of irreplaceable neurons, even in the",
+    "Funding Funding from the BBSRC, EPSRC, ESRC and MRC is gratefully acknowledged. References 1 Brayne C (2007) The elephant in the room: healthy brains in later life, epidemiology and public health. Nat Rev Neurosci ,8, 233239. 2 Gow J, Gilhooly M (2003) Risk Factors for Dementia and Cognitive Decline . Glasgow: NHS Health Scotland. 3 House of Lords (2005) Ageing: scientific aspects. London: The Stationery Office. 4 Stern PC, Carstensen LL (2000) The Aging Mind. Washington, DC: National Academy Press.",
+    "1124 the brain. Nature Reviews Neuroscience. Nat Rev Neurosci; 2012. pp. 225239.  1125 doi:10.1038/nrn3209 1126 75. van Praag X, Fleshner M, Schwartz MW, Mattson MP. Exercise, energy intake,  1127 glucose homeostasis, and the brain. J Neurosci. 2014;34: 1513915149.  1128 doi:10.1523/JNEUROSCI.2814-14.2014 1129 76. Rafalski VA, Brunet A. Energy metabolism in adult neural stem cell fate. Progress in  1130 Neurobiology. Prog Neurobiol; 2011. pp. 182203.  1131 doi:10.1016/j.pneurobio.2010.10.007",
+    "1124 the brain. Nature Reviews Neuroscience. Nat Rev Neurosci; 2012. pp. 225239.  1125 doi:10.1038/nrn3209 1126 75. van Praag X, Fleshner M, Schwartz MW, Mattson MP. Exercise, energy intake,  1127 glucose homeostasis, and the brain. J Neurosci. 2014;34: 1513915149.  1128 doi:10.1523/JNEUROSCI.2814-14.2014 1129 76. Rafalski VA, Brunet A. Energy metabolism in adult neural stem cell fate. Progress in  1130 Neurobiology. Prog Neurobiol; 2011. pp. 182203.  1131 doi:10.1016/j.pneurobio.2010.10.007",
+    "for the creation of redun-dancy in brain circuitry, which is associated with functionalreserve and resiliency. Brain function regulates most of thecompensatory strategy supporting maintenance of homeo-static equilibrium. Both of these processes are essential tohealthy aging and longevity.",
+    "of complex traits. It has been said that The brain is the chief architect,  orchestrator and driver of behavior; behavior, in turn, is the principal function of  the brain (Gomez -Marin et al., 2014, p. 1455) , and therefore to understand one  we need to understand the other. The brain and the behaviours that it causes are  highly complex traits influen ced by many factors including genes (Hager et al.,  2012; Hitzemann et al., 2013; McCarroll and Hyman, 2013) , environment (Carola"
+  ],
+  "answer": "The proper functioning of the brain is ensured by several factors:  1. **Precise Regulation of Molecular and Cellular Events**: The development of the brain is a highly complex process that relies on the precise regulation of molecular and cellular events, which are grounded in the spatiotemporal regulation of transcription [3].  2. **Energy Metabolism**: Energy metabolism plays a crucial role in the fate of adult neural stem cells, which is essential for maintaining brain function [8].  3. **Functional Reserve and Resiliency**: The creation of redundancy in brain circuitry is associated with functional reserve and resiliency, which are critical for maintaining homeostatic equilibrium and supporting healthy aging and longevity [9].  These factors collectively ensure that the brain works effectively.",
+  "question": "what ensures brains work"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_3 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_3
new file mode 100644
index 0000000..d731737
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_3
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2016 - A novel heat shock protein alpha 8 (Hspa8) molecular network mediating responses to stress- and ethanol-related behaviors.pdf",
+    "2012 - Needs Analysis of Genetics and Genomics in Communication Sciences and Disorders.pdf",
+    "2017 - Primer in Genetics and Genomics.pdf",
+    "2018 - Identification of non-HLA genes associated with development of islet autoimmunity and type.pdf",
+    "2020 - Mainstreaming genetics and genomics a systematic review.pdf",
+    "2009 - Basic Genetics and Genomics A Primer for Nurses.pdf",
+    "2010 - Genetic variants near TIMP3 and high-density.pdf",
+    "2004 - Errand Gabpab specify PGC1dependentoxidative phosphorylation gene expressionthat is altered in diabetic muscle.pdf",
+    "2010 - Genome-wide association identifies OBFC1as a locus involved in human leukocyte telomere biology.pdf",
+    "2010 - Genome-wide association identifies OBFC1as a locus involved in human leukocyte telomere biology.pdf"
+  ],
+  "extraction_id": [
+    "600a1af4-0f16-520c-a63f-7e0af523fa3c",
+    "b7b09b33-3c90-51c9-968c-d47809e9d964",
+    "53fa3a10-5290-5209-80ce-0655d2c602a5",
+    "631667de-f20a-59b6-af3c-924b612d21ea",
+    "0120a9f0-57fd-510d-b975-b1e1f870f9fb",
+    "2cafe5f4-79a3-5234-948d-d78c20b97650",
+    "12929889-6359-5c34-8997-95a41f6202a3",
+    "715eacf0-9e21-593f-b023-84a864eb801f",
+    "0ed3fd5b-86ce-5587-90b7-1e013a7bb8ad",
+    "ccda7fa4-0bd0-5af7-919c-47b435ad81ea"
+  ],
+  "document_id": [
+    "22bb099c-aeca-51e8-a82d-5d091d9f0936",
+    "c8a76cb1-506d-57e4-a18e-548e777898e2",
+    "6fae6815-e1b5-564b-81c7-39ed62bbd999",
+    "fb67c701-af96-57ad-b1e3-1309e1b53a52",
+    "ea0695f5-c52c-568b-ba97-8fa31405ef30",
+    "c37e2ace-171b-5776-8969-86eda9736481",
+    "da9c44fa-16a0-586b-8256-f4b91f4cfef9",
+    "259bea02-bb3a-57b6-8896-0b41d6cace05",
+    "76ba50b9-6eb2-51c7-8dd8-82d840d81219",
+    "76ba50b9-6eb2-51c7-8dd8-82d840d81219"
+  ],
+  "id": [
+    "chatcmpl-ADZIrY516c5O6uEqljwZdzIXywTvS",
+    "9e595bc1-a142-525c-97d4-5edde55c5bcf",
+    "1eed369d-2525-5621-b9a7-c344c2e48f32",
+    "1ef52b83-a34c-517e-b65f-b8d9c1acb79d",
+    "ab3b9b3a-2353-5730-8dd2-3b790ca7c5f7",
+    "d2fe7d2e-8d04-52ce-a8e6-de8437a83014",
+    "6cce33b0-9975-5727-8b53-2e3478cb3230",
+    "fb261aba-abc3-5e90-a322-61841f43f60c",
+    "ab7bbeb8-bc04-527e-8840-3bb5f905dffd",
+    "c8b4753f-417d-590b-b8b8-866322504fec",
+    "0b0f3258-b6bb-5de0-91d1-767ef5fce360"
+  ],
+  "contexts": [
+    "Neurogenetics",
+    "Genetics   Genetics is the study of individual genes and their protein products (Guttmacher &",
+    "genetics and genomics, article 1DNA, genes, and chromosomes. Biological Research for Nursing ,19, 717. Dueker, N. D., & Pericak-Vance, M. A. (2014). Analysis of genetic linkage data for Mendelian traits. Current Protocols in Human Genetics ,83, 1.4.11.4.31. Fu, M. R., Conley, Y. P., Axelrod, D., Guth, A. A., Yu, G., Fletcher, J., & Zagzag, D. (2016). Precision assessment of heterogeneity of lymphedema phenotype, genotypes and risk prediction. Breast , 29, 231240.",
+    "genetic factors. 371   372   373   374    375",
+    "GENETICS in MEDICINE |Volume 22 |Number 7 |July 2020 1153",
+    "to offspring. Genes are pieces of DNA, and most genes contain the  information for making a specific protein.  zGenetics - Genetics is a term that refers to the study of genes and  their role in inheritance - the way certain traits or conditions are  passed down from one generation to another.  zGenomics  - Genomics is a relatively new term that describes the  study of all of a person's genes including interactions of those genes  with each other and the person's environment.",
+    "www.pnas.org/cgi/doi/10.1073/pnas.0912702107 PNAS |April 20, 2010 |vol. 107 |no. 16 |74017406 GENETICS",
+    "GENETICS Downloaded from https://www.pnas.org by 41.90.188.152 on July 14, 2023 from IP address 41.90.188.152.",
+    "GENETICS Downloaded from https://www.pnas.org by 41.80.118.137 on October 17, 2023 from IP address 41.80.118.137.",
+    "GENETICS Downloaded from https://www.pnas.org by 41.80.118.137 on October 17, 2023 from IP address 41.80.118.137."
+  ],
+  "answer": "Genetics is the study of individual genes and their protein products [2]. It involves understanding how genes are inherited, how they function, and how they influence traits and conditions passed down from one generation to another [6]. This field also includes the analysis of genetic linkage data for Mendelian traits [3].",
+  "question": "genetics"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_30 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_30
new file mode 100644
index 0000000..c9827c7
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_30
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2008 - Rutter_s child and adolescent psychiatry-Blackwell Pub (2008).pdf",
+    "2013 - Neural-Immune Interactions in Brain Function and Alcohol Related Disorders.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2007 - Gene expression profiles in anatomically and functionally distinct regions.pdf",
+    "2009 - Multiscale Genomic Analysis of the Corticolimbic System_ Uncoveri (1).pdf",
+    "2009 - Neuroplasticity, Psychosocial Genomics.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf"
+  ],
+  "extraction_id": [
+    "5e06bd24-8977-582c-b01b-61be91612e1a",
+    "fb4ba6b4-c3ea-5671-9da8-15fcadccff59",
+    "d0222d2f-7e27-59de-9ad0-23febb3564f8",
+    "4d38ecad-88e4-5f52-8a99-55029773de79",
+    "b848d23b-0c65-5e44-b190-1ec8e5a76545",
+    "c755176c-961c-57f0-996c-662de89048d3",
+    "8cd38348-d367-5c85-829e-e465af8184cb",
+    "995b3eb6-e505-52a0-a142-ca507eb9a9ac",
+    "264526ff-3f41-5a6d-88af-6e237cea42cb",
+    "069a2a63-f01c-5235-a118-3744c21f2baa"
+  ],
+  "document_id": [
+    "59daba11-206e-5bbc-8833-9d1b661532b5",
+    "78271275-3409-5fc7-bbdd-53c484178e0b",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "d4a001e2-8cac-58cb-be8b-b9afa9382e01",
+    "3d0df5a3-7d7c-5edc-b94d-cae582f59c12",
+    "77549d17-8f07-5b62-8134-011a68f2ebd4",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c"
+  ],
+  "id": [
+    "chatcmpl-ADZMeEIq2Wv2GICWJZcSZAU1k5Qdz",
+    "34f059bf-1e74-580d-9b52-8c940ff0f302",
+    "fd7b6e37-2aba-525e-aa22-4a9cef18827d",
+    "f2dda7e1-1af6-54b0-8ffa-856313872579",
+    "96a2a72c-b239-58f0-b116-2b1eeb3e8434",
+    "b2d814c0-e515-54b9-b994-b457ca0e2739",
+    "45e53d76-dced-5f6b-abf2-c830b41c1c90",
+    "2fc8ee5e-7a5e-57cc-98e3-e9156aec2571",
+    "fbdf5982-c2f0-5577-bce3-bc8762aef713",
+    "2199f4c4-8126-54c8-a323-6704c96bc0f7",
+    "8b65f73a-2d73-53b2-b418-f8e485d58df3"
+  ],
+  "contexts": [
+    "areas that support pos-itive emotions and deactivate brain areas that are linked withaggression, fear and sadness (Diamond, 2004); this nding is consistent with the emotional prole associated with agreeableness.",
+    "Importantly, regions of the brain responsible for emotional regulation, executive  functioning, and their consequential behavioral outcomes are sensitive to  in  ammation  [  22  ] . The extended limbic system, primitively responsible for fear and  pleasure responses, stress, memory, and learning, has been shown to be modulated  by immune signaling. Early work established that there is a high density of IL-1  receptors in the dentate gyrus and pyramidal cell layer of the hippocampus, the",
+    "the midbrain structures are implicated in cardiacresponses to social stress (Wager et al, 2009 ). It is now evident that these same brain regions are involved in emotion regulation. Furthermore, the circuitry involved in physical pain and plea-sure appears to be activated by positive and negative socially induced emotion (Takahashi et al, 2009 ). The possibility therefore arises that positive well-being may be embodied in the acti- vation of neural circuitry in a reciprocal fashion",
+    "723732. Etkin, A., Egner, T., Peraza, D. M., Kandel, E. R., and Hirsch, J. (2006). Resolving emotional conict: a rolefor the rostral anterior cingulate cortex in modulatingactivity in the amygdala. Neuron, 51 , 871882. Fales, C. L., Barch, D. M., Rundle, M. M., Mintun, M. A., Snyder, A. Z. et al (2008). Altered emotional inter-ference processing in affective and cognitive-controlbrain circuitry in major depression. Biol Psychiatry, 63, 377384. Fanselow, M. S. (2000). Contextual fear, gestalt mem-",
+    "for cognitive processes such as learning,memory, and emotions.",
+    "expression of emotional behavior. Sensory inputs with emotional components are  transmitted to the amygdala where they are processed and fu rther relayed to other regions  to modulate autonomic and behavioral responses, and to form emotional memories  (LeDoux, 2000; Rosen, 2004). As a neural substrate of emotionality, many  neuropsychiatric disorders have been associated with structural changes i n the amygdala.  Individuals with genetically predisposed susceptibility to anxiety and depression have",
+    "components can act back upon its physical substrate. Thought, emotion, and action trigger neural activity, which can lead to a reorganization of the brain, shaping future psychosocial experience. From this perspective, we are not the passive products of neurophysiology and heredity; rather, through our behavior in the social environment, we become active agents in the con-struction of our own neurobiology and, ultimately, our own lives.",
+    "et al, 1995 ; Scher et al, 2005 ), (2) are less easily distracted from negative emotion process- ing (Ellenbogen et al, 2002 ; Lyubomirsky et al, 1998 ; Siegle et al, 2002 ; Wenzlaff and Bates, 1998 ), (3) show heightened stress hormone lev- els such as cortisol that may have deleterious effects on the brain (Sapolsky, 2000 ), and (4)",
+    "et al, 2000 ). Once activated, the amygdala sets in motion a cascade of responses to threat via pro-jections to the hypothalamus and prefrontal cor-tex (LeDoux, 1996 ). A neural region that is criti- cal for regulating responses to emotional stimuli is the ventrolateral prefrontal cortex (VLPFC;Hariri et al, 2002 ). Studies have shown that the labeling of negative affective states activates the right VLPFC and that increased activity inright VLPFC is associated with decreased activ-",
+    "tially participates in negative emotional states,although it also participates in positive emo- tional states (Zald, 2003 ). The amygdala orches- trates the somatomotor, visceral, and cognitiveresponses to threats by virtue of its connections with cortical brain structures above and hypotha- lamic and brainstem structures below it (LeDouxet al, 1990 ). The nucleus accumbens and ventral striatum participate in reward responses and pos- itive emotional states. Other structures that are"
+  ],
+  "answer": "Our brains maintain emotions through a complex interplay of various brain regions and neural circuits. Key areas involved include:  1. **Amygdala**: Sensory inputs with emotional components are transmitted to the amygdala, where they are processed and further relayed to other regions to modulate autonomic and behavioral responses, and to form emotional memories [6]. The amygdala orchestrates responses to threats and is involved in both negative and positive emotional states [10].  2. **Prefrontal Cortex**: The ventrolateral prefrontal cortex (VLPFC) is critical for regulating responses to emotional stimuli. Increased activity in the right VLPFC is associated with decreased activity in the amygdala, which helps in managing negative affective states [9].  3. **Extended Limbic System**: This system, which includes structures like the hippocampus, is primitively responsible for fear and pleasure responses, stress, memory, and learning. It is modulated by immune signaling, indicating that inflammation can affect emotional regulation [2].  4. **Midbrain Structures**: These structures are involved in emotion regulation and are activated by both physical pain and pleasure, as well as positive and negative socially induced emotions [3].  5. **Nucleus Accumbens and Ventral Striatum**: These areas participate in reward responses and positive emotional states [10].  Overall, the brain maintains emotions through the coordinated activity of these regions, which process, regulate, and respond to emotional stimuli, thereby shaping our emotional experiences and behaviors.",
+  "question": "how do our brains maintain emotions"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_31 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_31
new file mode 100644
index 0000000..0a28905
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_31
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2013 - Neural-Immune Interactions in Brain Function and Alcohol Related Disorders.pdf",
+    "2013 - Neural-Immune Interactions in Brain Function and Alcohol Related Disorders.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2014 - Genetic regulatory network analysis reveals that low density lipoprotein receptor-related protein 11 is involved in stress responses in mice.pdf",
+    "2021 - Prefrontal cortex VAMP1 gene network moderates the effect of the early environment on cognitive flexibility in children.pdf",
+    "2015 - Great Is Their Sin.pdf",
+    "2009 - Multiscale Genomic Analysis of the Corticolimbic System_ Uncoveri (1).pdf",
+    "2011 - Genetic Analysis of the Neurosteroid Deoxycorticosterone and Its Relation to Alcohol Phenotypes Identification of QTLs and Downstream Gene Regulation.pdf",
+    "2011 - Genetic Analysis of the Neurosteroid Deoxycorticosterone and Its Relation to Alcohol Phenotypes Identification of QTLs and Downstream Gene Regulation.pdf",
+    "2019 - Exploring the involvement of Tac2 in the mouse hippocampal stress response through gene networking.pdf"
+  ],
+  "extraction_id": [
+    "e4e689d6-5e01-50cb-bb0f-1d958542a343",
+    "87cb54ed-b246-52a8-8922-5baa4f2f5e7c",
+    "c83a0fd3-2bc2-510b-ba66-fad5dab1c430",
+    "a576772e-e17b-56fc-96b0-bdf8c913b2e8",
+    "8c989969-10c2-533e-ad71-5e9a54499798",
+    "2992ae99-13f8-5b72-9a5b-408a1ec77e32",
+    "bccdd21d-53b6-53c5-89ae-6508fa5ea4a9",
+    "f854fcfc-5758-5d5f-944d-d1db9e72ccdd",
+    "0b62b9d2-6622-5882-b3d6-e7f8482a927a",
+    "cd49980b-e59c-5d64-816d-3a8817f099fc"
+  ],
+  "document_id": [
+    "78271275-3409-5fc7-bbdd-53c484178e0b",
+    "78271275-3409-5fc7-bbdd-53c484178e0b",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "9e59e66c-6b3f-5c99-a12c-7bb6fd0d899f",
+    "976026ce-9e0c-5b0b-8469-abc8f92dbdf0",
+    "e5ae9710-3049-5327-82e4-e6626eb670c2",
+    "3d0df5a3-7d7c-5edc-b94d-cae582f59c12",
+    "4eef9c8d-17bf-5ed2-a90c-6bc64f7374b1",
+    "4eef9c8d-17bf-5ed2-a90c-6bc64f7374b1",
+    "8277ae11-6516-54d2-9723-73749d46db9b"
+  ],
+  "id": [
+    "chatcmpl-ADZMkkQgFtXLUsuTUrfbvqGBOuc3R",
+    "3667e7f7-c984-567e-9757-19d7827c2a52",
+    "db05c1b0-1a66-5a2c-9680-564167f95ffe",
+    "dfdcca45-79ae-5e00-bae0-175860786128",
+    "1900d276-5346-5041-b497-41b8f1dde22e",
+    "cc9faf66-a0d0-5427-9f84-004d1b450b5a",
+    "7c9bf714-0d21-5104-9aed-4bd1b191fbf4",
+    "b06f880b-97c9-5541-a76e-a5f37f31fa6a",
+    "010d5687-d237-51ca-87a1-e7e0af944e39",
+    "869496a0-2bff-569f-ba3a-03294ebf2e98",
+    "19df7543-5231-56d5-a59a-e342565b737d"
+  ],
+  "contexts": [
+    "pin-releasing hormone (CRH), adrenocorticotropic hormone (ACTH), and  glucocorticoids (GC), which are also called stress hormones. These hormones con- tribute to the regulation of immune responses and can also affect neuronal survival,  neurogenesis, synaptic plasticity, and behavioral responses  [  1,   2  ] . The HPA axis is  a three-tiered biological system that begins at the highest level with the release of  CRH from the hypothalamic paraventricular nucleus (PVN). CRH-expressing neu-",
+    "stressor in  uences the interleukin-1beta system, tumor necrosis factor-alpha, transforming growth factor-beta1, and neuropeptide mRNAs in speci  c brain regions. Brain Res Bull 51:187193       63.    Deak T et al (2005) Stress-induced increases in hypothalamic IL-1: a systematic analysis of  multiple stressor paradigms. Brain Res Bull 64:541556       64.    Hennessy MB et al (2004) Responses of guinea pig pups during isolation in a novel",
+    "stressful events. In rats and mice, the secretion of hypothalamicpituitaryadrenal hormones istypically greater, and increased HPA activity often persists into adulthood (Koehl et al, 1999 ). Basal levels of adrenal hormones are more typ-ically reported to be normal in primates, but there may be alterations in the diurnal hormone rhythm or an altered negative feedback, whichresults in protracted cortisol responses once acti-vated. Many effects of prenatal stress on brain",
+    "Y in depression and stress. Brain Research 1314, 194 205. Mozhui, K., Karlsson, R.M., Kash, T.L., Ihne, J., Norcross, M., Patel, S., Farrell, M.R., Hill, E.E., Graybeal, C., Martin, K.P., Camp, M., Fitzgerald, P.J., Ciobanu, D.C., Sprengel, R., Mishina, M., Wellman, C.L., Winder, D.G., Williams, R.W., Holmes, A., 2010. Strain differences in stress responsivity are associated with divergent amygdala gene expression and glutamate-mediated neuronal excitability. The Journal of",
+    "Neurobiology of Learning and Memory 185 (2021) 107509 21.Introduction  James McGaugh was one of the first neuroscientists to point to the  important influence of stress hormones on memory consolidation  (McGaugh, Gold, Van Buskirk, & Haycock, 1975 ). He and others  considered that hormones released by stressful experiences could  enhance memory consolidation, indicating particularly the hormones  epinephrine and glucocorticoids as memory modulators (McGaugh &",
+    "For example, stress is a functional state of psychosocial arousal that focuses and energizes us to confront the stressor, but chronic/toxic levels of stress lead to disruptive changes in brain architecture and dysregulation of stress response mechanisms, such as the hypothalamus-pituitary ( hpA) axis and the autonomic  nervous (ANS) system. Under chronic stress, the adrenal glands of mammals (including humans) release the steroid hormone cortisol. Cortisol acts by increas -",
+    "55:485494.  Herman JP, Ostrander MM, Mueller NK, Figueiredo H (2005). Limbic system  mechanisms of stress regulation: hypothalamo -pituitary -adrenocortical axis. Prog  Neuropsychopharmacol Biol Psychiatry 29:1201 1213.   Herry C, Bach DR, Esposito F, Di Salle F, P errig WJ, Scheffler K et al. (2007).  Processing of temporal unpredictability in human and animal amygdala. J Neurosci  27:5958 5966.   Hitzemann R, Malmanger B, Cooper S, Coulombe S, Reed C, Demarest K et al. (2002).",
+    "after restraint stress. Acute stress (like acute ethanol) activates the HPA axis and increases brain and circulating levels of GABAergic neuroactive steroids [1] as well as corticosterone, the major corticosteroid synthesized in rodents from DOC. GABAergic neuroactive steroids have anxiolytic properties when administered systemically [54,55]. Thus, we might have predicted that those strains with higher basal DOC levels would have been less",
+    "present in the brain as well as in the peripheral circulation. It issynthesized from progesterone, mainly in the adrenal zonafasciculata and it is precursor of both the glucocorticoidcorticosterone and the GABAergic neuroactive steroid (3 a,5a)- 3,21-dihydroxypregnan-20-one (tetrahydrodeoxycorticosterone,THDOC). These steroids are all elevated following acute stress[1] or ethanol administration in rats, and their elevation is blunted",
+    "plasticity and epigenetic regulation as a consequence of stress. Neuropharmacology 62, 3 12. McEwen, B.S., Nasca, C., Gray, J.D., 2016. Stress e ects on neuronal structure: hippo- campus, amygdala, and prefrontal cortex. Neuropsychopharmacology 41, 3 . Mozhui, K., Lu, L., Armstrong, W.E., Williams, R.W., 2012. Sex-speci c modulation of gene expression networks in murine hypothalamus. Front. Neurosci. 6, 63 . Navarro, V.M., 2013. Interactions between kisspeptins and neurokinin B. In: Kisspeptin"
+  ],
+  "answer": "During stressful experiences, our brains release several hormones, including corticotropin-releasing hormone (CRH), adrenocorticotropic hormone (ACTH), and glucocorticoids (GC) [1]. Additionally, cortisol, a steroid hormone, is released by the adrenal glands under chronic stress [6]. These hormones play significant roles in regulating immune responses, neuronal survival, neurogenesis, synaptic plasticity, and behavioral responses [1].",
+  "question": "what hormones do our brains release during stressful experiences?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_32 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_32
new file mode 100644
index 0000000..650d433
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_32
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2015 - Genetic dissection of sleep homeostasis.pdf",
+    "2019 - Leveraging genomics to uncover.pdf",
+    "2013 - Neural-Immune Interactions in Brain Function and Alcohol Related Disorders.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2012 - Genetic regulation of adult hippocampal neurogenesis A systems genetics approach using BXD recombinant inbred mouse strains.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2019 - Strain differences in maternal neuroendocrine and behavioral responses to stress and the relation to offspring cocaine responsiveness..pdf",
+    "2020 - Modeling the Genetic Basis of Individual Differences in Susceptibility to Gulf War Illness.pdf",
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf"
+  ],
+  "extraction_id": [
+    "cef725f8-c326-59f4-a65e-62d8c7bd5db5",
+    "c624519f-327a-5733-9e1e-94d5bec93fd7",
+    "f6556a02-048a-5e9b-ac7e-ed681db96345",
+    "f9be673c-af23-5d15-9087-37e818cf1a68",
+    "3c78be84-90fe-58ce-85e5-e85e2208057f",
+    "59789bd0-1ee6-51da-b2a1-94f847ff6c63",
+    "32902b1c-3a3a-5f5b-b651-a6fd0fa653a9",
+    "29253383-31a5-5fe1-8160-9d6091273a4d",
+    "1de7e365-88d0-5893-826e-7ac6a69b896e",
+    "5da98563-71dd-5d71-8303-b52f2fb8c6a7"
+  ],
+  "document_id": [
+    "ed971d1f-e77e-566b-b549-81cd0038834a",
+    "5da46d3b-fa82-57f6-b3e5-c82784347881",
+    "78271275-3409-5fc7-bbdd-53c484178e0b",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "c54da858-9620-588e-8e41-76a960af2ff6",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "d29d8018-09a1-53d4-8f07-9dd110c79b39",
+    "d235d186-3d1c-5cde-90d5-9c140cd920f4",
+    "17637a6f-804e-50e4-9cf5-37318e17f15c"
+  ],
+  "id": [
+    "chatcmpl-ADZMoelW4EZWflXHaXujPl4dX6GM9",
+    "bf56c010-06d1-598e-81cf-2a2603f0a883",
+    "76804170-ccb4-5e86-b9ba-533264556893",
+    "63c085a5-ad08-5f28-b3be-3e62b7739183",
+    "74ffa8aa-80dc-5e94-a373-c1af483d63f4",
+    "05e15635-52ee-5d80-9696-15cea22fb7e4",
+    "5ccf3333-4675-577f-bfce-5d5e72fd7c3f",
+    "6d2d21e3-a1c5-5a11-a7ca-7fc643cf8b36",
+    "8f5142d0-8efa-5fe8-b7bf-46dea42ec444",
+    "7b2a0384-586f-582f-93da-8fd64dc76095",
+    "2234517f-d2da-535b-8bb4-5ee5d33671e2"
+  ],
+  "contexts": [
+    "that corticosterone importantly amplies the SD induced changes",
+    "be used to predict corticosteroid response [200]. George etal.",
+    "we do not wish to dispute this viewpoint, it is interesting to note that anti- in  ammatory actions of CORT are most pronounced at high and supraphysiological  concentrations, whereas lower concentrations of CORT appear to have some  immune-potentiating effects (e.g.,  [  6  ] ). Whether these low-dose facilitation effects  relate more directly to the timing of CORT injection relative to cytokine measure- ments, or represent differential tissue sensitivity to glucocorticoids, remains to be",
+    "cortisol to the less bioactive cortisone (Seckl,1997 ). While the protection afforded by this bar- rier enzyme can be overwhelmed when cortisol levels get very high, it likely functions effec- tively when cortisol remains within the normalrange (Campbell and Murphy, 1997 ). There is now considerable interest in what types of events or other hormones might lower 11-HSD2 andthereby reduce the buffering benets it affords. On example is elevated catecholamine levels,",
+    "the balance between cell generation and cell death. Acute increase of corticosterone leads to decreased cell proliferation while chronic increase causes an increase in proliferation rate (Sapolsky et al., 2000). This discrepancy is due to the presence of two receptors with different binding affinities: the glucocorticoid receptor (GR) and mineralocorticoid receptor (MR). The GR present in",
+    "corticosterone dramatically reduce the delayed-type hypersensitivity response (Dhabhar andMcEwen, 1997 ,1999 ). Sorrells and Sapolsky (2007 ) have provided a thought provoking recent review, contrasting the well-established anti-inammatory aspect of glucocorticoids, with the mounting evidence for their pro-inammatory effects both in the periphery and in the brain fol-lowing chronic exposure. This pattern of results demonstrates that the acute stress response has",
+    "mature babies in order to stimulate lung maturation. As illustrated here, Dex readily bypasses the protective bar-rier enzyme 11 beta-hydroxysteroid dehydrogenase type2 (11-HSD2), which normally limits fetal exposure tomaternal cortisol by converting it to corticosterone, aless bioactive form. Some concerns linger about long-term effects of fetal exposure to high doses or sustainedcorticosteroid treatmentantenatal glucocorticoid therapies are warranted,",
+    "first session. Approximately 50 microliters was collected into lithium heparin-coated tubes  and then centrifuged for collection of plasma. Samples were stored at 80 degrees until  ready for processing. Plasma corticosterone concentration was measured with the use of the DetectX CORT  Enzyme Immunoassay kit (ArborAssays K014-H5, Ann Arbor, MI, USA). Room  temperature plasma samples were diluted 1:450 in assay buffer and processed according to",
+    "corticosterone in their drinking water as sole liquid source for seven days prior to i.p. treatment with 4 mg /kg diisopropylurophosphate. Data are mean consumption per day s.e.m. 4.2. Gene Expression in Response to Treatments 4.2.1. IL1b We observed wide variability in the e ect of CORT on expression of Il1b (Figure 2). ANOVA revealed a signicant main e ect for strain (F 32,122 =3.61, p<0.001). The main e ects for sex and",
+    "cortisol. Support has emerged for the basic tenets of this hypothesis in studies of both humans and animals (Avitsur et al, 2001 ; Miller and Chen, 2006 ; Miller et al, 2002 ,2009 ; Rohleder et al, 2001 ; Stark et al, 2001 ). Most of this work has made use of an ex vivo assay system, where leukocytes are stimulated with a bacterial prod- uct (endotoxin) in the presence of varying levelsof cortisol, which is expected to suppress their ability to synthesize inammatory cytokines. In"
+  ],
+  "answer": "Corticosterone has several uses and effects as indicated in the provided context:  1. **Amplification of Stress-Induced Changes**: Corticosterone importantly amplifies stress-induced changes [1]. 2. **Anti-inflammatory and Immune Effects**: At high and supraphysiological concentrations, corticosterone has pronounced anti-inflammatory actions. Conversely, at lower concentrations, it appears to have some immune-potentiating effects [3]. 3. **Impact on Cell Proliferation**: Acute increases in corticosterone lead to decreased cell proliferation, while chronic increases cause an increase in proliferation rate. This is due to the presence of two receptors with different binding affinities: the glucocorticoid receptor (GR) and the mineralocorticoid receptor (MR) [5]. 4. **Reduction of Delayed-Type Hypersensitivity Response**: Corticosterone dramatically reduces the delayed-type hypersensitivity response [6]. 5. **Measurement in Research**: Corticosterone levels can be measured in plasma samples using specific enzyme immunoassay kits, which is useful in various research settings [8].  These points highlight the diverse roles and applications of corticosterone in both physiological and research contexts.",
+  "question": "what is the use of corticosterone?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_4 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_4
new file mode 100644
index 0000000..35b4e62
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_4
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2004 - Nutritional genomics.pdf",
+    "2008 - (Infectious Disease) Karl A. Western (auth.), Vassil St. Georgiev PhD, Karl A. Western MD, John J. McGowan PhD (eds.) - National Institute of Allergy and Infectious Diseases, NIH_ Frontiers in Researc (3).pdf",
+    "2008 - Biotools for Determining the Genetics of Susceptibility to Infectious Diseases.pdf",
+    "2006 - Invited Review Microbial ecology in the age of genomics and metagenomics concepts, tools, and recent advances.pdf",
+    "2008 - Molecular profiling in the age of cancer genomics.pdf",
+    "2003 - Molecular profiling in the age.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf"
+  ],
+  "extraction_id": [
+    "713c3d5f-dea2-5d83-a4f9-8749e8c0a9e8",
+    "39d8f5af-dff0-58ab-b16f-f0b25f7ccdf0",
+    "16c8fbb0-ab2a-563f-a6b2-e0d8733b69fb",
+    "fe6eb7f0-9f09-50f8-a7a1-c71e507226d5",
+    "c08e6c0a-19fe-52ae-a715-8241e7b9baf8",
+    "167ddb29-f516-5670-9b89-a5d6c9eb930f",
+    "4c017db4-38d5-5116-b707-57e836fd043b",
+    "512cfd3a-f28e-5e11-8caa-6add0151a824",
+    "564e2e42-51ba-5b46-9375-b9ebee9ceabc",
+    "dd3030ab-0d42-521f-a769-8ee1b2fd0902"
+  ],
+  "document_id": [
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "f97bdb31-0a83-5046-a60c-81d633684557",
+    "4db8c752-c8e2-5f6d-a091-dc4f1d0c48bc",
+    "fcbbb3ce-6524-50e3-9f8d-c191dc551231",
+    "a666073a-eb22-53b0-bc94-550e775e456a",
+    "a3e71525-cef6-50ba-abd1-e3853ddf77fa",
+    "547b71a7-ac5a-52ca-b9db-04391e9e50c6",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0"
+  ],
+  "id": [
+    "chatcmpl-ADZIuGqxbDENLCoABzAMboXc8WUvK",
+    "9b81ec06-a109-5441-93bc-fd3e2d0b7862",
+    "2369f273-6321-5702-9c80-c29ae74935d0",
+    "3aafc5c1-96b5-5629-bc72-b00cf5070536",
+    "6081b16d-3380-5602-9daf-0500940fafbb",
+    "011b9d94-7b34-597c-9b89-74062b999132",
+    "32219bd7-c673-5deb-bb35-3bea4ae9bd3a",
+    "5cebb071-960c-5072-beb7-842815ae89bb",
+    "59772d2c-7eac-5a4b-b9ef-70735afda23e",
+    "aa406f0f-8f39-5189-9131-91345a876489",
+    "f03ab31a-1da3-50f1-9cfe-dc05cee18c05"
+  ],
+  "contexts": [
+    "is the   eld of bioinformatics.",
+    "the umbrella of bioinformatics or com-putational biology.",
+    "methods of computer-based information processing for ana-lyzing the structure and function of biologically important molecules. NCBI bioinformatics-related resources may be accessed through its home page at:   www.ncbi.nlm.nih.gov.          The NCBI has three principal branches:     1.     Computational Biology Branch (  http://www.ncbi.nlm. nih.gov/CBBresearch/)           2.     Information Engineering Branch (  http://www.ncbi.nlm. nih.gov/IEB/)",
+    "methods of computer-based information processing for ana-lyzing the structure and function of biologically important molecules. NCBI bioinformatics-related resources may be accessed through its home page at:   www.ncbi.nlm.nih.gov.          The NCBI has three principal branches:     1.     Computational Biology Branch (  http://www.ncbi.nlm. nih.gov/CBBresearch/)           2.     Information Engineering Branch (  http://www.ncbi.nlm. nih.gov/IEB/)",
+    "been successful in microbial ecological research withoutbioinformatics tools. Broadly defined, bioinformatics refersto the use of computers to seek patterns in the observedbiological data and to propose mechanisms for such patterns.As can be seen from below, bioinformatics not only canhelp us directly address experimental research objectives butalso can integrate information from various sources and seekspatterns not achievable through experimentation alone.",
+    "Since the first protein database was created by Margaret Dayhoffin 1965 in response to the increase in protein sequencing, therehas been an explosion of data from the different modalities. Foreach of the aforementioned levels, bioinformatics plays a crucialand intimate role in each of the steps. In general, there are threelarge categories of bioinformatics applications, including data-bases, algorithms and predictions. The category of databasesallows for the combining and organization of large amounts",
+    "Since the first protein database was created by Margaret Dayhoffin 1965 in response to the increase in protein sequencing, therehas been an explosion of data from the different modalities. Foreach of the aforementioned levels, bioinformatics plays a crucialand intimate role in each of the steps. In general, there are threelarge categories of bioinformatics applications, including data-bases, algorithms and predictions. The category of databasesallows for the combining and organization of large amounts",
+    "remit of the early bioinformaticist.1,2T o address these problems, the eld drew from the foundations of statistics, mathematics, physics, computer science and, of course, molecular biology. T oday, predictably, bioinformatics still reects the broad base on which it started, comprising an eclectic collection of scientic specialists. As a result of its inherent diversity, it is difcult to dene the scope of bioinformatics as a discipline. It may be even fruitless to try to draw hard boundaries around the eld.",
+    "remit of the early bioinformaticist.1,2T o address these problems, the eld drew from the foundations of statistics, mathematics, physics, computer science and, of course, molecular biology. T oday, predictably, bioinformatics still reects the broad base on which it started, comprising an eclectic collection of scientic specialists. As a result of its inherent diversity, it is difcult to dene the scope of bioinformatics as a discipline. It may be even fruitless to try to draw hard boundaries around the eld.",
+    "remit of the early bioinformaticist.1,2T o address these problems, the eld drew from the foundations of statistics, mathematics, physics, computer science and, of course, molecular biology. T oday, predictably, bioinformatics still reects the broad base on which it started, comprising an eclectic collection of scientic specialists. As a result of its inherent diversity, it is difcult to dene the scope of bioinformatics as a discipline. It may be even fruitless to try to draw hard boundaries around the eld."
+  ],
+  "answer": "Bioinformatics broadly refers to the use of computers to seek patterns in observed biological data and to propose mechanisms for such patterns [5]. It involves methods of computer-based information processing for analyzing the structure and function of biologically important molecules [3], [4]. The field draws from the foundations of statistics, mathematics, physics, computer science, and molecular biology, reflecting a broad base of scientific specialties [8], [9], [10].",
+  "question": "what is bioinformatics"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_5 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_5
new file mode 100644
index 0000000..962ae82
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_5
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2006 - Positional cloning of genes contributing to variability in nociceptive and analgesic phenotypes.pdf",
+    "2007 - QTL Mapping in Aging Systems.pdf",
+    "2005 -Knott- Regression based QTL mapping.pdf",
+    "2005 - Regression-based quantitative trait loci mapping robust, efficient and effective.pdf",
+    "2005 - Regression-based quantitative trait loci mapping robust, efficient and effective.pdf",
+    "2005 -Knott- Regression based QTL mapping.pdf",
+    "2007 - Using quantitative trait loci analysis to select plants for altered radionuclide accumulation.pdf",
+    "2008 - Genotype-phenotype relationships and the patterning of complex traits as exemplified in the mammalian dentition.pdf",
+    "2019 - Novel Genetic Loci Control L5 Vertebral Trabecular Bone and the Response to Low Calcium Intake in Growing BXD Recombinant Inbred Mice.pdf",
+    "2012 - Teaching Neuroinformatics with an Emphasis on Quantitative Locus Anlaysis.pdf"
+  ],
+  "extraction_id": [
+    "c2c33142-b1dc-5162-a2a1-b452d2385958",
+    "ace8317f-2e7a-5590-a8e6-5e961480c0fb",
+    "e12f12c8-b1e0-54fa-86f8-0bcdb580bca1",
+    "e8203703-d34a-5848-bf54-4e20eb6fc3c5",
+    "75b53145-3938-5fbe-9cca-0389a68e1955",
+    "26dd8d34-b134-5426-b717-61b8a3a0f752",
+    "9ca9216b-e4cb-52c2-a286-f7d5d37936b6",
+    "b672f393-c45d-5393-96ee-77934e21e9c3",
+    "92e2d87b-02c9-588b-bc3c-e1034c05826d",
+    "0184b980-f596-51d9-a1a5-dd9c8d4ba388"
+  ],
+  "document_id": [
+    "8ba88825-7473-52f8-8a1d-27f25644c4a2",
+    "35fbcd3c-97e8-57e5-b4c9-08dfbd4bce2e",
+    "cd41c63b-e5c2-5040-bbc5-ab20925b7d17",
+    "ba67a5b2-3dc7-57dc-8f8b-2d01433e58c2",
+    "ba67a5b2-3dc7-57dc-8f8b-2d01433e58c2",
+    "cd41c63b-e5c2-5040-bbc5-ab20925b7d17",
+    "682e6f43-10d4-5772-a69a-26e774606ba7",
+    "f6e866b8-b233-5862-bfb8-9949d0dabb97",
+    "de8dda5e-0e2f-5aa9-bb13-851c526b36a5",
+    "f36cbb2c-90f3-5544-8ce8-52b2004f6b49"
+  ],
+  "id": [
+    "chatcmpl-ADZIye9JJrA436MgjlTpeY9z4NFZS",
+    "1ec396e1-0218-5f22-8db7-8653770944fb",
+    "e3149a33-9780-5f50-b582-142cdae5a5d3",
+    "ef0bab2a-db4a-57ac-9f75-32ec8c4a8f87",
+    "62ec26e1-3c71-558d-9378-e920e47edb08",
+    "5b07b911-a624-52ed-8506-ab14cb16a2eb",
+    "297470d7-ce20-5685-af94-a8ed5c68386b",
+    "543c9c0c-e8f5-59d8-b1e0-22172ace332e",
+    "1aa1e57d-cced-59d2-ac5b-9f3be7be2355",
+    "adbe8575-3c00-53e6-bb98-e86b8d01c7c5",
+    "22a5b128-d4d2-5fad-a60a-162c1d9a3369"
+  ],
+  "contexts": [
+    "(although quite demanding) process offollowing the trait across multiple generations by  tracing its coinheritance with genetic markers (a technique referred to as linkage  mapping). Finding loci responsible for variability in a quantitative trait (quantitative trait  locus mapping, or QTL mapping) is much more difficult, as there are many more sources  of variation to capture. lnbred mouse strains are the optimum starting point for QTL",
+    "Genetic linkage analysis can be used to identify regions of the genome that contain genes that predispose to the observed quantitative trait, leading to iden-tification of QTLs. A significant QTL means that different genotypes at a poly-morphic marker locus are associated with different trait values. Linkage isdetermined by the log of odds (LOD) scores or likelihood ratio statistics (LRS)(seeNote 1 ). To calculate a LOD score or an LRS score for a selected quanti-",
+    "quantitative trait loci in crosses between outbred linesusing least squares. Genetics 136, 11951207. Haseman, J. K. & Elston, R. C. 1972 The investigation of linkage between a quantitative trait and a marker locus.Behav. Genet. 2, 319. Henshall, J. M. & Goddard, M. E. 1999 Multiple trait mapping of quantitative trait loci after selective genotypingusing logistic regression. Genetics 151, 885894. Jansen, R. C. 1993 Interval mapping of multiple quantitative trait loci. Genetics 135, 205211.",
+    "quantitative trait loci in crosses between outbred linesusing least squares. Genetics 136, 11951207. Haseman, J. K. & Elston, R. C. 1972 The investigation of linkage between a quantitative trait and a marker locus.Behav. Genet. 2, 319. Henshall, J. M. & Goddard, M. E. 1999 Multiple trait mapping of quantitative trait loci after selective genotypingusing logistic regression. Genetics 151, 885894. Jansen, R. C. 1993 Interval mapping of multiple quantitative trait loci. Genetics 135, 205211.",
+    "Keywords: quantitative trait loci mapping; regression; structured outbred populations 1. HISTORY The idea of using markers associated with a trait of interest, for example, to predict the performance of individuals in the trait, is not new. Initially, however, the markers used were not identied at the molecular level but rather through the phenotype, for example, coat colour or by the use of simple biochemicalprocedures such as blood groups. An early implemen-",
+    "Keywords: quantitative trait loci mapping; regression; structured outbred populations 1. HISTORY The idea of using markers associated with a trait of interest, for example, to predict the performance of individuals in the trait, is not new. Initially, however, the markers used were not identied at the molecular level but rather through the phenotype, for example, coat colour or by the use of simple biochemicalprocedures such as blood groups. An early implemen-",
+    "tions between markers and phenotype. Once allelic effects at each locus are identified, different techniques can be used to position precise loci (i.e., QTL) influencing the trait. These techniques include marker regression (30), interval mapping (31), and multiple mapping strategies (32). Marker regression locates QTL with respect to all markers simultaneously by regression onto the marker means. It also estimates the additive (and dominance) effects, tests their signif-",
+    "successful in identifying genes for simple traits.  Quantitative trait mapping and genome wide  association studies identify chromosomal regions referred to as quantitative trait loci (QTLs) that  are statistically associated with the trait.  Usually there are several such associations, each on the  order of megabases (Mb) in length containing the usual diversity of single nucleotide  polymorphisms (SNPs), one to two thousand per Mb, and there has been little success identifying",
+    "markers reveal potential gene locations regulating the trait of interest as known as quant itative  trait loci  (QTL s). Historically, this approach has been successful in identifying genes that are  responsible for rare, monogenic bone diseases. More recently, much denser maps of SNP s allow  researchers  to perform genome -wide linkage analysis for  complex trait s like bone phenotypes .  However, several difficulties preventing the discovery of causal genes include genetic",
+    "Quantitative Trait Locus (QTL) analysis, which links  phenotype to loci on chromosomes that likely had an  impact on the phenotype.  Students then are able to sift  through a list of genes in the region(s) of the chromosome  identified by the QT L analysis and find a candidate gene  that has relatively high expression in the brain region of  interest.  Once such a candidate gene is identified,  students can find out more information about the gene,"
+  ],
+  "answer": "The process of finding a genetic marker followed by a quantitative trait loci (QTL) involves several steps:  1. **Linkage Mapping**: The initial step is to trace the trait across multiple generations by examining its coinheritance with genetic markers. This technique is known as linkage mapping [1]. Genetic linkage analysis is used to identify regions of the genome that contain genes predisposing to the observed quantitative trait [2].  2. **Identification of QTLs**: Once regions of interest are identified through linkage mapping, the next step is to pinpoint the loci responsible for variability in the quantitative trait. This process is known as quantitative trait locus mapping or QTL mapping [1]. QTL mapping is more challenging due to the numerous sources of variation that need to be captured [1].  3. **Statistical Analysis**: To determine linkage, statistical methods such as the log of odds (LOD) scores or likelihood ratio statistics (LRS) are used [2]. A significant QTL indicates that different genotypes at a polymorphic marker locus are associated with different trait values [2].  4. **Techniques for Positioning QTLs**: Various techniques can be employed to position precise loci influencing the trait. These include marker regression, interval mapping, and multiple mapping strategies [7]. Marker regression, for instance, locates QTLs with respect to all markers simultaneously by regression onto the marker means and estimates the additive and dominance effects [7].  5. **Genome-Wide Association Studies (GWAS)**: In some cases, genome-wide association studies are conducted to identify chromosomal regions referred to as QTLs that are statistically associated with the trait [8]. These regions often contain a high density of single nucleotide polymorphisms (SNPs) [8].  6. **Candidate Gene Identification**: After identifying the QTL regions, researchers can sift through the list of genes in these regions to find candidate genes that may have a significant impact on the phenotype [10].  By following these steps, researchers can effectively identify genetic markers and subsequently map quantitative trait loci, providing insights into the genetic basis of complex traits.",
+  "question": "Explain the process of finding a genetic marker followed by a quantitative trait loci."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_6 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_6
new file mode 100644
index 0000000..45c7db9
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_6
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Germline de novo mutation clusters arise.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Germline de novo mutation clusters arise.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf"
+  ],
+  "extraction_id": [
+    "af805fbb-a39f-5a29-a0b0-9add1126b553",
+    "5f52d45a-991b-54c3-92ae-37dd96e31a42",
+    "5f52d45a-991b-54c3-92ae-37dd96e31a42",
+    "403bbc25-ce94-5a4f-a409-436cc02fb204",
+    "5f52d45a-991b-54c3-92ae-37dd96e31a42",
+    "907c33dd-34b8-51f5-a91f-fb83cf11f7f9",
+    "403bbc25-ce94-5a4f-a409-436cc02fb204",
+    "5f52d45a-991b-54c3-92ae-37dd96e31a42",
+    "c07e5efe-7d80-547e-847b-eef61bb661cc",
+    "8cba1054-1540-57ee-a5c4-350f5555081f"
+  ],
+  "document_id": [
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "f2b2ca83-a34f-5f99-b9f2-357b2ddbe136",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "f2b2ca83-a34f-5f99-b9f2-357b2ddbe136",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858"
+  ],
+  "id": [
+    "chatcmpl-ADZJ5u5h9f6SgdrxrixAsqUmOQgLr",
+    "72da6034-227d-5dac-9ef6-90c246ec2b40",
+    "66e5e009-5496-5e18-bfbe-9a9567cad60c",
+    "2f2342b3-4c07-5bfd-80c6-8bc47fead6b6",
+    "ab92961e-c267-5e56-aeb9-0d03fd0a4102",
+    "fb421292-e4ea-510b-8a69-48e12e6e6a43",
+    "3b5635bb-8308-5c6b-8ee0-d65293257362",
+    "788b6b85-7ef2-5805-bc0c-d8af71332e0d",
+    "4802fb82-204d-57b6-b24f-5683f3731aea",
+    "c8e7e683-487f-5075-bbef-126ca0203c6c",
+    "5da6f433-231d-586b-a057-558a4c68f741"
+  ],
+  "contexts": [
+    "Genes 2018 ,9, 615 18 of 20 97. McFarlane, R.J.; Humphrey, T.C. A role for recombination in centromere function. Trends Genet. 2010 ,26, 209213. [CrossRef] 98. Talbert, P .B.; Henikoff, S. Centromeres convert but dont cross. PLoS Biol. 2010 ,8, e1000326. [CrossRef] 99. Durfy, S.J.; Willard, H.F. Concerted Evolution of Primate Alpha Satellite DNA Evidence for an Ancestral Sequence Shared by Gorilla and Human X Chromosome Satellite. J. Mol. Biol. 1990 ,216, 555566. [CrossRef]",
+    "4.1. Recombination and Repair at Centromeres: Errors in Copying and Mending Highly Repetitive DNA Why are centromeres so cold?, asked Andy Choo in his review of centromeres [ 96]. He was referring to centromere DNA as being cold to recombination. While maternal and paternal chromosomes suffer multiple DNA double-stranded breaks (DSBs) to induce recombination and exchange of genetic information by crossing over during meiosis, centromere loci are refractory",
+    "exacerbates centromere rearrangements [ 54], indicating that there may be active mechanisms to suppress centromeric recombination and these may, at least in part, involve core centromeric proteins. Centromere alpha-satellite DNA is estimated to represent between 3% and 10% of the human genome [ 101], reviewed in [ 19]. During each round of replication, unperturbed cells suffer over 40 DNA DSBs [ 102], of which at least half are repaired by homologous recombination (HR) in S-phase and G2,",
+    "347357 (1998).  31. Baudat, F. et al. PRDM9 is a major determinant of meiotic recombination hotspots in humans and mice. Science  327, 836840 (2010).  32. Kong, A. et al. Recombination rate and reproductive success in humans. Nat.Genet.  36, 12031206 (2004).  33. Ottolini, C. S. et al. Genome-wide maps of recombination and chromosome segregation in human oocytes and embryos show selection for maternal recombination rates. Nat. Genet.  47, 727735 (2015).",
+    "to this process. This led to the assumption that centromeres do not undergo recombination and that the repetitive arrays are maintained as stable. However, this clashed with the notion that centromeres very origin stems from recombination to create the repetitive array, where multiple short- and long-range recombination events may be responsible for the generation and reiteration of blocks of highly homogenized alpha-satellite DNA throughout the centromere [ 97,98]. Furthermore, in addition",
+    "of these DSBs through recombination-dependent pathways, such as homologous recombination (HR),  may disrupt centromere integrity in several ways: (1) Crossover between sister chromatids will lead  to sister chromatid exchange (SCE), which has been  reported at human cent romeres. (2) Search for  the homologous sequence may erroneously identify an  identical or nearly identical sequence within  the same chromatid downstream or upstream of the break site. Recombination between these two",
+    "higher in regions of high recombination. Trends Genet.  18, 337340 (2002). 26. Webster, M. T. & Hurst, L. D. Direct and indirect consequences of meiotic recombination: implications for genome evolution. Trends Genet.  28,  101109 (2012).  27. Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. Nature  500, 415421 (2013).",
+    "to chromosome-specic alpha-satellites, certain centromeric sequences are shared by all chromosomes, evidence that formation of these arrays is dominated by interchromosomal exchanges [ 8,98100]. This invites new questions about the stability of centromere DNA outside of meiosis. Indeed, our recent analysis has shown that centromeres can undergo recombination during a single round of cell division in primary human cells [ 54]. Depletion of CENP-A and other CCAN proteins",
+    "shown to play a role in DNA repair (reviewed in [ 125]), and in vitro experiments show that this hybridization may facilitate DSB repair by bridging the broken DNA fragments in a Rad52-dependent manner during recombination [126]. Centromeres have been suggested [ 127,128], but not proven, to behave like fragile sites of the human genome. Common fragile sites are described as genomic loci where ongoing replication collides",
+    "Cell Biol. 2016 ,17, 1629. [CrossRef] [PubMed] 54. Giunta, S.; Funabiki, H. Integrity of the human centromere DNA repeats is protected by CENP-A, CENP-C, and CENP-T. Proc. Natl. Acad. Sci. USA 2017 ,114, 19281933. [CrossRef] [PubMed] 55. Giunta, S. Centromere Chromosome Orientation Fluorescent in situ Hybridization (Cen-CO-FISH) Detects Sister Chromatid Exchange at the Centromere in Human Cells. Bio-Protocol 2018 ,8. [CrossRef]"
+  ],
+  "answer": "Recombination in human centromeres is generally suppressed, making these regions \"cold\" to recombination events [2]. This suppression is thought to be actively maintained by core centromeric proteins [3]. Despite this, there is evidence suggesting that centromeres can undergo recombination, particularly through mechanisms that involve homologous recombination (HR) during DNA repair processes [6]. For instance, crossover between sister chromatids, known as sister chromatid exchange (SCE), has been reported at human centromeres [6]. Additionally, centromeres may experience recombination during a single round of cell division in primary human cells [8]. This indicates that while recombination is generally suppressed in centromeres, it can still occur under certain conditions, potentially impacting centromere integrity and stability [5], [6], [8].",
+  "question": "What about recombination in human centromeres?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_7 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_7
new file mode 100644
index 0000000..5d05576
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_7
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Germline de novo mutation clusters arise.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2017 - Human female meiosis revised new.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf"
+  ],
+  "extraction_id": [
+    "5f52d45a-991b-54c3-92ae-37dd96e31a42",
+    "af805fbb-a39f-5a29-a0b0-9add1126b553",
+    "907c33dd-34b8-51f5-a91f-fb83cf11f7f9",
+    "5f52d45a-991b-54c3-92ae-37dd96e31a42",
+    "5f52d45a-991b-54c3-92ae-37dd96e31a42",
+    "403bbc25-ce94-5a4f-a409-436cc02fb204",
+    "c07e5efe-7d80-547e-847b-eef61bb661cc",
+    "5f52d45a-991b-54c3-92ae-37dd96e31a42",
+    "8e3ed969-da49-5ba4-8382-dc2714b01497",
+    "8cba1054-1540-57ee-a5c4-350f5555081f"
+  ],
+  "document_id": [
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "f2b2ca83-a34f-5f99-b9f2-357b2ddbe136",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "95817342-4fb7-543f-beab-f3d7d0a1dd80",
+    "262df0d6-ad68-544a-88ed-b4568f305858"
+  ],
+  "id": [
+    "chatcmpl-ADZJEypA9e5sRA4lUD0c4IqRsBYeu",
+    "66e5e009-5496-5e18-bfbe-9a9567cad60c",
+    "72da6034-227d-5dac-9ef6-90c246ec2b40",
+    "3b5635bb-8308-5c6b-8ee0-d65293257362",
+    "2f2342b3-4c07-5bfd-80c6-8bc47fead6b6",
+    "fb421292-e4ea-510b-8a69-48e12e6e6a43",
+    "ab92961e-c267-5e56-aeb9-0d03fd0a4102",
+    "c8e7e683-487f-5075-bbef-126ca0203c6c",
+    "4802fb82-204d-57b6-b24f-5683f3731aea",
+    "dfa6d21d-2407-5738-84df-95b68469c263",
+    "5da6f433-231d-586b-a057-558a4c68f741"
+  ],
+  "contexts": [
+    "4.1. Recombination and Repair at Centromeres: Errors in Copying and Mending Highly Repetitive DNA Why are centromeres so cold?, asked Andy Choo in his review of centromeres [ 96]. He was referring to centromere DNA as being cold to recombination. While maternal and paternal chromosomes suffer multiple DNA double-stranded breaks (DSBs) to induce recombination and exchange of genetic information by crossing over during meiosis, centromere loci are refractory",
+    "Genes 2018 ,9, 615 18 of 20 97. McFarlane, R.J.; Humphrey, T.C. A role for recombination in centromere function. Trends Genet. 2010 ,26, 209213. [CrossRef] 98. Talbert, P .B.; Henikoff, S. Centromeres convert but dont cross. PLoS Biol. 2010 ,8, e1000326. [CrossRef] 99. Durfy, S.J.; Willard, H.F. Concerted Evolution of Primate Alpha Satellite DNA Evidence for an Ancestral Sequence Shared by Gorilla and Human X Chromosome Satellite. J. Mol. Biol. 1990 ,216, 555566. [CrossRef]",
+    "of these DSBs through recombination-dependent pathways, such as homologous recombination (HR),  may disrupt centromere integrity in several ways: (1) Crossover between sister chromatids will lead  to sister chromatid exchange (SCE), which has been  reported at human cent romeres. (2) Search for  the homologous sequence may erroneously identify an  identical or nearly identical sequence within  the same chromatid downstream or upstream of the break site. Recombination between these two",
+    "exacerbates centromere rearrangements [ 54], indicating that there may be active mechanisms to suppress centromeric recombination and these may, at least in part, involve core centromeric proteins. Centromere alpha-satellite DNA is estimated to represent between 3% and 10% of the human genome [ 101], reviewed in [ 19]. During each round of replication, unperturbed cells suffer over 40 DNA DSBs [ 102], of which at least half are repaired by homologous recombination (HR) in S-phase and G2,",
+    "to this process. This led to the assumption that centromeres do not undergo recombination and that the repetitive arrays are maintained as stable. However, this clashed with the notion that centromeres very origin stems from recombination to create the repetitive array, where multiple short- and long-range recombination events may be responsible for the generation and reiteration of blocks of highly homogenized alpha-satellite DNA throughout the centromere [ 97,98]. Furthermore, in addition",
+    "347357 (1998).  31. Baudat, F. et al. PRDM9 is a major determinant of meiotic recombination hotspots in humans and mice. Science  327, 836840 (2010).  32. Kong, A. et al. Recombination rate and reproductive success in humans. Nat.Genet.  36, 12031206 (2004).  33. Ottolini, C. S. et al. Genome-wide maps of recombination and chromosome segregation in human oocytes and embryos show selection for maternal recombination rates. Nat. Genet.  47, 727735 (2015).",
+    "shown to play a role in DNA repair (reviewed in [ 125]), and in vitro experiments show that this hybridization may facilitate DSB repair by bridging the broken DNA fragments in a Rad52-dependent manner during recombination [126]. Centromeres have been suggested [ 127,128], but not proven, to behave like fragile sites of the human genome. Common fragile sites are described as genomic loci where ongoing replication collides",
+    "to chromosome-specic alpha-satellites, certain centromeric sequences are shared by all chromosomes, evidence that formation of these arrays is dominated by interchromosomal exchanges [ 8,98100]. This invites new questions about the stability of centromere DNA outside of meiosis. Indeed, our recent analysis has shown that centromeres can undergo recombination during a single round of cell division in primary human cells [ 54]. Depletion of CENP-A and other CCAN proteins",
+    "Studying the direct link between re combination and sister chromatid dynamics with combined live cell imaging and genomics will likely yieldimportant insight into the impact that centromeric and telomeric cross- overs have on chromosome segregation.Reconstructing the bivalent con guration from MeioMaps: recombination and its linkwith chromosome segregation The combined assessment of haplotypes that are determined by recombination also allowed the rst direct correlations between",
+    "Cell Biol. 2016 ,17, 1629. [CrossRef] [PubMed] 54. Giunta, S.; Funabiki, H. Integrity of the human centromere DNA repeats is protected by CENP-A, CENP-C, and CENP-T. Proc. Natl. Acad. Sci. USA 2017 ,114, 19281933. [CrossRef] [PubMed] 55. Giunta, S. Centromere Chromosome Orientation Fluorescent in situ Hybridization (Cen-CO-FISH) Detects Sister Chromatid Exchange at the Centromere in Human Cells. Bio-Protocol 2018 ,8. [CrossRef]"
+  ],
+  "answer": "Recombination in human centromeres is a complex process that is generally suppressed to maintain centromere integrity. Here are the key points about how recombination works in human centromeres:  1. **Recombination Suppression**: Centromere DNA is often referred to as being \"cold\" to recombination, meaning that it is generally refractory to the recombination processes that occur during meiosis in other parts of the genome [1].  2. **Potential Disruptions**: When recombination does occur, it can disrupt centromere integrity in several ways. For instance, crossover between sister chromatids can lead to sister chromatid exchange (SCE), which has been reported at human centromeres. Additionally, the search for homologous sequences may erroneously identify nearly identical sequences within the same chromatid, leading to inappropriate recombination events [3].  3. **Mechanisms to Suppress Recombination**: There are active mechanisms to suppress recombination at centromeres, which may involve core centromeric proteins. This suppression is crucial because recombination-dependent pathways, such as homologous recombination (HR), could otherwise exacerbate centromere rearrangements [4].  4. **Recombination in Centromere Formation**: Despite the general suppression, recombination is believed to play a role in the origin and formation of centromeres. Multiple short- and long-range recombination events may be responsible for generating and reiterating blocks of highly homogenized alpha-satellite DNA throughout the centromere [5].  5. **Recombination Outside Meiosis**: Recent analyses have shown that centromeres can undergo recombination during a single round of cell division in primary human cells, indicating that recombination at centromeres is not entirely restricted to meiosis [8].  In summary, while recombination at human centromeres is generally suppressed to maintain their structural integrity, it can occur under certain conditions and may play a role in the formation and evolution of centromeric DNA.",
+  "question": "How does recombination work in human centromeres?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_8 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_8
new file mode 100644
index 0000000..6e73b0c
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_8
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2018 - Germline de novo mutation clusters arise.pdf",
+    "2008 - Mechanisms of Disease genetic insights into the etiology of type 2 diabetes and obesity.pdf",
+    "2018 - Germline de novo mutation clusters arise.pdf",
+    "2008 - Loci Related to Metabolic-Syndrome Pathways Including LEPR.pdf",
+    "2003 - Haplotypes and the systematic analysis of genetic variation in genes and genomes.pdf",
+    "2018 - Repetitive Fragile Sites Centromere Satellite DNA.pdf",
+    "2003 - Haplotypes and the systematic analysis of genetic variation in genes and genomes.pdf",
+    "2003 - Haplotypes and the systematic analysis of genetic variation in genes and genomes.pdf",
+    "2020 - Prospective avenues for human population genomics and disease mapping in southern Africa.pdf",
+    "2016 - A genetic method for dating ancient genomes provides.pdf"
+  ],
+  "extraction_id": [
+    "403bbc25-ce94-5a4f-a409-436cc02fb204",
+    "0fa3ac68-ea06-5d95-b3fb-f224d40e38a9",
+    "403bbc25-ce94-5a4f-a409-436cc02fb204",
+    "74f21fa4-31ff-5aa6-b806-1ffc73b79801",
+    "de271b3e-86e8-5405-8e15-a54376db728b",
+    "af805fbb-a39f-5a29-a0b0-9add1126b553",
+    "e764c7b0-e155-5358-a5c9-a168508a32ea",
+    "de271b3e-86e8-5405-8e15-a54376db728b",
+    "4927bfe6-f007-5ad0-88c8-f9ae4bde540f",
+    "fcf5296e-6be4-5789-b1e1-ac57fef15119"
+  ],
+  "document_id": [
+    "f2b2ca83-a34f-5f99-b9f2-357b2ddbe136",
+    "87ceda8f-0ce6-5678-9ade-96a40a991647",
+    "f2b2ca83-a34f-5f99-b9f2-357b2ddbe136",
+    "c7583131-8c05-576b-a458-577021374b5d",
+    "97525d6c-b50d-5826-84eb-71ddd52aea27",
+    "262df0d6-ad68-544a-88ed-b4568f305858",
+    "97525d6c-b50d-5826-84eb-71ddd52aea27",
+    "97525d6c-b50d-5826-84eb-71ddd52aea27",
+    "e74e3b9c-14d2-54c3-9319-7dc1b45db41e",
+    "5a5e67ea-4830-5fe8-95c3-ccfcc8324036"
+  ],
+  "id": [
+    "chatcmpl-ADZJSm9AzU2UQT9a71JIfAJoKh0Fh",
+    "ab92961e-c267-5e56-aeb9-0d03fd0a4102",
+    "46297c04-59ef-5f94-af20-5e83036b5ea9",
+    "788b6b85-7ef2-5805-bc0c-d8af71332e0d",
+    "9df97195-cdb6-5271-8dd2-89a421f6281a",
+    "94686ace-46ce-51f1-9b26-07c27baca6b9",
+    "72da6034-227d-5dac-9ef6-90c246ec2b40",
+    "9ee9c9e1-70ed-512a-bd20-9f967829f75a",
+    "acfd48ac-6d04-5691-b2b7-6ebe179c0f0b",
+    "4bcb02fe-e0e1-5e2b-b0c7-7d27bb03b73c",
+    "aa3c4d11-71cb-5941-a6b0-56f9358ba565"
+  ],
+  "contexts": [
+    "347357 (1998).  31. Baudat, F. et al. PRDM9 is a major determinant of meiotic recombination hotspots in humans and mice. Science  327, 836840 (2010).  32. Kong, A. et al. Recombination rate and reproductive success in humans. Nat.Genet.  36, 12031206 (2004).  33. Ottolini, C. S. et al. Genome-wide maps of recombination and chromosome segregation in human oocytes and embryos show selection for maternal recombination rates. Nat. Genet.  47, 727735 (2015).",
+    "Genet  39: 977983 33 Myers S et al. (2005) A fine-scale map of recombination  rates and hotspots across the human genome.  Science  310: 321324REVIEW Nature.indt   1 Nature.indt   1 28/11/07   9:46:50 am 28/11/07   9:46:50 am",
+    "higher in regions of high recombination. Trends Genet.  18, 337340 (2002). 26. Webster, M. T. & Hurst, L. D. Direct and indirect consequences of meiotic recombination: implications for genome evolution. Trends Genet.  28,  101109 (2012).  27. Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. Nature  500, 415421 (2013).",
+    "D.R., and Donnelly, P. (2004). The ne-scale structure ofrecombination rate variation in the human genome. Science 304, 581584. 33. Winckler, W., Myers, S.R., Richter, D.J., Onofrio, R.C., McDo- nald, G.J., Bontrop, R.E., McVean, G.A., Gabriel, S.B., Reich, D., Donnelly, P., et al. (2005). Comparison of ne-scale recom- bination rates in humans and chimpanzees. Science 308, 107111. 1192 The American Journal of Human Genetics 82, 11851192, May 2008",
+    "www.pharmaco-genomics.com 569REVIEW 48. Reich DE, Schaffner SF , Daly MJ  et al. :  Human chromosome sequence variation and the influence of gene history, mutation  and recombination. Nat. Genet.  32, 135-142  (2002).   The authors provide evidence that  recombination hot spots may represent a  general feature of the human genome and play a major role in shaping genetic  variation in humans. 49. Wall JD, Pritchard JK: Haplotype blocks  and linkage disequilibrium in the human",
+    "Genes 2018 ,9, 615 18 of 20 97. McFarlane, R.J.; Humphrey, T.C. A role for recombination in centromere function. Trends Genet. 2010 ,26, 209213. [CrossRef] 98. Talbert, P .B.; Henikoff, S. Centromeres convert but dont cross. PLoS Biol. 2010 ,8, e1000326. [CrossRef] 99. Durfy, S.J.; Willard, H.F. Concerted Evolution of Primate Alpha Satellite DNA Evidence for an Ancestral Sequence Shared by Gorilla and Human X Chromosome Satellite. J. Mol. Biol. 1990 ,216, 555566. [CrossRef]",
+    "Variations on a theme: cataloguing human  DNA sequence variation. Science  278, 1580- 1581 (1997). 37. Jeffreys AJ, Kauppi L, Neumann R: Intensely punctate meiotic recombination in the class II region of the major  histocompatibility complex. Nat. Genet.  29,  217-222 (2001). 38. Chakravarti A, Buetow KH, Antonarakis SE   et al.: Nonuniform recombination within  the human beta-globin gene cluster. Am. J.  Hum. Genet.  36, 1239-1258 (1984). 39. Smith RA, Ho PJ, Clegg JB, Kidd, JR,",
+    "genome. Nat. Rev. Genet.  4, 587-597  (2003).   Important review, including discussion of  the recently proposed haplotype-block  model of LD. 50. Nachman MW: Variation in recombination  rate across the genome: evidence and  implications. Curr. Opin. Genet. Dev.  12,  657-663 (2002). 51. Kong A, Gudbjartsson DF , Sainz J  et al. : A  high-resolution recombination map of the human genome. Nat. Genet.  31, 241-247  (2002). 52. Sabeti PC, Reich DE, Higgins JM  et al. :",
+    "Recombination maps are often used for admixture mapping (Browning and Browning 2007). A recombination map is a genetic map that illustrates the variation of the recombina-tion rate across a region of the genome or the entire genome (Myers etal. 2005). It is dependent on the underlying dis-tribution of recombination events that occur between suc-cessive generations within a given population (Kong etal. 2010). The presence and activity of the PRDM9 zinc finger protein in the population under study, the ratio",
+    "31. Fu Q, et al. (2015) An early modern human from Romania with a recent Neanderthal ancestor. Nature 524(7564):216 219. 32. Baudat F, et al. (2010) PRDM9 is a major determinant of meiotic recombination hotspots in humans and mice. Science 327(5967):836 840. 33. Lesecque Y, Glmin S, Lartillot N, Mouchiroud D, Duret L (2014) The red queen model of recombination hotspots evolution in the light of archaic and modern human ge- nomes. PLoS Genet 10(11):e1004790."
+  ],
+  "answer": "Recombination in the human genome is a complex process with several key aspects highlighted in the provided context:  1. **PRDM9 and Recombination Hotspots**: PRDM9 is identified as a major determinant of meiotic recombination hotspots in humans and mice [1], [10]. This protein plays a crucial role in determining where recombination events are likely to occur during meiosis.  2. **Recombination Rate and Reproductive Success**: There is a relationship between recombination rate and reproductive success in humans, suggesting that recombination may have evolutionary implications [1].  3. **Genome-wide Maps of Recombination**: Detailed maps of recombination rates and hotspots across the human genome have been created, providing insights into the fine-scale structure of recombination rate variation [2], [4]. These maps are essential for understanding genetic variation and are used in admixture mapping [9].  4. **Variation in Recombination Rates**: There is significant variation in recombination rates across different regions of the human genome, which has implications for genetic diversity and evolution [8].  5. **Influence on Genetic Variation**: Recombination hotspots are a general feature of the human genome and play a major role in shaping genetic variation [5].  6. **Selection for Maternal Recombination Rates**: Genome-wide maps of recombination in human oocytes and embryos show selection for maternal recombination rates, indicating that recombination patterns can be subject to evolutionary pressures [1].  These points collectively highlight the importance of recombination in shaping the human genome, influencing genetic diversity, and having evolutionary consequences.",
+  "question": "What about recombination in the human genome?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_9 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_9
new file mode 100644
index 0000000..764ad85
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_cs_gn_9
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2016 - Coming of age ten years of next.pdf",
+    "2020 - Precision and Personalized Medicine How Genomic.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2014 - Computational tools to aid the design and development of a genetic reference population.pdf",
+    "2015 -Pandey- Functional Analysis of Genomic Variation and Impact on Molecular.pdf",
+    "2015 - Functional Analysis of Genomic Variation and Impact on Molecular and Higher Order Phenotypes.pdf"
+  ],
+  "extraction_id": [
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "de09f30d-e9ba-5379-8c7a-85b2cd2ed6c8",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "c14d1c74-a14a-5037-8d3f-f32a60faa9a5",
+    "a744f8ce-7920-5fb8-acce-912f70112924",
+    "f4e989e5-c3d0-5d5c-b8c3-95894a14630b",
+    "fa426831-7c04-56c1-a191-1ebbc35342ed"
+  ],
+  "document_id": [
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "cd11028a-933b-52a0-9534-c173323056ef",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "70cbde25-6406-5a31-91ae-57f430e8f267",
+    "93381e23-494b-5bc2-9d09-ef315506601f",
+    "263d327b-f5db-54e4-a215-b3f8a51cd7d6"
+  ],
+  "id": [
+    "chatcmpl-ADZJdRAJz1ZAYluBcJm58tFEyX9Ng",
+    "98010acc-fd11-5d33-bced-626ef29f2896",
+    "8477a774-dddb-5541-b8d7-d51a7e56b0af",
+    "d2540614-9397-5e3e-8b5f-ad328ca973b2",
+    "13a6b6f9-4a9a-5eb9-ac79-d986d9e613f0",
+    "97f2aa12-623b-53ec-9793-5834311a37dd",
+    "1e324977-2ca5-5062-8a09-7659d516e899",
+    "3e782f01-a06e-51b6-ac8a-0e0a56939d08",
+    "da667832-cd2f-5af6-a0a8-a17542b0a2e2",
+    "d1158643-3625-5855-a03d-eec4ac96eb4d",
+    "cdf2b80f-1509-50a2-9cb2-a36dd6f3f2cc"
+  ],
+  "contexts": [
+    "FURTHER INFORMATION 10X Genomics: http://www.10xgenomics.com 454 Sequencing: http://www.454.com Advances in Genome Biology and Technology (AGBT):  http://www.agbt.org BGISEQ500: http://seq500.com/en/portal/Sequencer.shtml Illumina: http://www.illumina.com Ion Torrent: https://www.thermofisher.com/us/en/home/ brands/ion-torrent.html Oxford Nanopore Technologies: https://www.nanoporetech. com Pacific Biosciences: http://www.pacb.com Personal Genome Project: http://www.personalgenomes.org",
+    "36. Sequencing, H.G. Finishing the euchromatic sequence of the human genome. Nature 2004 ,431, 931945. 37. Heather, J.M.; Chain, B. The sequence of sequencers: The history of sequencing DNA. Genomics 2016 ,107, 18. [CrossRef] 38. Rothberg, J.M.; Leamon, J.H. The development and impact of 454 sequencing. Nat. Biotechnol. 2008 ,26, 11171124. [CrossRef] [PubMed] 39. Shendure, J.; Ji, H. Next-generation DNA sequencing. Nat. Biotechnol. 2008 ,26, 11351145. [CrossRef] [PubMed]",
+    "sequencing. Genome Res. 20, 11651173 (2010). 64. English,A.C. etal.  Assessing structural variation in a  personal genome-towards a human reference diploid  genome. BMC Genomics 16, 286 (2015). 65. Carneiro,M.O. etal.  Pacific Biosciences sequencing  technology for genotyping and variation discovery in  human data. BMC Genomics 13, 375 (2012). 66. Quail,M.A. etal.  A tale of three next generation  sequencing platforms: comparison of Ion T orrent,  Pacific Biosciences and Illumina MiSeq sequencers.",
+    "22. Karow, J. Qiagen launches GeneReader NGS System  atAMP; presents performance evaluation by broad.  GenomeWeb  [online], https:// www.genomeweb.com/ molecular-diagnostics/qiagen-launches-genereader- ngs-system-amp-presents-performance-evaluation   (4Nov 2015). 23. Smith,D.R. & McKernan,K. Methods of producing  and sequencing modified polynucleotides . US Patent  8058030 (2011). 24. Margulies,M. etal.  Genome sequencing in  microfabricated high-density picolitre reactors. Nature  437, 376380 (2005).",
+    "160. Glenn,T .C. Field guide to next-generation DNA  sequencers. Mol. Ecol. Resour. 11, 759769 (2011). 161. Karow,J. At AGBT , 10X Genomics launches GemCode  platform; shipments slated for Q2 as firm battles IP  lawsuits.  GenomeWeb  [online], https://www. genomeweb.com/sample-prep/agbt-10x-genomics- launches-gemcode-platform-shipments-slated-q2-firm- battles-ip-lawsuits  (2Mar 2015). Competing interests statement The authors declare competing interests: see Web version  for  details. FURTHER INFORMATION",
+    "sequencing. Bioinformatics 31, 20402042 (2015). 46. Qiagen.  Oncology insights enabled by knowledge base- guided panel design and the seamless workflow of the  GeneReader NGS system  Press Release. Qiagen   [online], http://www.genereaderngs.com/PROM-9192- 001_1100403_WP_GeneReader_NGS_0116_NA.pdf   (2016). 47. Forgetta,V. etal.  Sequencing of the Dutch elm disease  fungus genome using the Roche/454 GS-FLX Titanium  System in a comparison of multiple genomics core",
+    "DNA), and provide the means to link sequences containing applications. First,  base- callers like Phred (4,5) extract raw sequences from raw data. There are  also contig assemblers like Phrap (University of Washington, http://bozeman. mbt.washington.edu/phrap.docs/phrap.html ) or CAP3 (6) that assemble frag- ments to contigs and packages like consed (7) or GAP4 (8), which are used to  finish sequencing projects. These programs are not explained in detail here.",
+    "sequencing data to solutions from the genotyping array data. iv               PREVIEW",
+    "11 BIOINFORMATIC CHALLENGES FOR GENOMIC MEDICINE      Processing and managing of high-throughput sequence data     High throughput sequencing offers severa l advantages relative to array-based  genotyping or expression assays. First, unlike genotyping arrays, whole genome sequencing is not limited to interrogating onl y known sequence variants. Similarly, RNA- sequencing (RNA-seq) enables expression quanti fication of novel transcripts that are not",
+    "11 BIOINFORMATIC CHALLENGES FOR GENOMIC MEDICINE      Processing and managing of high-throughput sequence data     High throughput sequencing offers severa l advantages relative to array-based  genotyping or expression assays. First, unlike genotyping arrays, whole genome sequencing is not limited to interrogating onl y known sequence variants. Similarly, RNA- sequencing (RNA-seq) enables expression quanti fication of novel transcripts that are not"
+  ],
+  "answer": "### How to Guide for Genetic Sequencing  Genetic sequencing is a powerful tool used to determine the order of nucleotides in DNA. This guide will walk you through the basic steps involved in genetic sequencing, from sample preparation to data analysis.  #### Step 1: Sample Preparation 1. **Collect DNA Sample**: Obtain a DNA sample from the organism or tissue of interest. This can be done using various methods such as blood draws, buccal swabs, or tissue biopsies. 2. **Extract DNA**: Use a DNA extraction kit to isolate DNA from the collected sample. Ensure the DNA is of high quality and free from contaminants.  #### Step 2: Library Preparation 1. **Fragment DNA**: Break the DNA into smaller fragments. This can be done using mechanical shearing or enzymatic digestion. 2. **Add Adapters**: Ligate specific adapters to the ends of the DNA fragments. These adapters are necessary for the sequencing process and allow the fragments to bind to the sequencing platform.  #### Step 3: Amplification 1. **PCR Amplification**: Amplify the DNA fragments using Polymerase Chain Reaction (PCR). This step increases the quantity of DNA to ensure there is enough material for sequencing.  #### Step 4: Sequencing 1. **Choose a Sequencing Platform**: Select an appropriate sequencing platform based on your needs. Popular platforms include Illumina, Ion Torrent, Pacific Biosciences, and Oxford Nanopore Technologies [1], [5], [6]. 2. **Load the Sample**: Load the prepared DNA library onto the sequencing platform. Each platform has specific protocols for loading samples. 3. **Run the Sequencer**: Initiate the sequencing run. The platform will read the DNA fragments and generate raw sequence data.  #### Step 5: Data Analysis 1. **Base Calling**: Use base-calling software to convert raw data into nucleotide sequences. Tools like Phred are commonly used for this purpose [7]. 2. **Sequence Assembly**: Assemble the short DNA sequences into longer contiguous sequences (contigs) using assemblers like Phrap or CAP3 [7]. 3. **Quality Control**: Assess the quality of the assembled sequences. Tools like consed or GAP4 can help in finishing sequencing projects and ensuring accuracy [7].  #### Step 6: Interpretation 1. **Variant Calling**: Identify genetic variants such as single nucleotide polymorphisms (SNPs) and insertions/deletions (indels) using variant calling software. 2. **Annotation**: Annotate the identified variants to understand their potential impact on genes and biological functions. 3. **Data Management**: Store and manage the sequencing data using bioinformatics tools and databases. High-throughput sequencing generates large amounts of data that require efficient processing and management [9], [10].  #### Additional Resources - For more detailed protocols and information on specific sequencing technologies, refer to resources from companies like 10X Genomics, Illumina, and Pacific Biosciences [1]. - For historical context and advancements in sequencing technologies, see references such as Heather and Chain (2016) and Shendure and Ji (2008) [2].  By following these steps, you can successfully perform genetic sequencing and analyze the resulting data to gain insights into the genetic makeup of your sample.",
+  "question": "Create a how to guide for genetic sequencing"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_1 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_1
new file mode 100644
index 0000000..a9796ef
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_1
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2007 - Combining classical trait and microarray data to dissect transcriptional regulation a case study.pdf",
+    "2017 - Genomic regulation of type 2 diabetes endophenotypes Contribution.pdf",
+    "2010 - Neural tube defect genes and maternal diabetes during pregnancy.pdf",
+    "2009 - Prioritizing genes for follow-up from genome wide association studies using information on gene expression in tissues relevant for type 2 diabetes mellitus.pdf",
+    "2022 - System Genetics in the Rat Family.pdf",
+    "2022 - Systems genetics in the rat HXBBXH family identifies Tti2 as a pleiotropic quantitative trait gene for adult hippocampal neurogenesis and serum glucose.pdf",
+    "2022 -Senko- Hippocampal neurogenesis serum glucose.pdf",
+    "2022 -Senko- System Genetics in the Rat HXB\uf022BXH Family.pdf",
+    "2020 - ChREBP downregulates SNAT2 amino acid transporter expression through interactions with SMRT in response to a high-carbohydrate diet.pdf",
+    "2015 - Targeted Allelic Expression.pdf"
+  ],
+  "extraction_id": [
+    "1e5ec803-ae2d-5bbd-8d40-438fb1ec1eab",
+    "a0845748-d229-56b1-8666-5fd7708267b4",
+    "eaa27c67-ef56-5b12-8dc0-a656cc36c529",
+    "543f1861-21f2-52de-88e5-fa81a7b6ef64",
+    "ec24c99e-4654-5fb7-a1ed-ec3f8a941711",
+    "184f8279-2ea5-5f18-8e15-2804ee9e62d5",
+    "c597d023-1a22-5849-8c4f-9f3448c22962",
+    "a56d014f-d78d-582c-845d-2b10823f5424",
+    "a575ca7c-aa73-5b6a-a152-0ff08ddec434",
+    "37df3b54-130c-5424-90f6-af59ecb5cdf8"
+  ],
+  "document_id": [
+    "bb54e43d-7f70-5ee2-a5b9-0e20000dfd97",
+    "fef1ae33-b3af-50ea-909c-f1b57f7fe981",
+    "aa74b552-7e06-5596-8dec-298c40ad558c",
+    "4b1a56e7-6821-5504-b6da-27dcdf57c6a5",
+    "426b5aeb-1550-5039-8f2a-bd83d17c8648",
+    "e6323aba-6fec-500b-99e3-a41c2e7f17ff",
+    "bac2ab98-4317-59ed-99ef-deda8c22786d",
+    "c67a6829-954a-5202-85fb-7524b03fab28",
+    "fbfc6093-648c-55f7-9fc8-2ec4964278f1",
+    "a0f46d1e-81be-5b29-9082-86c1114c3edd"
+  ],
+  "id": [
+    "chatcmpl-ADYmUfKwQ32pLN2HQWzuhXDWOhemk",
+    "08c0f648-0618-56cb-935a-c627000943f4",
+    "1b2895af-da13-52dd-9fd2-133a43b98b5f",
+    "39d6e4a1-5bbd-5f35-80b2-d3c205a5457c",
+    "2a71b5a3-67d8-55d8-97f8-cb34cbfcaa41",
+    "1e08685d-0f9d-5ead-84c1-e97fe346e025",
+    "4c381a87-dc30-5d3a-95a9-a32255cfe571",
+    "e8e69e50-076e-5459-ac5a-8e267fa33e13",
+    "0be84448-80cf-52bd-a84c-668a9ac49b20",
+    "6b49a027-22fc-59c5-aa87-3155663fd003",
+    "0feb3ea0-bd53-5e94-8a65-8cd2bdecdf0e"
+  ],
+  "contexts": [
+    "Lan H, Rabaglia ME, Stoehr JP, Nadler ST, Schueler KL et al (2003) Gene expression proles of nondiabetic and diabetic obese mice suggest a role of hepatic lipogenic capacity in diabetes susceptibility. Diabetes 52:688700Theor Appl Genet (2008) 116:683690 689 123",
+    "Effects of high fat feeding on liver gene expression in diabetic goto-kakizaki rats, Gene Regul. Syst. Bio 6 (2012) 151 e168. [23] P.J. Kaisaki, G.W. Otto, J.F. McGouran, A. Toubal, K. Argoud, H. Waller-Evans, C. Finlay, S. Cald /C19erari, M.T. Bihoreau, B.M. Kessler, D. Gauguier, R. Mott, Ge- netic control of differential acetylation in diabetic rats, PLoS One 9 (2014) e94555 . [24] S.P. Wilder, P.J. Kaisaki, K. Argoud, A. Salhan, J. Ragoussis, M.T. Bihoreau,",
+    "Figure 2. Diabetes increases the variability of gene expression levels in other experimental paradigms. ( A) Microarray data from gene",
+    "also showed differential expression in the liver, where it regulates a number of genes involved in both glucose andlipid metabolism. These results add further support to aTable 3: Numbers of genes for which expressi on levels in pancreas, skel etal muscle, adipose tissue or  liver were altered in dia betes as  compared to controls P < 0.01 (DGI) P < 0.05 (DGI) P < 0.01 (WTCCC) 11 42 P < 0.05 (WTCCC) 30 115 P < 0.01 in DGI and P < 0.05 in WTCCC or  P < 0.01 in WTCCC and P < 0.05 in DGI60",
+    "toSHR wild type littermates. Liver, together with skeletal muscle and adipose tissue, aredeci- sive organs inmaintaining glucose homeostasis and, hence, thedevelopment ofinsulin resis- tance [75]. Functional analysis ofdifferentially expressed genes intheliver identified networks ofgenes and potential regulators whose activation and inhibition could explain insulin resis- tance and dysglycemia intheheterozygous animals. Wealso recorded significant upregulation",
+    "toSHR wild type littermates. Liver, together with skeletal muscle and adipose tissue, aredeci- sive organs inmaintaining glucose homeostasis and, hence, thedevelopment ofinsulin resis- tance [75]. Functional analysis ofdifferentially expressed genes intheliver identified networks ofgenes and potential regulators whose activation and inhibition could explain insulin resis- tance and dysglycemia intheheterozygous animals. Wealso recorded significant upregulation",
+    "toSHR wild type littermates. Liver, together with skeletal muscle and adipose tissue, aredeci- sive organs inmaintaining glucose homeostasis and, hence, thedevelopment ofinsulin resis- tance [75]. Functional analysis ofdifferentially expressed genes intheliver identified networks ofgenes and potential regulators whose activation and inhibition could explain insulin resis- tance and dysglycemia intheheterozygous animals. Wealso recorded significant upregulation",
+    "toSHR wild type littermates. Liver, together with skeletal muscle and adipose tissue, aredeci- sive organs inmaintaining glucose homeostasis and, hence, thedevelopment ofinsulin resis- tance [75]. Functional analysis ofdifferentially expressed genes intheliver identified networks ofgenes and potential regulators whose activation and inhibition could explain insulin resis- tance and dysglycemia intheheterozygous animals. Wealso recorded significant upregulation",
+    "mRNA in diabetic liver. Biochem Biophys Res Commun 290: 903-908, 2002. 712 42. Watson PJ, Fairall L, and Schwabe JW . Nuclear hormone receptor co-repressors: 713 structure and function. Mol Cell Endocrinol 348: 440-449, 2012. 714 43. Yamashita H, Takenoshita M, Sakurai M, Bruick RK, Henzel WJ, Sh illinglaw 715 W, Arnot D, and Uyeda K . A glucose-responsive transcr iption factor that regulates 716 carbohydrate metabolism in the liver. Proc Natl Acad Sci U S A 98: 9116-9121, 2001. 717",
+    "impacts gene expression in a cell type-dependent manner. Science 2009;325:1246 1250diabetes.diabetesjournals.org Locke and Associates 1491Downloaded from http://diabetesjournals.org/diabetes/article-pdf/64/4/1484/580211/db140957.pdf by Kenya Institution user on 11 July 2023"
+  ],
+  "answer": "Gene expression in the liver is significantly affected by diabetes in several ways:  1. **Hepatic Lipogenic Capacity**: In diabetic obese mice, gene expression profiles suggest a role of hepatic lipogenic capacity in diabetes susceptibility, indicating that genes involved in lipid metabolism are differentially expressed in the liver of diabetic individuals [1].  2. **High Fat Feeding**: In diabetic Goto-Kakizaki rats, high fat feeding alters liver gene expression, which implies that dietary factors can further influence the gene expression changes associated with diabetes [2].  3. **Variability of Gene Expression**: Diabetes increases the variability of gene expression levels in the liver, affecting the regulation of genes involved in both glucose and lipid metabolism [3], [4].  4. **Glucose Homeostasis and Insulin Resistance**: The liver, along with skeletal muscle and adipose tissue, plays a crucial role in maintaining glucose homeostasis. Functional analysis of differentially expressed genes in the liver has identified networks of genes and potential regulators whose activation and inhibition could explain insulin resistance and dysglycemia in diabetic conditions [5], [6], [7], [8].  5. **Carbohydrate Metabolism**: A glucose-responsive transcription factor that regulates carbohydrate metabolism in the liver is affected by diabetes, indicating changes in the expression of genes involved in glucose metabolism [9].  Overall, diabetes leads to significant changes in the expression of genes in the liver, affecting pathways related to lipid and glucose metabolism, insulin resistance, and overall metabolic homeostasis.",
+  "question": "How is gene expression in the liver affected by diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_10 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_10
new file mode 100644
index 0000000..639579e
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_10
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2004 - A New Look at Viruses in Type 1 Diabetes.pdf",
+    "2020 - Integration of genomics and transcriptomics predicts diabetic retinopathy susceptibility genes.pdf",
+    "2016 - Integrated multi-omics of the human gut microbiome in a case study of familial type 1 diabetes.pdf",
+    "2003 - A functional polymorphism in the promoterenhancer region of the FOXP3Scurfin gene associated with type 1 diabetes.pdf",
+    "2005 - Pathway analysis of coronary atherosclerosis.pdf",
+    "2003 -Genetic epidemiology of type 1 diabetes.pdf",
+    "2018 - The human gut microbiome in early-onset type 1 diabetes from the TEDDY study.pdf",
+    "2011 - Type 1 Diabetes Etiology, Immunology.pdf",
+    "2017 - Type 1 diabetes mellitus.pdf",
+    "2004 - Diabetes Genes a.pdf"
+  ],
+  "extraction_id": [
+    "8bbb4581-dc07-5410-9737-6d249f3740f6",
+    "018ac588-c327-5122-9c18-18f4d0df0f14",
+    "092a9b75-9985-5876-a650-59bc3f0d10fb",
+    "aacbb5a1-c294-5568-ba02-3d4342091e86",
+    "858559b5-74d3-585a-9f45-ffa065ecb0f7",
+    "84a487be-a531-5f09-b2d5-d0525c59d581",
+    "9cca2fe6-7584-5d28-91f3-e06edca7ed54",
+    "388e7eec-4204-59b5-a42d-e56a9032da0b",
+    "d342e632-c951-519a-b0de-505f3515403d",
+    "48f690af-58fa-59e1-a0ca-ce421aaa356c"
+  ],
+  "document_id": [
+    "38edad91-ff31-504e-91d8-eac3833615b0",
+    "699a10ff-44d7-5cb3-bc25-ec5ba85cb751",
+    "f0405966-38bf-5a04-aa2c-1474b11362bb",
+    "4a3964a4-0aea-58ee-b749-33e0d8c62228",
+    "fa9c400b-fbfa-54ce-a801-7594b489e42d",
+    "cbc7f2d3-3f65-50ba-b281-96dd1c77f2c0",
+    "36096262-86f1-5c7e-bea1-4abbc610a974",
+    "3c9823cd-3615-53b6-96c8-b7d2123d3eb0",
+    "8e8b9b6e-8dfb-5aae-8c61-5f53bd4e0242",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa"
+  ],
+  "id": [
+    "chatcmpl-ADYnJbAeICzXtvQR31T420R0p7Xn0",
+    "d156068e-31f6-5464-8ef1-eb5e7c58aa8d",
+    "b205daa9-4723-5641-9ed4-428d83cf7758",
+    "e7e8ef7b-bad0-54bc-814d-d947ea04756b",
+    "c851d17c-1ad0-5b9a-8820-ad45d0e4b075",
+    "0f6e6870-960c-560d-ad61-36c1d4d9970f",
+    "f6fd1d0f-d88e-55f7-8ed6-bba917a65b8f",
+    "00b43e01-2296-528e-82e1-5671bffe784d",
+    "0790a91d-f1c5-519a-9b0e-73a9f73b8da4",
+    "5daae5a1-9163-5850-874b-ea63ecdd4f87",
+    "65247182-02f3-501c-94d4-36f4893ff703"
+  ],
+  "contexts": [
+    "disordering particular lymphocyte subsets [57]. Viral anti-body-free BB rats show an increased frequency and accel-erated onset of diabetes, suggesting that infection may havea protective effect against the development of diabetes bythese animals [230]. Thus, we speculate that infection orimmune stimulation in humans may also reduce the pen-etrance of susceptibility genes, which could account for thelow concordance rate between identical twins of less than40% for the development of T1D [13]. Conclusion",
+    "ished immune responsiveness, a well-characterized feature of diabetes ( Shanmugam et al., 2003 ; Mowat and Baum, 1971 ). Further, we considered that the genetic component of an individuals response to glucose may influence their susceptibility to diabetic complications like retinopathy. Cell lines from individuals with diabetes with and without retinopathy reveal differences in the response to glucose at a molec-",
+    "diabetes. ISME J. 5,8291 (2011). 30. Brown, C. T. et al. Gut microbiome metagenomics analysis suggests a functional model for the development of autoimmunity for type 1 diabetes.PLoS ONE 6,e25792 (2011). 31. Endesfelder, D. et al. Compromised gut microbiota networks in children with anti-islet cell autoimmunity. Diabetes 63,2006 2014 (2014). 32. Kostic, A. D. et al. The dynamics of the human infant gut microbiome in development and in progression toward type 1 diabetes. Cell Host Microbe 17, 260273 (2015).",
+    "+T cells related to diabetes-associated",
+    "the innate immune system (8, 36, 37) are known to play important roles in the development of diabetes itself, no study to date has linked these ideas with the",
+    "same or related viruses might complete the process of immune-mediated b-cell destruction. Alternatively, chil- dren genetically predisposed to develop autoimmunediabetes might have an altered immune system that is more likely to respond to viral exposures with strongly detectable antibody levels against certain viral antigens.If so, the detectable levels of antibodies to multiple viral antigens in diabetic patients would not indicate a causal",
+    "with -cell autoimmunity and those without. Diabetes 62, 12381244 (2013).  9. Mario, E. et al. Gut microbial metabolites limit the frequency of autoimmune  T cells and protect against type 1 diabetes. Nat. Immunol. 18, 552562  (2017).  10. Needell, J. C. & Zipris, D. The role of the intestinal microbiome in type 1 diabetes pathogenesis. Curr. Diab. Rep. 16, 89 (2016).  11. Davis-Richardson, A. G. et al. Bacteroides dorei dominates gut microbiome prior",
+    "141. Filippi CM, Estes EA, Oldham JE, von Herrath MG. Immuno- regulatory mechanisms triggered by viral infections protect fromtype 1 diabetes in mice. J Clin Invest 119: 15151523, 2009. 142. Filippi CM, von Herrath MG. Viral trigger for type 1 diabetes: pros and cons. Diabetes 57: 28632871, 2008. 143. Flohe SB, Wasmuth HE, Kerad JB, Beales PE, Pozzilli P. A wheat-based, diabetes-promoting diet induces a Th1-type cytokinebias in the gut of NOD mice. Cytokine 21: 149154, 2003.",
+    "12451252 (2008). 77. Hofer,J.  etal.  Elevated proportions of recent thymic  emigrants in children and adolescents with type1  diabetes. Rejuvenation Res. 12, 311320 (2009). 78. Wong,F.S. How does Bcell tolerance contribute to  the protective effects of diabetes following induced  mixed chimerism in autoimmune diabetes? Diabetes  63, 18551857 (2014). 79. Roep,B.O. & Peakman,M. Antigen targets of type1  diabetes autoimmunity. Cold Spring Harb. Perspect.  Med.  2, a007781 (2012).",
+    "Immune Hypothesis: Evidence supporting an immune pathogenesis  is strongest for diabetic autonomic neuropathy. Autonomic ganglia heavily  infiltrated by lymphocytes, plasma cells, and macrophages were found at  autopsy in five type 1 diabetics with symptomatic autonomic neuropathy.  Striking cervical sympathetic ganglia atrophy was reported in another with  severe sensory and autonomic neuropathy.32  Autoimmune pathogenesis may be involved in proximal diabetic"
+  ],
+  "answer": "The immune system is intricately related to diabetes, particularly Type 1 Diabetes (T1D), through several mechanisms:  1. **Autoimmunity and \u03b2-cell Destruction**: T1D is characterized by the immune system mistakenly attacking and destroying insulin-producing \u03b2-cells in the pancreas. This autoimmune response is a central feature of T1D [6].  2. **Immune Responsiveness**: Diabetes is associated with diminished immune responsiveness, which can affect the body's ability to manage glucose levels and may influence susceptibility to diabetic complications [2].  3. **Role of Infections**: Infections or immune stimulation may have a protective effect against the development of diabetes. For instance, viral infections have been shown to trigger immunoregulatory mechanisms that protect against T1D in animal models [1], [8].  4. **Gut Microbiome and Autoimmunity**: The gut microbiome plays a significant role in the development of autoimmunity related to T1D. Alterations in gut microbiota networks have been observed in children with anti-islet cell autoimmunity, suggesting a link between gut health and immune responses in diabetes [3], [7].  5. **Immune Pathogenesis in Complications**: Evidence supports an immune pathogenesis for diabetic complications such as autonomic neuropathy, where immune cells infiltrate autonomic ganglia, leading to nerve damage [10].  In summary, the immune system's malfunction, particularly through autoimmunity and altered immune responses, is a key factor in the development and progression of diabetes, especially T1D.",
+  "question": "How is the immune system related to diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_11 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_11
new file mode 100644
index 0000000..af8ae20
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_11
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2015 - Type 2 Diabetes Mellitus and the Association of Candidate Genes.pdf",
+    "2009 - Analysis of 19 genes for association with type I diabetes in the Type I Diabetes Genetics Consortium families..pdf",
+    "2009 - Genome-wide association study and meta-analysis find that over 40 loci affect risk of type 1 diabetes.pdf",
+    "2015 - Transcript Expression Data from Human.pdf",
+    "2016 - Systematic Evaluation of Genes and Genetic Variants Associated with Type 1 Diabetes Susceptibility.pdf",
+    "2010 - Twelve type 2 diabetes susceptibility loci identified.pdf",
+    "2022 - A genome-wide functional genomics approach uncovers genetic determinants of immune phenotypes in type 1 diabetes.pdf",
+    "2021- Genome\u2010wide search for genes affecting the age at diagnosis of type 1.pdf",
+    "2008 - Shared and Distinct Genetic Variants in Type 1 Diabetes.pdf",
+    "2023 - Childhood adiposity and novel subtypes of adult-onset diabetes a Mendelian randomisation and genome-wide genetic correlation study.pdf"
+  ],
+  "extraction_id": [
+    "e1e1abb2-882f-5ba4-a51b-3b9bfc4df5aa",
+    "8ae199fd-0820-54c6-8d5c-aea5bf5fb895",
+    "a58e318d-3358-518c-ac23-6dd4d7b000f2",
+    "8fb04ac0-460b-58d3-ad43-2c7720bfd87e",
+    "082f1c10-0745-5d70-a176-336fc972319c",
+    "6912cf22-46e3-540b-bafe-f4951ec2bd70",
+    "fc30d552-be59-5ddf-9bac-e247d536ed96",
+    "bc2a4183-8ca7-5b72-8e03-25f4933ecc8b",
+    "20a53007-acf3-5317-89d5-1d69f1845d62",
+    "defbf2fb-7aa7-538d-b6ac-81ecd607179c"
+  ],
+  "document_id": [
+    "3236fdee-e304-5b88-921f-52e52dc5afa3",
+    "a0e27a2d-a07b-5b4d-a93a-907303dd8876",
+    "e6566ede-0c5a-51d1-aac9-e6e1695e937a",
+    "2b30d4f3-9ec3-574f-9a36-709b0e09c3f2",
+    "e4288a56-0280-5681-8eb4-4f52b3160451",
+    "8be48d47-68bd-5bec-844d-7ddd3e624442",
+    "368e0215-393e-5bec-a87c-e976adaa3ca5",
+    "a98a972b-8b50-58c8-9126-1883a96b1a09",
+    "4a655174-c16b-54d5-901a-6508d638cc23",
+    "fff2bd78-2ac2-5672-b8fd-ed82ab7c910b"
+  ],
+  "id": [
+    "chatcmpl-ADYnQqyvhC46GEUXw2f9p5UR309ef",
+    "d8d64729-8353-5fd3-938f-c7e0467698f9",
+    "07f9090f-101c-5b89-ab7b-a072dbf1ed4b",
+    "96b66f03-33dd-5a88-91c8-e0aa13cbcf3d",
+    "9a17c246-a9c7-5c13-92ef-5d551c7439e9",
+    "f3e96e99-cd95-5c0c-92c5-72d6edf2f6ff",
+    "9a6042ed-f076-51c3-b0f3-3d8b94e9852f",
+    "123d1a9a-12c9-59a2-8f3e-083220452036",
+    "6fc3a7f1-bd7e-55d9-be9b-1c6f5fb5452e",
+    "ca60f298-62fe-5fcc-a833-8439733cfae2",
+    "81df736a-3450-53da-9421-57f7d29e3218"
+  ],
+  "contexts": [
+    "Imran Ali Khan et al., Genetic Variants in Indian Diabetes Patients  www.jcdr.net Journal of Clinical and Diagnostic Research. 2015 Nov, Vol-9(11): GC01-GC05 44of the pancreas and islets during embryonic growth [3]. Genetic  variants in this gene are associated with increased risk of T2DM in a  variety of study populations [28,29].  In the first published GWAS for T2DM, SLC30A8 (rs13266634) was  revealed to be associated with diabetes (OR, 1.26; p = 5.0  10-7).",
+    "diabetes and celiac disease. N Engl J Med 2008; 359: 27672777. 11 Fung E, Smyth DJ, Howson JM, Cooper JD, Walker NM, Stevens H et al. Analysis of 17 autoimmune disease-associated variants in type 1 diabetes identifies 6q23/TNFAIP3 as asusceptibility locus. Genes Immun 2008; 10: 188191. 12 Cooper JD, Smyth DJ, Smiles AM, Plagnol V, Walker NM, Allen JE et al. Meta-analysis of genome-wide association study data identifies additional type 1 diabetes risk loci. Nat Genet 2008; 40: 13991401.",
+    "10. Smyth, D.J. et al. Shared and distinct genetic variants in type 1 diabetes and celiac disease. N. Engl. J. Med. 359, 27672777 (2008). 11. Fung, E. et al. Analysis of 17 autoimmune disease-associated variants in type 1 diabetes identies 6q23/TNFAIP3 as a susceptibility locus. Genes Immun. 10, 188191 (2009). 12. Cooper, J.D. et al. Meta-analysis of genome-wide association study data identies additional type 1 diabetes risk loci. Nat. Genet. 40, 13991401 (2008).",
+    "14. Pasquali L, Gaulton KJ, Rodriguez-Segui SA, Mularoni L, Miguel-Escalada I, et al. (2014) Pancreatic islet enhancer clusters enriched in type 2 diabetes risk-associated variants. Nat Genet 46: 136 143. doi:10.1038/ng.2870 PMID: 24413736 15. Fairfax BP, Humburg P, Makino S, Naranbhai V, Wong D, et al. (2014) Innate immune activity condi- tions the effect of regulatory variants upon monocyte gene expression. Science 343: 1246949. doi: 10. 1126/science.1246949 PMID: 24604202",
+    "The Journal of Immunology Systematic Evaluation of Genes and Genetic Variants Associated with Type 1 Diabetes Susceptibility Ramesh Ram,*,Munish Mehta,*,Quang T. Nguyen,*,Irma Larma,*, Bernhard O. Boehm,,xFlemming Pociot,{Patrick Concannon,,#and Grant Morahan*, Genome-wide association studies have found >60 loci that confer genetic susceptibility to type 1 diabetes (T1D). Many of these are",
+    "disease and type II diabetes. Genes Immun.  10, 654658 (2009). 41. Hindorff, L.A. et al. Potential etiologic and functional implications of genome-wide  association loci for human diseases and traits. Proc. Natl. Acad. Sci. USA  106,  93629367 (2009). 42. Nicolson, T.J. et al.  Insulin storage and glucose homeostasis in mice null for the  granule zinc transporter ZnT8 and studies of the type 2 diabetes-associated variants.  Diabetes  58, 20702083 (2009).",
+    "The composition and activity of the human immune system is under genetic control, and people  with certain changes in their genes are more susceptible than others to develop type 1 diabetes.  Previous studies have identified around 60 locations in the human DNA (known as loci) associated  with the condition, but it remains unclear how these loci influence the immune system and whether  diabetes will emerge. Chu, Janssen, Koenen et al. explored how variations in genetic information can influence the",
+    "mellitus-associated genetic variants contribute to overlapping immune regulatory networks. Front Genet 2018; 9:535. 13 Syreeni A, Sandholm N, Cao J et al. Genetic determinants of glycated hemoglobin in type 1 diabetes. Diabetes 2019; 68: 858 67. 14 Sidore C, Busonero F, Maschio A et al. Genome sequencing elucidates Sardinian genetic architecture and augmentsGenes affecting type 1 diabetes diagnosis age / A. Syreeni et al .",
+    "Genetic Variants in Type 1 Diabetes and Celiac Disease n engl j med 359;26 www.nejm.org december 25, 2008 2777Kalev I, Oselin K, Prlist P, et al. CC-26.  chemokine receptor CCR5-del32 mutation  as a modifying pathogenetic factor in type I diabetes. J Diabetes Complications 2003;17:387-91. Szalai C, Csszr A, Czinner A, et al.  27.  Chemokine receptor CCR2 and CCR5 polymorphisms in children with insulin-dependent diabetes mellitus. Pediatr Res 1999;46:82-4. Yang B, Houlberg K, Millward A, De - 28.",
+    "13(1):2337. https://doi.org/10.1038/s41467-022-29932-y 5. Burgess S, Butterworth A, Thompson SG (2013) Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol 37(7):658 665. https://doi. org/10.1002/gepi.21758 6. Cousminer DL, Ahlqvist E, Mishra R et al (2018) First genome- wide association study of latent autoimmune diabetes in adults reveals novel insights linking immune and metabolic diabetes. Diabetes Care 41(11):2396 2403. https://doi.org/10.2337/dc18-"
+  ],
+  "answer": "The genomic variants associated with immune system components and diabetes include:  1. **SLC30A8 (rs13266634)**: This variant is associated with an increased risk of Type 2 Diabetes Mellitus (T2DM) [1]. 2. **6q23/TNFAIP3**: Identified as a susceptibility locus for Type 1 Diabetes (T1D) [2], [3]. 3. **Pancreatic islet enhancer clusters**: These clusters are enriched in T2DM risk-associated variants [4]. 4. **Regulatory variants affecting monocyte gene expression**: These variants are conditioned by innate immune activity [4]. 5. **>60 loci**: Genome-wide association studies have identified over 60 loci that confer genetic susceptibility to T1D [5], [7]. 6. **CCR5-del32 mutation**: This mutation in the chemokine receptor CCR5 is a modifying pathogenetic factor in T1D [9]. 7. **CCR2 and CCR5 polymorphisms**: These polymorphisms are found in children with insulin-dependent diabetes mellitus [9]. 8. **Novel insights linking immune and metabolic diabetes**: Identified through the first genome-wide association study of latent autoimmune diabetes in adults [10].  These variants highlight the complex interplay between genetic factors influencing both the immune system and diabetes.",
+  "question": "What are the genomic variants associated with immune system components and diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_12 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_12
new file mode 100644
index 0000000..e684657
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_12
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2010 - Metabolic Footprint of Diabetes A Multiplatform.pdf",
+    "2012 - Novel biomarkers for pre\u2010diabetes identified by metabolomics.pdf",
+    "2010 - Metabolic Footprint of Diabetes A Multiplatform.pdf",
+    "2010 - Metabolic Footprint of Diabetes A Multiplatform.pdf",
+    "2011 - Emerging Applications of Metabolomic.pdf",
+    "2015 - Genetics, genomics and personalized medicine in Type 2 Diabetes.pdf",
+    "2011 - The eMERGE Network A consortium of biorepositories linked to electronic medical records data for conducting genomic studies.pdf",
+    "2011 - Biomarkers for the Prediction of Type 2 Diabetes.pdf",
+    "2010 - Metabolic Footprint of Diabetes A Multiplatform.pdf",
+    "2009 - Metabolomics Applied to Diabetes Research.pdf"
+  ],
+  "extraction_id": [
+    "8ad05bed-b0fd-52d4-badf-a98233a0808d",
+    "c775612c-c80b-5a50-9417-d6fd89ec07ee",
+    "2359c12d-8263-5183-a350-fff365318805",
+    "97b6d492-9139-50ec-9685-53a803f5c995",
+    "df823d9a-e2de-5dab-b336-af4682b9ce70",
+    "92a2a3c7-ed41-5394-b716-fdbf5c198a86",
+    "a35d4e2a-ce04-536d-b88a-8f273aa03f40",
+    "75f979f5-425b-563c-b4ba-ec3a971f356a",
+    "6d77a75e-68a4-5c27-b387-449f7f9f9487",
+    "380e9a2e-8f9f-5f9e-ba20-3695b1c60fda"
+  ],
+  "document_id": [
+    "b199607e-293e-56e8-88c8-e0716d1ee9eb",
+    "d93e3562-3419-51a6-86db-8247a9e69361",
+    "b199607e-293e-56e8-88c8-e0716d1ee9eb",
+    "b199607e-293e-56e8-88c8-e0716d1ee9eb",
+    "10c69e6a-3771-5cc6-a915-a31556dec650",
+    "d8b85c3e-62f3-5e67-99b0-d0a2f225aff0",
+    "3a174301-2941-578f-8ed6-f16d88fd2230",
+    "c68d29dd-eaa1-53f8-bc0d-aa85b2f39352",
+    "b199607e-293e-56e8-88c8-e0716d1ee9eb",
+    "a6ae2fb6-88ae-588f-a98d-b6092f886ed9"
+  ],
+  "id": [
+    "chatcmpl-ADYnYRyH5dd9Q9xzg35pmgGcm27tO",
+    "13c68218-4920-5df7-a0b4-017298c9001a",
+    "393e2363-48e6-56ad-94d6-39b1915b2f5a",
+    "3df2fa36-b9aa-51c6-9e36-acfcef1310b6",
+    "ead10261-182f-5ab1-9af0-ce8a17677d4a",
+    "024eea85-c974-51fc-8def-89db09ba56b0",
+    "cef34be2-673e-553f-9c92-1ecef8edec4f",
+    "5c7dc6d7-800e-5c77-ac61-bd8e3086754c",
+    "3b9547ce-8316-5256-a68b-256058b3ee79",
+    "06da63dc-6a8d-5682-80e0-7d37b66cdf6f",
+    "0cb19f85-21d9-54f1-81a4-43969ac050e8"
+  ],
+  "contexts": [
+    "allows the detection of systemic metabolic imbalances, thereby providing a disease specific picture of human physiology. doi:10.1371/journal.pone.0013953.g003Metabolomics of Diabetes PLoS ONE | www.plosone.org 9 November 2010 | Volume 5 | Issue 11 | e13953",
+    "Metabolomics studies allow metabolites involved in disease mechanisms to be discovered by monitoring metabolite level changes in predisposed individuals compared with healthy ones (Shaham et al, 2008; Newgard et al, 2009; Zhao et al, 2010; Pietilainen et al, 2011; Rhee et al, 2011; Wang et al,2 0 1 1 ; Cheng et al, 2012; Goek et al, 2012). Altered metabolite levels may serve as diagnostic biomarkers and enable preventive action. Previous cross-sectional metabolomics studies of T2D",
+    "doi:10.1371/journal.pone.0013953.t006Metabolomics of Diabetes PLoS ONE | www.plosone.org 8 November 2010 | Volume 5 | Issue 11 | e13953",
+    "monitoring and preventing progression to costly co-morbidities. The principal concept of metabolomics being able to find some metabolites differing in a control and a type 2 diabetic group is established. It is not our goal here to show this once again. The questions we ask are rather How well are different approaches suited to attain this goal? and What are optimal settings under which such studies can be successful?. Others have already investigated these questions before [16,17,18]. However, we",
+    "H, Raftery D, Nair KS. Quantitative me-tabolomics by H-NMR and LC-MS/MSconrms altered metabolic pathways in diabetes. PLoS ONE 2010;5:e10538 2. Li LO, Hu YF, Wang L, Mitchell M, Berger A, Coleman RA. Early hepatic insulin re-sistance in mice: a metabolomics analysis.Mol Endocrinol 2010;24:657 666 3. Bain JR, Stevens RD, Wenner BR, Ilkayeva O, Muoio DM, Newgard CB. Metabolomicsapplied to diabetes research: moving frominformation to knowledge. Diabetes 2009; 58:2429 2443",
+    "70 Zhang Q, Fillmore TL, Schepmoes AA et al. Serum proteomics reveals systemic dysregulation of innate immunity in Type 1 diabetes. J. Exp. Med. 210(1), 191203 (2013). 71 Roberts LD, Koulman A, Griffin JL. Towards metabolic biomarkers of insulin resistance and Type 2 diabetes: progress from the metabolome. Lancet Diabetes Endocrinol.   2(1), 6575 (2014).  \t Illustrates\tpotential\tmetabolic\tbio-markers\twhich\tmay\tbe\t used\tto\tdetect\tpeople\tat-risk\tfor\tT2D/insulin\tresistance,",
+    "Serum or plasma concentrations of sugars and sugar metabo- lites (e.g., glucose, mannose, desoxyhexose, and 1,5-anhy-droglucoitol), ketone bodies ( -hydroxybutyrate),  lipids  (e.g., phosphatidyl-cholines and nonesterified fatty acids),  branched-chain amino acids, and other metabolites were found  to be associated with insulin resistance or diabetes status (see  Supplementary Data  online for full references). A proof-of- concept multi-platform, metabolome-wide study based on the",
+    "Serum or plasma concentrations of sugars and sugar metabo- lites (e.g., glucose, mannose, desoxyhexose, and 1,5-anhy-droglucoitol), ketone bodies ( -hydroxybutyrate),  lipids  (e.g., phosphatidyl-cholines and nonesterified fatty acids),  branched-chain amino acids, and other metabolites were found  to be associated with insulin resistance or diabetes status (see  Supplementary Data  online for full references). A proof-of- concept multi-platform, metabolome-wide study based on the",
+    "Conclusions/Significance: Our study depicts the promising potential of metabolomics in diabetes research by identification of a series of known and also novel, deregulated metabolites that associate with diabetes. Key observations include perturbations of metabolic pathways linked to kidney dysfunction (3-indoxyl sulfate), lipid metabolism (glyceropho-",
+    "with signicant limitations and potential for misuse oftechnologies and overinterpretation of data. Here we seekto provide a critical evaluation of progress to date inapplication of metabolomics technologies for the under-standing of diabetes and obesity mechanisms, for sub-classication of different forms of diabetes to assist intailoring of therapeutic strategies, and for more detailedevaluation of the safety and efcacy of drugs used totreat the disease.Overview of current metabolomics"
+  ],
+  "answer": "The role of the immune system in the metabolomics of diabetes and associated conditions is highlighted in the context provided. Specifically, it is noted that serum proteomics has revealed systemic dysregulation of innate immunity in Type 1 diabetes [6]. This indicates that the immune system plays a significant role in the metabolic alterations observed in diabetes. The identification of metabolic biomarkers related to immune system dysregulation can help in detecting individuals at risk for Type 2 diabetes and insulin resistance [6]. This connection underscores the importance of understanding immune system interactions within the broader metabolic landscape of diabetes.",
+  "question": "What is the role of the immune system in the metabolomics of diabetes and associated conditions?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_13 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_13
new file mode 100644
index 0000000..4b3a146
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_13
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2022 - A genome-wide functional genomics approach uncovers genetic determinants of immune phenotypes in type 1 diabetes.pdf",
+    "2020 - Whole blood co-expression modules associate with metabolic traits and type 2 diabetes an IMI-DIRECT study.pdf",
+    "2020 - Polygenic inheritance, GWAS, polygenic risk scores,and the search for functional variants.pdf",
+    "2022 - A genome-wide functional genomics approach uncovers genetic determinants of immune phenotypes in type 1 diabetes.pdf",
+    "2022 - A genome-wide functional genomics approach uncovers genetic determinants of immune phenotypes in type 1 diabetes.pdf",
+    "2010 - Comparative genetic analysis of inflammatory.pdf",
+    "2016 - Effects of the genome on immune regulation in type 1 diabetes.pdf",
+    "2018 - The genetic architecture of type 1 diabetes mellitus.pdf",
+    "2022 - A genome-wide functional genomics approach uncovers genetic determinants of immune phenotypes in type 1 diabetes.pdf",
+    "2018 - The genetic architecture of type 1 diabetes mellitus.pdf"
+  ],
+  "extraction_id": [
+    "f716d630-c82a-5987-ab4b-1b8a20d81dfa",
+    "fbf3d28b-b05c-51a2-b902-94f17ff51d7b",
+    "3615b8f4-612d-52e5-8581-8c0d97b2a845",
+    "f4c97581-4139-5397-9f3f-ccbb39846d93",
+    "bf2387f1-5389-54e8-897e-84575efee7f1",
+    "1c95778f-bf36-5398-b891-85533d60c80c",
+    "a744412e-5003-5732-9a73-f1f5267aa715",
+    "5b8b3673-7fd4-5989-9982-a6d5ea374c8d",
+    "32bf7dd7-d271-577f-9146-71da2681ec98",
+    "5b8b3673-7fd4-5989-9982-a6d5ea374c8d"
+  ],
+  "document_id": [
+    "368e0215-393e-5bec-a87c-e976adaa3ca5",
+    "a3f00a6f-be97-51ce-9198-87f6469ce2db",
+    "39ab8f23-a31d-561c-ba90-65b99f64b83e",
+    "368e0215-393e-5bec-a87c-e976adaa3ca5",
+    "368e0215-393e-5bec-a87c-e976adaa3ca5",
+    "ab74ea2b-684a-5f6f-b77b-f3dbd4de86e8",
+    "9fcfc0c3-80b5-515c-9263-a1a17cfa9a4c",
+    "341261db-b38a-5bd2-8d8d-fc04a0b3da30",
+    "368e0215-393e-5bec-a87c-e976adaa3ca5",
+    "341261db-b38a-5bd2-8d8d-fc04a0b3da30"
+  ],
+  "id": [
+    "chatcmpl-ADYnd7yDDFmHs6an7OU6ShUeBDi9c",
+    "f799cd21-0ae3-5c3a-b3d4-9dfa4f5cfcef",
+    "8090d389-97c8-521f-8139-1947014f3d5e",
+    "87c0635a-b18d-58dd-8e92-ef98d713b870",
+    "0cd28c43-f17e-5e9e-8fa9-b81ef89264c3",
+    "50474cf9-286a-50eb-b344-2837cc7c43a6",
+    "5f2de1ce-56f7-501d-a0e0-54991c1324f7",
+    "82f7da2f-7abf-59e1-b259-46a01b375f1c",
+    "acc9b87c-583b-5ba6-bc6f-b833d2e8d2cb",
+    "9b1cf5ca-d793-5c2a-a2db-c88f44ac6ec4",
+    "ce911802-af16-57a4-90e8-e3257a9ee7af"
+  ],
+  "contexts": [
+    "'&'.+* .%(\"'.+ * $$* ! \f\r \t\f\u000b '&'.+* .%(\"'.+ * $$*\t\u000b r Figure 2. Impact of type 1 diabetes (T1D) genome- wide association studies (GWAS) single- nucleotide polymorphisms (SNPs) on immune phenotypes.  (A)Quantile- quantile (Q- Q) plots of quantitative trait locus (QTL) profiles of 62 T1D GWAS loci grouped by cell populations. The distribution of p- values",
+    "diseases, including T2D. Many of the module-QTL locioverlap with GWAS hits for immune-related pheno- types, suggesting that the modules described here might be of importance in the context of inflammatory dis- eases. Similar analyses should be performed for co- expression modules in other more T2D-relevant tissues to provide further insight into the causal networks underlying T2D aetiology. Similarly, network rewiring in T2D might be more strongly detectable in other tissues",
+    "(58)], revealing some interesting possible candidate functionalgenes other than those associated with the HLA and related sys-tems. In addition, early GWAS on type 1 diabetes by Todd et al.(23) revealed suggestive functional effects of non-HLA variants involved in immune functions. Another interesting application of",
+    "Research article       Genetics and Genomics | Medicine Chu, Janssen, Koenen etal. eLife 2022;11:e73709. DOI: https://doi.org/10.7554/eLife.73709  9 of 17Genetic regulation of immune phenotypes in T1D To further explore potential genetic regulation of immune phenotypes on the whole- genome level,  we performed QTL mapping in 300DM. This identified nine genome- wide significant QTLs (p- value  < 5  108) associated with immune- cell proportion, including four associated with T cell subpopu-",
+    "studies (r2> 0.8) and performed a chi- square test on clinical status by using PLINK 1.9. Samples in  300DM were taken as cases and samples in 500FG as controls. Impact of T1D GWAS loci on immune phenotypes To detect the impact of T1D GWAS loci on immune- cell populations, we grouped all traits into four  categories (B cells, T cells, monocytes, and NK cells), and counted the number of suggestive associ- ations (p- value < 0.05) between the 63 top SNPs from T1D GWAS loci and immune- cell traits. 1000",
+    "In the present study, we interrogated GWAS data sets on CD, UC and T1D for known susceptibility loci implicated inthese diseases. Our comparative analysis serves several impor-tant roles: rst, the ability to identify additional susceptibilityloci for one disease by testing known loci for another disease,similar to previous studies ( 12,13). This approach increases statistical power by limiting the number of hypotheses",
+    "Conclusions A major challenge is to translate GWAS ndings intocausal variants and target genes. The Immunochipeffort has greatly contributed to our understanding of disease mechanisms by identifying pathways, which could not be linked to diabetes by existing hypotheticalmodels. Diabetes is probably a much more diverse disease than the current subdivision into T1DM and T2D implies and a more precise subdivisioninto subgroups may also pave the way for a more",
+    "edge of the role(s) of genetic variation (SNPs) in population-level sus-ceptibility to T1D ( Ram et al., 2016a ). However, GWAS analyses do not automatically determine the particular gene(s) in a speci c locus that are mechanistically associated with disease pathogenesis, or elucidate the manner in which disease gene(s) interact ( Zhong  et al., 2010). The diculty associated with ascribing functional impacts to SNPs is partly explained by the fact that most disease-associated SNPs identi ed by",
+    "(Supplementary file 1C). We next investigated whether these genetic risk loci for T1D affect immune parameters and func- tion. The quantile- quantile plot of the association of the 63 T1D GWAS loci with different cell types  and cytokines illustrates an inflated deviation from an expected uniform distribution (Figure 2A,  Figure2figure supplement 1). We further tested whether this deviation can be explained by chance",
+    "Fadason et al. demonstrated that functionally relevant type 2 diabetes- associated SNPs are spatially linked with speci c changes in the ex- pression levels of genes within disease-associated tissues ( Fadason et al., 2017 ). Similarly, a study demonstrated that integrating chro- matin interactions with GWAS analyses is important in elucidatingcausal genes that modulate regulatory networks in autoimmune dis- eases ( McGovern et al., 2016). As such, the spatial organization of DNA"
+  ],
+  "answer": "Yes, the landscape of QTL (Quantitative Trait Loci) and GWAS (Genome-Wide Association Studies) hits can be used to dissect the role of the immune system in diabetes and its complications. Several pieces of evidence from the provided context support this:  1. **Impact on Immune Phenotypes**: GWAS SNPs for type 1 diabetes (T1D) have been shown to impact immune phenotypes. For example, QTL profiles of 62 T1D GWAS loci grouped by cell populations reveal the distribution of p-values, indicating significant associations between these loci and immune cell traits [1].  2. **Overlap with Immune-Related Phenotypes**: Many module-QTL loci overlap with GWAS hits for immune-related phenotypes, suggesting that these genetic modules are important in the context of inflammatory diseases, including diabetes [2].  3. **Genetic Regulation of Immune Phenotypes**: QTL mapping in a study identified nine genome-wide significant QTLs associated with immune-cell proportions, including T cell subpopulations, indicating a genetic regulation of immune phenotypes in T1D [4].  4. **Impact on Immune-Cell Populations**: Analysis of T1D GWAS loci showed suggestive associations between top SNPs and immune-cell traits, categorized into B cells, T cells, monocytes, and NK cells, further highlighting the impact of these loci on immune cell populations [5].  5. **Comparative Analysis of Susceptibility Loci**: Comparative analysis of GWAS data sets for diseases like T1D, Crohn's disease (CD), and ulcerative colitis (UC) helps identify additional susceptibility loci and increases statistical power, which is crucial for understanding the genetic basis of immune-related complications in diabetes [6].  6. **Pathway Identification**: The Immunochip effort has contributed to understanding disease mechanisms by identifying pathways linked to diabetes, which were not previously associated with the disease, indicating the complexity and diversity of diabetes and its immune-related aspects [7].  7. **Functional Impacts of SNPs**: Although GWAS analyses do not automatically determine the specific genes associated with disease pathogenesis, they provide insights into how disease genes interact and affect immune parameters and functions [8], [9].  In summary, the integration of QTL and GWAS data provides valuable insights into the genetic regulation of immune phenotypes and their role in diabetes and its complications, supporting the use of these landscapes for dissecting the immune system's involvement in the disease.",
+  "question": "Can the landscape of QTL and GWAS hits be used to dissect the role of immune system in diabetes and complications?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_2 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_2
new file mode 100644
index 0000000..6b99815
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_2
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Recent Developments in the Genetic and Genomic Basis of Type 2 Diabetes.pdf",
+    "2017 - Type 1 diabetes mellitus.pdf",
+    "2010 - Genetics of Type 1 Diabetes What\u2019s Next.pdf",
+    "2018 - Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls.pdf",
+    "2011 - Type 2 diabetes and obesity genomics and the clinic.pdf",
+    "2010 - Genome-wide analysis of transcriptional regulation in the murine liver.pdf",
+    "2010 - Genome-wide analysis of transcriptional regulation in the murine liver.pdf",
+    "2015 - Transcript Expression Data from Human.pdf",
+    "2004 - Interaction and Association Analysis of a Type 1 Diabetes Susceptibility Locus.pdf",
+    "2019 - IRS1\u2010 rs10498210 GA and CCR5\u201059029 AG polymorphisms in patients with type 2 diabetes in Kurdistan.pdf"
+  ],
+  "extraction_id": [
+    "1213249d-8ed3-5d13-9137-f11b87a7a78b",
+    "39b6a474-b721-509f-bbc3-094dc1f49634",
+    "5557d2db-b55a-59c9-8fe7-89b196a28617",
+    "43eecb5d-aca2-5c3e-9351-afbef000a795",
+    "10685e4c-eb4c-562a-a64a-d98e83c12c0b",
+    "151aa443-b9af-55db-9a30-adc4440ac7ef",
+    "7cbef74a-2d81-5a3a-a4d4-dfacdb86e632",
+    "867d0b1b-16a1-53ea-b014-3c204b9001a5",
+    "d71343b2-f7c5-52b4-96f9-bcc98f97fe81",
+    "843f5b60-2702-59e4-b237-02d002200e6d"
+  ],
+  "document_id": [
+    "7d051350-d939-5183-be22-742727573a75",
+    "8e8b9b6e-8dfb-5aae-8c61-5f53bd4e0242",
+    "261cbb40-ed6b-554c-a70d-db6b9f14cf74",
+    "af63c74d-a204-5f9f-9a32-3451b112e5ba",
+    "5086a525-124e-5a45-b75a-657d67a3250a",
+    "8a115c1b-662c-5062-b77f-bbde0532bbe9",
+    "8a115c1b-662c-5062-b77f-bbde0532bbe9",
+    "2b30d4f3-9ec3-574f-9a36-709b0e09c3f2",
+    "4246f8d0-69e8-56cf-9674-d379467dfb61",
+    "18afbfee-ddee-54b3-88cc-342812a65d09"
+  ],
+  "id": [
+    "chatcmpl-ADYmdeow5Femrvb7YWgDS0ML3p64y",
+    "6f00dd31-490e-53cd-81b3-c56e13bd7edd",
+    "fa4e127f-2c54-592b-a478-152bc74e7351",
+    "f0c9d05b-7999-5cb7-bb48-0666cf74aec0",
+    "55dca975-78ec-594d-8a30-a0849b683089",
+    "126bf287-0f5e-52a9-abac-ad59ad3ea153",
+    "90565c2b-fdb6-5b0f-a710-9086a4cfcd2b",
+    "ceb7bd13-b917-566f-8e17-40dd523afd42",
+    "226e2873-a0bf-554d-9576-7fca5f2ffc0f",
+    "a495dcc8-5cee-58a9-9f15-95be8fbc9b6a",
+    "997a967e-6428-51c9-9847-24d16f11f9f1"
+  ],
+  "contexts": [
+    "associated with increased fasting plasma glucose levels and type2 diabetes risk. Nat Genet. 2009;41(1):89 94. 23. Rees M, Wincovitch S, Schultz J, Waterstradt R, Beer N, Baltrusch S, et al. Cellular characterisation of the GCKR P446L variant associated with type 2 diabe tes risk. Diabetologia. 2012;55 (1):114 22. 24. Nejentsev S, Walker N, Riches D, Egholm M, Todd J, et al. Rare variants of IFIH1 , a gene implicated in antiviral responses, protect against type 1 diabetes. Science. 2009;324(5925):387 9.",
+    "HLAlinked genes in juvenile diabetes mellitus.  Br.Med. J. 3, 133135 (1975). 52. Erlich,H.A.  etal.  Next generation sequencing reveals  the association of DRB3*02:02 with type 1 diabetes.  Diabetes  62, 26182622 (2013). 53. CaillatZucman,S.  etal.  Agedependent HLA genetic  heterogeneity of type1 insulindependent diabetes  mellitus. J.Clin. Invest. 90, 22422250 (1992). 54. Cucca,F.  etal.  The distribution of DR4 haplotypes  inSardinia suggests a primary association of typeI",
+    "holdt R, Akolkar B, Erlich HA, Hilner JE, Julier C, Morahan G, Nerup J,Nierras CR, Chen WM, Rich SS, Type 1 Diabetes Genetics Consortium. Ahuman type 1 diabetes susceptibility locus maps to chromosome 21q22.3.Diabetes 2008;57:2858 2861 58. Nejentsev S, Walker N, Riches D, Egholm M, Todd JA. Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1diabetes. Science 2009;324:387389 59. Altshuler D, Daly M. Guilt beyond a reasonable doubt. Nat Genet 2007;39: 813 815",
+    "because of their presumed roles in immune signalling, considered to be a major feature of T1D-susceptibility. These include ERBB3 (receptor tyrosine-protein kinase erbB-3 precursor) at 12q13 and SH2B3/LNK (SH2B adaptor protein 3), TRAFD1 (TRAF-type zinc finger domain containing 1) and PTPN11 (protein tyrosine phos- phatase, non-receptor type 11) at 12q24. For these signal regions in",
+    "Nejentsev S, Walker N, Riches D, Egholm M, Todd JA (2009) Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1 diabetes. Science 324:387389 Nicolson TJ, Bellomo EA, Wijesekara N, Loder MK, Baldwin JM, Gyulkhandanyan AV, Koshkin V, Tarasov AI, Carzaniga R, Kronenberger K, Taneja TK, da Silva Xavier G, Libert S,",
+    "7   (Wellcome Trust Case Control Consortium 2007) .  Separate work that examined liver gene  expression in a smaller cohort of human samples with and without Type I diabetes found  that ERBB3  did not have a cis -eQTL but that a flanking gene, R PS26, did.  Since the disease  phenotype and RPS26  both had QTLs in the same location, this suggested the RPS26  was a  stronger candidate than ERBB3 .  The authors then used mouse liver and adipose expression",
+    "models.  A genome wide association study in a large human population proposed the  receptor typrosine kinase ERBB3  as the best candidate gene near a QTL for Type I diabetes",
+    "61. Nejentsev S, Walker N, Riches D, Egholm M, Todd JA (2009) Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1 diabetes. Science 324: 387 389. doi: 10.1126/science. 1167728 PMID: 19264985 62. Nica AC, Ongen H, Irminger JC, Bosco D, Berney T, et al. (2013) Cell-type, allelic, and genetic signa- tures in the human pancreatic beta cell transcriptome. Genome Res 23: 1554 1562. doi: 10.1101/gr. 150706.112 PMID: 23716500",
+    "gene is associated with insulin-dependent diabetes mellitus. Diabetes 33:176 183, 1984 3. Nistico L, Buzzetti R, Pritchard L, Van der Auwera B, Giovannini C, Bosi E, Larrad M, Rios M, Chow C, Cockram C, Jacobs K, Mijovic C, Bain S,Barnett A, Vandewalle C, Schuit F, Gorus F, Tosi R, Pozzilli P, Todd J: TheCTLA-4 gene region of chromosome 2q33 is linked to, and associated with,type 1 diabetes: Belgian Diabetes Registry. Hum Mol Genet 5:1075 1080, 1996",
+    "One of these genes associated with type 2 diabetes is the  insulin receptor substrate 1 (IRS1, OMIM association num-ber, 147545) (Alharbi, Khan, Abotalib, & AlHakeem, 2014; Alharbi, Khan, Munshi et al., 2014; Brender et al., 2013;  Brunetti, Chiefari, & Foti, 2014) and another is the CC motif chemokine receptor5(CCR5, OMIM association num-ber, 601373) (Balistreri et al., 2007; Mokubo et al., 2006;  Muntinghe et al., 2009). Insulin initiates a wide range of growth and metabolic ef-"
+  ],
+  "answer": "Yes, the genes SH2B3, IFIH1, and ERBB3 are related to diabetes:  1. **IFIH1**: This gene is implicated in antiviral responses and has been found to protect against type 1 diabetes [1], [3], [5], [8]. 2. **ERBB3**: This gene is considered to have a role in immune signaling, which is a major feature of type 1 diabetes susceptibility [4]. Additionally, it has been proposed as a candidate gene near a QTL for type 1 diabetes [7]. 3. **SH2B3**: This gene is also considered to have a role in immune signaling related to type 1 diabetes susceptibility [4].  Thus, all three genes, SH2B3, IFIH1, and ERBB3, are related to diabetes.",
+  "question": "Is any of the genes SH2B3, IFIH1 or ERBB3  related to diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_3 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_3
new file mode 100644
index 0000000..2951e1a
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_3
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2010 - The Genetics of Type 2 Diabetes.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf"
+  ],
+  "extraction_id": [
+    "c9f99f8c-8432-59fa-8392-38eba8a63d8b",
+    "55a150e2-71ee-5b88-bbbb-22eecc2ff5e4",
+    "1e4b6f6f-e67a-53ef-8af6-78aa4c9ce112",
+    "23654fd6-b738-5c3b-9e6b-9b29ea47df38",
+    "76e4f0c7-ffb4-5643-b8d7-46cc584cfbf8",
+    "2618e650-f07c-5d21-a8f9-8f0dcdd51fd6",
+    "2618e650-f07c-5d21-a8f9-8f0dcdd51fd6",
+    "2618e650-f07c-5d21-a8f9-8f0dcdd51fd6",
+    "4f65affc-d78a-5048-86f3-22c0e12d298c",
+    "4f65affc-d78a-5048-86f3-22c0e12d298c"
+  ],
+  "document_id": [
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "a1d211d4-279e-51d7-b2b2-33bc2763d089",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40"
+  ],
+  "id": [
+    "chatcmpl-ADYmiyBbswpmJDsIM9Hrs7jxPruIX",
+    "430a5d63-0af9-569e-b9db-2cc4262f1d77",
+    "0337dd9b-7228-5664-aff3-fa20c0f04d6d",
+    "c6c8b56a-1f17-5d96-bea4-787673f11be9",
+    "e5ac3a01-3a53-5239-b1d0-26ee4e72f37b",
+    "b9f46fff-157d-5007-ae86-987d9b5022b5",
+    "0940a12a-fee5-57ca-9a9f-ce720b43119e",
+    "0752929d-fb98-5c2f-b47f-e493f25ac70d",
+    "d8db4432-bb66-59de-bb9e-c0667ec9010c",
+    "f5909a51-0d41-5aee-ac5a-8d47550ef094",
+    "adf9b377-f569-5f08-be4b-4d9d1913990c"
+  ],
+  "contexts": [
+    "understood. It seems that interactions between multiple genes and environmental factors may play a role.  One of these factors is dietary factors. There is evidence supporting the role  of nutrient- gene interactions   in DM pathophysiology  [5]. Thus, a greater understanding of potential gene -nutrient interactions may  be relevant for DM prevention and treatment.  Nutrigenetics and nutrigenomics are defined as the science of the effects of genetic variation on",
+    "nutrition  [12] . The identi  cation of gene variants that contribute  both to variation in fetal growth and to the susceptibility to T2DM, however, suggests that this metabolic   programming   could also be partly genetically determined  [13] .   These complex interactions between genes and environment  complicate the task of identifying any single genetic susceptibility factor for T2DM. Three general approaches have been adopted",
+    "Nutrients 2014, 6 5340    However, while the a pplication of these technologies is becoming more accessible, analysis of the  complex large data sets that are generated  presents multiple challenges.   The aim of the present review was to provide insights regarding the role of nutrient -gene interactions  in DM pathogenesis, prevention and treatment. In addition, we explored how an individuals genetic  makeup can affect nutrient metabolism and the response to nutrient intake, potentially leading to DM.",
+    "Nutrients 2014, 6 5343    3. Gene -Nutrient or Dietary Patter n Interactions in T he Development of T2DM   Recently, several studies have d emonstrated the  significant effects of genotype by environment  interactions on T2D M [48,49] . However, further clarification of the role of these interactions at the  genome -wide level could help predict disease risk more accurately and facilitate the development of",
+    "in nutritional epidemiology: applications, needs and  new horizons .Hum Genet 125, 507525. Kaput, J., Noble, J., Hatipoglu, B., et al. ( 2007) Application of nutrigenomic concepts to type 2 diabetes melli-tus.Nutr Metab Cardiovasc Dis 17,89103. Ordovas, J.M., Kaput, J., and Corella, D. ( 2007) Nutrition in the genomics era: cardiovascular disease risk and  the Mediterranean diet .Mol Nutr Food Res 51, 12931299. van Ommen, B., El-Sohemy , A., Hesketh, J., et al . ( 2010)",
+    "dietary patterns according to genetic variations, the role of gene -nutrient interactions, gene - diet-phenotype interactions and epigenetic modifications caused by nutrients; these studies  will facilitate an understanding of the early molecular events that occur in DM and will contribute to the identification of better biomarke rs and diagnostics tools. In particular, this",
+    "Abstract:  Diabetes mellitus (DM) is considered a global pandemic, and the incidence of  DM continues to grow worldwide. Nutrients and dietary patterns are central issues in the  prevention, development and treatment of this disease. The pathogenesis of DM is not  comp letely understood, but nutrient -gene interactions at different levels, genetic predisposition  and dietary factors appear to be involved. Nutritional genomics studies generally focus on",
+    "approach will help to develop tailored diets that maximize the use of nutrients and other functional ingredients present in food, which will aid in the prevention and delay of DM and its complications. This rev iew discusses the current state of nutrigenetics, nutrigenomics and  epigenomics research on DM. Here, we provide an overview of the role of gene variants and nutrient interactions, the importance of nutrients and dietary patterns on gene expression, OPEN ACCESS",
+    "It was previously report ed that food intake is a key component that affects the incidence of DM. Thus,  the identification and analysis of nutrient/gene interactions are necessary steps to understand DM etiopathogenesis. In general, nutrients can affect gene expression via different mechanisms: ( i) directly;  (ii) through their metabolites and ( iii) through signal tran sduction molecules (Figure 1).",
+    "Nutrients 2014, 6 5347    3.4. Importance of Genotype by Macronutrient Interactions for T2DM -Related Traits   Recently, using genome -wide complex trait anal ysis, the genome -environment contribution of   14 dietary factors (glycemic load, total energy, protein, total fat, SF A, MUFA, PUFA, n- 3 PUFA,   n-6 PUFA, n-3:n-6 PUFA, carbohydrate, alcohol intake, trans fat and fiber) to the total phenotypic  variance of 4 T2DM -related traits (fasting glucose, fasting insulin, HOMA -IR and HOMA of  cell"
+  ],
+  "answer": "Genomics can be used to better understand the nutritional factors of diabetes through the study of nutrient-gene interactions, which play a significant role in the pathophysiology of diabetes mellitus (DM) [1]. Nutrigenetics and nutrigenomics are fields that focus on how genetic variations affect nutrient metabolism and the body's response to nutrient intake, which can influence the development and progression of DM [3].   By identifying gene variants that contribute to both fetal growth and susceptibility to Type 2 Diabetes Mellitus (T2DM), researchers can better understand how metabolic programming is partly genetically determined [2]. This understanding can help in predicting disease risk more accurately and developing personalized dietary recommendations to prevent and manage DM [4].  Moreover, studying gene-nutrient interactions can reveal early molecular events in DM, contributing to the identification of better biomarkers and diagnostic tools [6]. This approach can also facilitate the development of tailored diets that maximize the use of nutrients and functional ingredients in food, aiding in the prevention and delay of DM and its complications [8].   Overall, genomics provides insights into how nutrients can affect gene expression through various mechanisms, such as directly, through their metabolites, or through signal transduction molecules [9]. This comprehensive understanding can lead to more effective prevention and treatment strategies for diabetes based on individual genetic profiles.",
+  "question": "nutrition is a factor for diabetes. how can genomics be use to better understand nutritional factors of diabetes"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_4 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_4
new file mode 100644
index 0000000..f140ead
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_4
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2018 - Ethical_Social_and_Legal_Consequences.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2007 - Bioethnic Conscription Genes, Race.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2010 - The Genetics of Type 2 Diabetes.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf"
+  ],
+  "extraction_id": [
+    "2618e650-f07c-5d21-a8f9-8f0dcdd51fd6",
+    "6cc02cc8-628d-58b8-b9d3-8fc049773c22",
+    "1e4b6f6f-e67a-53ef-8af6-78aa4c9ce112",
+    "76e4f0c7-ffb4-5643-b8d7-46cc584cfbf8",
+    "79223305-26a8-51d5-a962-bb7d925810d7",
+    "2618e650-f07c-5d21-a8f9-8f0dcdd51fd6",
+    "c9f99f8c-8432-59fa-8392-38eba8a63d8b",
+    "23654fd6-b738-5c3b-9e6b-9b29ea47df38",
+    "55a150e2-71ee-5b88-bbbb-22eecc2ff5e4",
+    "4f65affc-d78a-5048-86f3-22c0e12d298c"
+  ],
+  "document_id": [
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "3fb80410-0b56-5c01-b3d6-9388b6029a77",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "d90126d9-fd87-5b38-87f7-08415f690836",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "a1d211d4-279e-51d7-b2b2-33bc2763d089",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40"
+  ],
+  "id": [
+    "chatcmpl-ADYmomdJKgdvZlepHclGpceBX99vV",
+    "0752929d-fb98-5c2f-b47f-e493f25ac70d",
+    "20d914cb-135d-5033-adc4-61aa7468f8df",
+    "c6c8b56a-1f17-5d96-bea4-787673f11be9",
+    "b9f46fff-157d-5007-ae86-987d9b5022b5",
+    "9bbce823-83c5-5258-af26-f79575042496",
+    "d8db4432-bb66-59de-bb9e-c0667ec9010c",
+    "430a5d63-0af9-569e-b9db-2cc4262f1d77",
+    "e5ac3a01-3a53-5239-b1d0-26ee4e72f37b",
+    "0337dd9b-7228-5664-aff3-fa20c0f04d6d",
+    "f5909a51-0d41-5aee-ac5a-8d47550ef094"
+  ],
+  "contexts": [
+    "Abstract:  Diabetes mellitus (DM) is considered a global pandemic, and the incidence of  DM continues to grow worldwide. Nutrients and dietary patterns are central issues in the  prevention, development and treatment of this disease. The pathogenesis of DM is not  comp letely understood, but nutrient -gene interactions at different levels, genetic predisposition  and dietary factors appear to be involved. Nutritional genomics studies generally focus on",
+    "ABSTRACT    Genomics has contributed to a better understanding of many disorders including diabetes. The  following article looks at the ethical, social and legal consequences of genomic medicine and  predictive genetic testing for diabetes. This is currently a field in its nascent stage and developing  rapidly all over the world. The various ethical facets of genomic medicine in diabetes like its effects",
+    "Nutrients 2014, 6 5340    However, while the a pplication of these technologies is becoming more accessible, analysis of the  complex large data sets that are generated  presents multiple challenges.   The aim of the present review was to provide insights regarding the role of nutrient -gene interactions  in DM pathogenesis, prevention and treatment. In addition, we explored how an individuals genetic  makeup can affect nutrient metabolism and the response to nutrient intake, potentially leading to DM.",
+    "in nutritional epidemiology: applications, needs and  new horizons .Hum Genet 125, 507525. Kaput, J., Noble, J., Hatipoglu, B., et al. ( 2007) Application of nutrigenomic concepts to type 2 diabetes melli-tus.Nutr Metab Cardiovasc Dis 17,89103. Ordovas, J.M., Kaput, J., and Corella, D. ( 2007) Nutrition in the genomics era: cardiovascular disease risk and  the Mediterranean diet .Mol Nutr Food Res 51, 12931299. van Ommen, B., El-Sohemy , A., Hesketh, J., et al . ( 2010)",
+    "at the expense of understanding the social context and determinants of the disease.Biogenetic views tend to trump sociological views in the diabetes research imaginary ofconsortium members. However, the genetic epidemiologists who make up part of thediabetes consortium are not ignorant of the effects of proper diet and adequate exercise.Take away the television and the automobile and diabetes would all but disappear, quipped the head of one lab. Neither are researchers unsympathetic to those who sufferfrom",
+    "approach will help to develop tailored diets that maximize the use of nutrients and other functional ingredients present in food, which will aid in the prevention and delay of DM and its complications. This rev iew discusses the current state of nutrigenetics, nutrigenomics and  epigenomics research on DM. Here, we provide an overview of the role of gene variants and nutrient interactions, the importance of nutrients and dietary patterns on gene expression, OPEN ACCESS",
+    "understood. It seems that interactions between multiple genes and environmental factors may play a role.  One of these factors is dietary factors. There is evidence supporting the role  of nutrient- gene interactions   in DM pathophysiology  [5]. Thus, a greater understanding of potential gene -nutrient interactions may  be relevant for DM prevention and treatment.  Nutrigenetics and nutrigenomics are defined as the science of the effects of genetic variation on",
+    "Nutrients 2014, 6 5343    3. Gene -Nutrient or Dietary Patter n Interactions in T he Development of T2DM   Recently, several studies have d emonstrated the  significant effects of genotype by environment  interactions on T2D M [48,49] . However, further clarification of the role of these interactions at the  genome -wide level could help predict disease risk more accurately and facilitate the development of",
+    "nutrition  [12] . The identi  cation of gene variants that contribute  both to variation in fetal growth and to the susceptibility to T2DM, however, suggests that this metabolic   programming   could also be partly genetically determined  [13] .   These complex interactions between genes and environment  complicate the task of identifying any single genetic susceptibility factor for T2DM. Three general approaches have been adopted",
+    "It was previously report ed that food intake is a key component that affects the incidence of DM. Thus,  the identification and analysis of nutrient/gene interactions are necessary steps to understand DM etiopathogenesis. In general, nutrients can affect gene expression via different mechanisms: ( i) directly;  (ii) through their metabolites and ( iii) through signal tran sduction molecules (Figure 1)."
+  ],
+  "answer": "Diabetes mellitus (DM) is a global pandemic with increasing incidence worldwide, where nutrients and dietary patterns play a crucial role in its prevention, development, and treatment [1]. The pathogenesis of DM involves complex interactions between genetic predisposition and dietary factors, which are not yet fully understood [1]. Genomics offers a promising avenue to elucidate these interactions by studying nutrient-gene interactions at various levels [1]. Nutritional genomics, including nutrigenetics and nutrigenomics, focuses on how genetic variations affect nutrient metabolism and the body's response to nutrient intake, potentially influencing DM pathogenesis [3]. By identifying gene variants and understanding their interactions with dietary factors, genomics can help predict disease risk more accurately and facilitate the development of personalized dietary interventions [8]. This approach aims to develop tailored diets that maximize the use of nutrients and functional ingredients in food, aiding in the prevention and delay of DM and its complications [6]. Overall, genomics provides valuable insights into the role of gene-nutrient interactions in DM, paving the way for more effective prevention and treatment strategies [7].",
+  "question": "nutrition is a factor for diabetes. construct an abstract about how can genomics be use to better understand nutritional factors of diabetes"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_5 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_5
new file mode 100644
index 0000000..2939139
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_5
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - Pathomechanisms of Type 2 Diabetes Genes.pdf",
+    "2015 - Diabetes mellitus The epidemic of the century.pdf",
+    "2012 - Type 2 Diabetes Genetics Beyond GWAS.pdf",
+    "2007 - Recent development in pharmacogenomics from candidate genes to genome-wide association studies.pdf",
+    "2014  - Dorothy Hodgkin Lecture 2014 Understanding genes identified by genome\u2010wide association.pdf",
+    "2015 - Diabetes mellitus The epidemic of the century.pdf",
+    "2015 - Type 2 Diabetes Mellitus and the Association of Candidate Genes.pdf",
+    "2007 - A German genome-wide linkage scan for type 2 diabetes supports the existence of a metabolic syndrome locus on chromosome 1p36.13 and a type 2 diabetes locus on chromosome 16p12.pdf",
+    "2014 - Nutrigenetics and Nutrigenomics Insights into Diabetes Etiopathogenesis.pdf",
+    "2007 - TCF7L2 the biggest story in diabetes genetics since HLA.pdf"
+  ],
+  "extraction_id": [
+    "eff1d167-9689-5c26-9a12-c66714696d86",
+    "36f9d4f2-293e-53e3-8b4b-12571af6669a",
+    "a3a875fa-e55b-52d0-b9bf-72b96330c393",
+    "f2fa55c2-fbca-5f7b-a744-deb279bf9369",
+    "86253f12-bb43-5236-bfb1-df5dff759f6d",
+    "7cfe9f29-a0ee-56d3-be3b-1b238a43bc07",
+    "5ffb710d-ca19-5415-bbb6-34b3f85bf47f",
+    "198c5f2d-fc43-5744-9cd8-4222c8fa8ab8",
+    "25187f10-04b3-51c6-8f4c-d4e480353fa2",
+    "0f7bd536-46b9-52e2-927e-a8309d541066"
+  ],
+  "document_id": [
+    "cf8ec75c-8ffe-5baa-830d-ac7a4a5964bd",
+    "e114dd28-fd39-56df-bdeb-8806474a6c10",
+    "d59a38d7-889b-51b5-b896-c305c82a2169",
+    "fe012b74-6516-5503-a88a-dc8071869632",
+    "11d0cb98-a00f-53f1-92e3-e1be17002c02",
+    "e114dd28-fd39-56df-bdeb-8806474a6c10",
+    "3236fdee-e304-5b88-921f-52e52dc5afa3",
+    "f8a85060-6303-5228-ba89-7ee8701bda9f",
+    "ce4f171c-494c-53f2-a770-c3edd3561c40",
+    "018b8646-b894-5b7d-a8c8-932a2bc13ca8"
+  ],
+  "id": [
+    "chatcmpl-ADYmusmSLbuh68YcOiU3r0KXSi3Ub",
+    "3e678e3c-ad2c-5884-9c88-7f4d54c914bd",
+    "6617e15c-ab52-596c-b628-60ec5a7001e7",
+    "1eb3a215-002b-528b-a954-bb9e2419ea6f",
+    "e456e587-e172-5ae9-b68e-98e38c5052c2",
+    "5d936c2c-faf7-5b0f-92e1-c3f8f43b3011",
+    "ed5d8e9e-859e-5256-a7b5-468c1f7837a2",
+    "263f6b22-d314-5653-bbef-3f0e3e09839b",
+    "05e76af5-c67b-50ca-a06a-a603d6d4b35e",
+    "fc63f56e-f1fb-56e0-9e62-b4bdcefb5a53",
+    "c21b7f01-ff01-5561-8016-c4432d844baf"
+  ],
+  "contexts": [
+    "single nucleotide polymorphisms in TCF7L2 are reproduc-ibly associated with type 2 diabetes and reduce the insulinresponse to glucose in nondiabetic individuals. Diabetes55:28902895 135. Cauchi S, Meyre D, Dina C, Choquet H, Samson C, Gallina S, Balkau B, Charpentier G, Pattou F, StetsyukV, Scharfmann R, Staels B, Fru  hbeck G, Froguel P 2006 Transcription factor TCF7L2 genetic study in the Frenchpopulation: expression in human /H9252-cells and adipose tissue",
+    "L. Mechanisms by which common variants in the TCF7L2 gene  increase risk of type 2 diabetes. J Clin Invest  2007; 117: 2155-2163  [PMID: 17671651 DOI: 10.1172/JCI30706] 164 Gloyn AL , Braun M, Rorsman P. Type 2 diabetes susceptibility  gene TCF7L2 and its role in beta-cell function. Diabetes  2009; 58:  800-802 [PMID: 19336690 DOI: 10.2337/db09-0099] 165 da Silva Xavier G , Loder MK, McDonald A, Tarasov AI, Carzaniga  R, Kronenberger K, Barg S, Rutter GA. TCF7L2 regulates late",
+    "transcription factor 7-like 2 ( TCF7L2 ) gene confers risk of type 2 diabetes. Nat Genet. 2006; 38:320323. [PubMed: 16415884] 172. Gloyn AL, Noordam K, Willemsen MA, Ellard S, Lam WW, et al. Insights into the biochemical and genetic basis of glucokinase activation from naturally occurring hypoglycemia mutations. Diabetes. 2003; 52:24332440. [PubMed: 12941786] 173. Pearson ER, Donnelly LA, Kimber C, Whitley A, Doney AS, et al. Variation in TCF7L2",
+    "2 (TCF7L2 ) gene confers risk of Type 2  diabetes. Nat. Genet.  38(3), 320323  (2006). 143Florez JC, Jablonski KA, Bayley N et al.   TCF7L2 polymorphisms and progression to diabetes in the Diabetes Prevention Program. N. Engl. J. Med.  355(3),  241250 (2006). 144Damcott CM, Pollin TI, Reinhart LJ et al.   Polymorphisms in the transcription factor 7-like 2 ( TCF7L2 ) gene are associated with",
+    "rs7903146 and rs12255372 in intron 3 of the TCF7L2 gene [20], associated with a ~45% increase in Type 2 diabetes risk per allele. As such, the TCF7L2 locus presently repre- sents the strongest known genetic determinant of Type 2diabetes. Risk allele carriers show impaired insulin produc-tion [21] and b-cell dysfunction in vitro [22]. TCF7L2 (previously referred to as TCF-4) is a high-mobility group box-containing transcription factor involved in Wingless-type MMTV integration site (Wnt)",
+    "genes which also play a significant role in the risk and  pathogenesis of the disease[158,159]. The association  of TCF7L2  gene variants with type 2 diabetes and  its mechanism of action received special attention  by several investigators[161,162]. Over expression of the protein was shown to decrease the sensitivity of  beta islet cells to secrete insulin[163,164] and was more  precisely involved in the regulation of secretary granule  fusion that constitute a late event in insulin secretion",
+    "et al. Variant of transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2  diabetes. Nat Genet . 2006;38:320-23.   Sladek R, Rocheleau G, Rung J, Dina C, Shen L, Serre D, et al. A genome- [9] wide association study identifies novel risk loci for type 2 diabetes. Nature .  2007;445:881-85.  Kirchhoff K, Machicao F, Haupt A, Schafer SA, Tschritter O, Staiger H, et al. [10] Polymorphisms in the TCF7L2, CDKAL1 and SLC30A8 genes are associated",
+    "transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2diabetes. Nat Genet 38:320 3231422 Diabetologia (2007) 50:1418 1422",
+    "approximately double odds ratio compared to most other diabetes susceptibility polymorphisms. TCF7L2  is a transcription factor involved in the Wnt signaling pathway that is ubiquitously expressed,  and it has been observed that TCF7L2  risk alleles result in the overexpression of TCF7L2  in   pancreatic  cells. This overexpression causes reduced nutrient -induced insulin secretion, which results  in a direct predisposition to T2DM as well as an indirect predisp osition via an increase in hepatic glucose",
+    "diabetes. The gene seems to be widely expressed [ 18] and the transcription factor product is known to be involved in the Wnt signalling cascade. Current evidence strongly supports the idea that the predominant effect of TCF7L2 dysfunction on type 2 diabetes development is mediated through impairment of insulin secretion [ 11,1517,20], a finding that would be consistent, for example, with theknown effects of other (non-homologous) TCF genes (TCF1 [also known as HNF1A ] and TCF2 [also known as"
+  ],
+  "answer": "Yes, the gene TCF7L2 is involved in diabetes, specifically type 2 diabetes. Multiple sources in the provided context confirm this association:  1. Single nucleotide polymorphisms in TCF7L2 are reproducibly associated with type 2 diabetes and reduce the insulin response to glucose in nondiabetic individuals [1]. 2. Common variants in the TCF7L2 gene increase the risk of type 2 diabetes [2]. 3. The TCF7L2 gene confers risk of type 2 diabetes [3], [4]. 4. Variants in the TCF7L2 gene are associated with a significant increase in type 2 diabetes risk, representing the strongest known genetic determinant of the disease [5]. 5. Overexpression of TCF7L2 protein decreases the sensitivity of beta islet cells to secrete insulin, which is involved in the regulation of insulin secretion [6]. 6. TCF7L2 risk alleles result in overexpression in pancreatic cells, causing reduced nutrient-induced insulin secretion and a predisposition to type 2 diabetes [9]. 7. The predominant effect of TCF7L2 dysfunction on type 2 diabetes development is mediated through impairment of insulin secretion [10].  These references collectively support the involvement of TCF7L2 in the pathogenesis and risk of type 2 diabetes.",
+  "question": "Is the gene TCF7L2 involved in diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_6 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_6
new file mode 100644
index 0000000..dacccec
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_6
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Recent Developments in the Genetic and Genomic Basis of Type 2 Diabetes.pdf",
+    "2017 - Type 1 diabetes mellitus.pdf",
+    "2010 - Genetics of Type 1 Diabetes What\u2019s Next.pdf",
+    "2018 - Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls.pdf",
+    "2011 - Type 2 diabetes and obesity genomics and the clinic.pdf",
+    "2015 - Transcript Expression Data from Human.pdf",
+    "2010 - Genome-wide analysis of transcriptional regulation in the murine liver.pdf",
+    "2010 - Genome-wide analysis of transcriptional regulation in the murine liver.pdf",
+    "2013 - The CTRB12 Locus Affects Diabetes Susceptibility.pdf",
+    "2009 - Genome-Wide Linkage Scan in Gullah-Speaking African American Families.pdf"
+  ],
+  "extraction_id": [
+    "1213249d-8ed3-5d13-9137-f11b87a7a78b",
+    "39b6a474-b721-509f-bbc3-094dc1f49634",
+    "5557d2db-b55a-59c9-8fe7-89b196a28617",
+    "43eecb5d-aca2-5c3e-9351-afbef000a795",
+    "10685e4c-eb4c-562a-a64a-d98e83c12c0b",
+    "867d0b1b-16a1-53ea-b014-3c204b9001a5",
+    "151aa443-b9af-55db-9a30-adc4440ac7ef",
+    "7cbef74a-2d81-5a3a-a4d4-dfacdb86e632",
+    "97905c56-72e3-546d-ba0e-dfe1023f0c27",
+    "766aa9ab-312f-5d62-be31-860ba0697180"
+  ],
+  "document_id": [
+    "7d051350-d939-5183-be22-742727573a75",
+    "8e8b9b6e-8dfb-5aae-8c61-5f53bd4e0242",
+    "261cbb40-ed6b-554c-a70d-db6b9f14cf74",
+    "af63c74d-a204-5f9f-9a32-3451b112e5ba",
+    "5086a525-124e-5a45-b75a-657d67a3250a",
+    "2b30d4f3-9ec3-574f-9a36-709b0e09c3f2",
+    "8a115c1b-662c-5062-b77f-bbde0532bbe9",
+    "8a115c1b-662c-5062-b77f-bbde0532bbe9",
+    "5cd422c9-d4bc-5a96-8af8-00561458e67b",
+    "bd1d6b45-3929-5bd8-a677-d143381a7da5"
+  ],
+  "id": [
+    "chatcmpl-ADYmzCY2UpsFw8AXX7kNO9LvCTBMY",
+    "6f00dd31-490e-53cd-81b3-c56e13bd7edd",
+    "fa4e127f-2c54-592b-a478-152bc74e7351",
+    "f0c9d05b-7999-5cb7-bb48-0666cf74aec0",
+    "55dca975-78ec-594d-8a30-a0849b683089",
+    "126bf287-0f5e-52a9-abac-ad59ad3ea153",
+    "226e2873-a0bf-554d-9576-7fca5f2ffc0f",
+    "90565c2b-fdb6-5b0f-a710-9086a4cfcd2b",
+    "ceb7bd13-b917-566f-8e17-40dd523afd42",
+    "487d6a88-44ef-520e-a910-5b4b89416880",
+    "d4d61f22-5ba2-5ef1-a497-167894bf1c7f"
+  ],
+  "contexts": [
+    "associated with increased fasting plasma glucose levels and type2 diabetes risk. Nat Genet. 2009;41(1):89 94. 23. Rees M, Wincovitch S, Schultz J, Waterstradt R, Beer N, Baltrusch S, et al. Cellular characterisation of the GCKR P446L variant associated with type 2 diabe tes risk. Diabetologia. 2012;55 (1):114 22. 24. Nejentsev S, Walker N, Riches D, Egholm M, Todd J, et al. Rare variants of IFIH1 , a gene implicated in antiviral responses, protect against type 1 diabetes. Science. 2009;324(5925):387 9.",
+    "HLAlinked genes in juvenile diabetes mellitus.  Br.Med. J. 3, 133135 (1975). 52. Erlich,H.A.  etal.  Next generation sequencing reveals  the association of DRB3*02:02 with type 1 diabetes.  Diabetes  62, 26182622 (2013). 53. CaillatZucman,S.  etal.  Agedependent HLA genetic  heterogeneity of type1 insulindependent diabetes  mellitus. J.Clin. Invest. 90, 22422250 (1992). 54. Cucca,F.  etal.  The distribution of DR4 haplotypes  inSardinia suggests a primary association of typeI",
+    "holdt R, Akolkar B, Erlich HA, Hilner JE, Julier C, Morahan G, Nerup J,Nierras CR, Chen WM, Rich SS, Type 1 Diabetes Genetics Consortium. Ahuman type 1 diabetes susceptibility locus maps to chromosome 21q22.3.Diabetes 2008;57:2858 2861 58. Nejentsev S, Walker N, Riches D, Egholm M, Todd JA. Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1diabetes. Science 2009;324:387389 59. Altshuler D, Daly M. Guilt beyond a reasonable doubt. Nat Genet 2007;39: 813 815",
+    "because of their presumed roles in immune signalling, considered to be a major feature of T1D-susceptibility. These include ERBB3 (receptor tyrosine-protein kinase erbB-3 precursor) at 12q13 and SH2B3/LNK (SH2B adaptor protein 3), TRAFD1 (TRAF-type zinc finger domain containing 1) and PTPN11 (protein tyrosine phos- phatase, non-receptor type 11) at 12q24. For these signal regions in",
+    "Nejentsev S, Walker N, Riches D, Egholm M, Todd JA (2009) Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1 diabetes. Science 324:387389 Nicolson TJ, Bellomo EA, Wijesekara N, Loder MK, Baldwin JM, Gyulkhandanyan AV, Koshkin V, Tarasov AI, Carzaniga R, Kronenberger K, Taneja TK, da Silva Xavier G, Libert S,",
+    "61. Nejentsev S, Walker N, Riches D, Egholm M, Todd JA (2009) Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1 diabetes. Science 324: 387 389. doi: 10.1126/science. 1167728 PMID: 19264985 62. Nica AC, Ongen H, Irminger JC, Bosco D, Berney T, et al. (2013) Cell-type, allelic, and genetic signa- tures in the human pancreatic beta cell transcriptome. Genome Res 23: 1554 1562. doi: 10.1101/gr. 150706.112 PMID: 23716500",
+    "7   (Wellcome Trust Case Control Consortium 2007) .  Separate work that examined liver gene  expression in a smaller cohort of human samples with and without Type I diabetes found  that ERBB3  did not have a cis -eQTL but that a flanking gene, R PS26, did.  Since the disease  phenotype and RPS26  both had QTLs in the same location, this suggested the RPS26  was a  stronger candidate than ERBB3 .  The authors then used mouse liver and adipose expression",
+    "models.  A genome wide association study in a large human population proposed the  receptor typrosine kinase ERBB3  as the best candidate gene near a QTL for Type I diabetes",
+    "and 16p13.2 (near TMEM114 ) have not previously been implicated in b-cell function, type 2 diabetes susceptibility, or related phenotypes. However, in publically available gene expression data from the MuTHER consortium, rs4148941 acts as eQTL for CHST3 in lymphoblast cell lines ( P=5310251) and SPOCK2 in both adipose tissue (P=1310221) and lymphoblast cell line ( P=331024) (22). Given the additional trend toward association with GLP-1 RA treatment response in diabetic patients, further",
+    "IGFBP1, and IGFBP3. The IGF pathway is nowsuspected to play a role in diabetes because of observedassociations with IGF2BP2 (2729)."
+  ],
+  "answer": "Yes, the genes SH2B3, IFIH1, and ERBB3 are related to diabetes:  1. **IFIH1**: Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1 diabetes [1], [3], [5], [6].  2. **ERBB3**: ERBB3 is considered to be a major feature of type 1 diabetes susceptibility due to its role in immune signaling [4]. Additionally, a genome-wide association study proposed ERBB3 as the best candidate gene near a QTL for type 1 diabetes [8].  3. **SH2B3**: SH2B3 (also known as LNK) is considered to be a major feature of type 1 diabetes susceptibility due to its role in immune signaling [4].  Thus, all three genes, SH2B3, IFIH1, and ERBB3, are related to diabetes.",
+  "question": "Is any of the genes SH2B3, IFIH1 or ERBB3 related to diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_7 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_7
new file mode 100644
index 0000000..c7d4dcb
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_7
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "2011 - Prioritizing candidate disease genes by network-based boosting of genome-wide association data.pdf",
+    "2020 - Insights into pancreatic islet cell dysfunction from type 2 diabetes mellitus genetics..pdf",
+    "2007 - Integrative analysis for finding genes and networks involved in diabetes and other complex diseases.pdf",
+    "2010 - Genetics of Type 1 Diabetes What\u2019s Next.pdf",
+    "2015 - Biological interpretation of genome-wide association studies using predicted gene functions.pdf",
+    "2011 - Shared Genomics of Type 2 and Gestational Diabetes Mellitus.pdf",
+    "2010 - Common Inherited Variation in Mitochondrial Genes.pdf",
+    "1999 - Linkage of Type 2 Diabetes Mellitus and of Age at Onset to a Genetic Location.pdf",
+    "2019 - Genome-wide association study of type 2 diabetes in Africa.pdf"
+  ],
+  "extraction_id": [
+    "f7fe5916-4f25-5740-8737-f668f216575d",
+    "dffdea93-109e-5114-8795-e0fc66d6d3ed",
+    "f7013243-3e5f-509d-a414-edc4d7f27bc2",
+    "f13b4fee-14f4-5827-9482-3692165c8ce6",
+    "e5a38afd-cb9c-5552-9edd-3e9043d4f30d",
+    "0b09c4c7-a276-517f-a6e1-9388032fe622",
+    "29039cd9-9414-59e9-b97c-14f6f71ec4a2",
+    "8e91b32f-a873-5dc7-927d-52786cc44aa8",
+    "69b05acc-0a98-51de-a69c-1e46ca1c0ba3",
+    "ef39a6c5-9067-59e8-84ab-8b89071510d5"
+  ],
+  "document_id": [
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "db0aa4b3-66ec-5d51-be72-2a1289db944a",
+    "2a386c81-8f24-5993-8e48-0e89d7fb4fec",
+    "b91aeacf-6e83-52ac-beb6-034ad77cab18",
+    "261cbb40-ed6b-554c-a70d-db6b9f14cf74",
+    "8f9f62fd-9423-55b3-abf9-24cde0d2e775",
+    "bef0cabe-0bca-5715-9ffc-0b825744fbcf",
+    "9a5c8cba-06cb-5280-871f-1bbe128c3dc4",
+    "631b1f41-1064-5fc1-87f9-8a3c9f24ee9d",
+    "a7e4b6f4-fbb6-5dde-b638-d0d694c8ce87"
+  ],
+  "id": [
+    "chatcmpl-ADYn4g7NCIHEHW87vnQFVH1QRLe6y",
+    "e81d17bd-858c-52b7-8c02-2076e59afe20",
+    "18817608-0557-5acb-a091-9bc4d3640f7e",
+    "65941ce7-c762-5ae5-b1cd-4c62d8caddac",
+    "2e004b17-d266-50d9-be7f-33b523e59e54",
+    "375e0eba-87cf-5081-9f39-da1938e8be9e",
+    "b3455bcd-494e-5288-93ae-2fd761dd4157",
+    "51114ced-f323-57b9-87fb-30094a97642c",
+    "65daaa1d-b4e7-5d6c-aa4f-56b8a88bc1d7",
+    "ec145460-62ed-5375-b1a9-6231f94db4b9",
+    "e633c6eb-1fc6-5430-a324-f652c7f3e082"
+  ],
+  "contexts": [
+    "9.    Ehm MG, Karnoub MC, Sakul H, Gottschalk K,  Holt DC, Weber JL, American Diabetes Association GENNID Study Group. Genetics of NIDDM, et al. Genome wide search for type 2 diabetes susceptibil-ity genes in four American populations. Am J Hum Genet. 2000;66:187181.      10.    McCarthy M, Zeggini E. Genome-wide association  studies in type 2 diabetes. Curr Diab Rep. 2009;9:16471.      11.    Hivert MF, Jablonski KA, Perreault L, Saxena R,",
+    "that from orthologous genes of yeast, worm, and fly. The resulting HumanNet gene network can be accessed through a web interface (http://www.functionalnet.org/humannet). Using this interface, researchers can easily search the network using a set of seedTable 1. Selected top-ranked Crohns disease and type 2 diabetes genes for which network data added support to GWAS evidence, measured as an increase in odds (prior =1.7 for each) Crohns disease",
+    "twins. Diabetologia 30, 763768 (1987). 3. Neel, J. V. in The Genetics of Diabetes Mellitus    (eds W. Creutzfeldt, J. Kbberling, & J. V. Neel) 1-11 (Springer, 1976). 4. International HapMap Consortium, etal. A second generation human haplotype map of over 3.1 million  SNPs. Nature 449, 851861 (2007). 5. Sabeti, P . C. etal. Genome-wide detection and  characterization of positive selection in human  populations. Nature 449, 913918 (2007). 6. Genomes Project, C. etal. A global reference",
+    "Genome Biology  2007, 8:R253Open Access2007Bergholdtet al.Volume 8, Issue 11, Article R253Research Integrative analysis for finding genes and networks involved in  diabetes and other complex diseases Regine Bergholdt*, Zenia M Strling, Kasper Lage, E Olof Karlberg,  Pll  lason, Mogens Aalund, Jrn Nerup*, Sren Brunak,  Christopher T Workman and Flemming Pociot* Addresses: *Steno Diabetes Center, Niels Steensensvej 2, DK-2820 Gentofte, Denmark. Center for Biological Sequence Analysis, Technical",
+    "77. Bergholdt R, Brorsson C, Lage K, Nielsen JH, Brunak S, Pociot F. Expression proling of human genetic and protein interaction networks intype 1 diabetes. PLoS One 2009;4:e6250 78. Bergholdt R, Storling ZM, Lage K, Karlberg EO, Olason PI, Aalund M, Nerup J, Brunak S, Workman CT, Pociot F. Integrative analysis for ndinggenes and networks involved in diabetes and other complex diseases.Genome Biol 2007;8:R253 79. Oresic M, Simell S, Sysi-Aho M, Na nto -Salonen K, Seppa nen-Laakso T,",
+    "31. Saxena, R. et al. Genome-wide association analysis identies loci for type 2 diabetes and triglyceride levels. Science 316, 13311336 (2007). 32. Franke, L. et al. Reconstruction of a functional human gene network, with an application for prioritizing positional candidate genes. Am. J. Hum. Genet. 78, 10111025 (2006). 33. Su, Z., Marchini, J. & Donnelly, P. HAPGEN2: simulation of multiple disease SNPs. Bioinformatics 27,23042305 (2011).",
+    "Genetic exploration of GDM is in its initial stage. The genetics of GDM,  focusing on human association studies with candidate genes common to both T2DM and GDM is elegantly summarized by Robitaille and Grant (2008). The purpose of this chapter is to provide a comprehensive overview to include recent literature on susceptible gene variants that may contribute to both GDM and T2DM.   SEARCH STRATEGIES   A systematic literature search using PubMed was performed to identify stud-",
+    "Human Molecular Genetics 16(1): 3649, 2007). The DiabetesGenetics Initiative (DGI) study was used for the analysis, as we had access to genotype data in this study. The unadjusted gene p-value, P BestSNP g is the association p-value of the best regional SNP for gene g(y-axis in A). Phenotype permutation analysis was used as the gold standard to test goodness of gene score correction as it corrects forall confounders without requiring a priori knowledge of the confounders ( P Gene",
+    "version 2.0: users manual. PGL tech rep 2. Population Ge-netics Laboratory, Department of Genetics, Southwest Foun-dation for Biomedical Research, San Antonio Elbein SC (1997) The genetics of human noninsulin-dependent (type 2) diabetes mellitus. J Nutr 127:1891S1896S Elbein S, Hoffman M, Leppert M, Hasstedt S (1997) Linkage of fasting glucose in relatives of an NIDDM sib pair tomarkers on chromosome 9p. Diabetes 57 Suppl 1:51A Elston RC (1998) Methods of linkage analysisand the as-",
+    "Diabetes Study (DDS): a platform for chronic disease research.Glob Health Epidemiol Genom 1:e2. https://doi.org/10.1017/ gheg.2015.3 17. Genomes Project C, Auton A, Brooks LD et al (2015) A global reference for human genetic variation. Nature 526:68 74 18. Howie BN, Donnelly P, Marchini J (2009) A flexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet 5(6):e1000529. https://doi. org/10.1371/journal.pgen.1000529"
+  ],
+  "answer": "To use GeneNetwork to find genes related to diabetes in humans, you can follow these steps:  1. **Access the HumanNet Gene Network**: The HumanNet gene network can be accessed through its web interface at [http://www.functionalnet.org/humannet](http://www.functionalnet.org/humannet) [2].  2. **Search Using Seed Genes**: Once on the HumanNet interface, you can search the network using a set of seed genes. These seed genes can be genes already known to be associated with diabetes or genes you are interested in exploring further [2].  3. **Review Network Data**: The network data can provide additional support to genome-wide association studies (GWAS) evidence. For example, the network data can help measure the increase in odds for certain genes being associated with type 2 diabetes [2].  4. **Integrative Analysis**: You can perform integrative analysis to find genes and networks involved in diabetes. This involves combining data from various sources to identify potential genetic interactions and pathways related to diabetes [4].  By following these steps, you can effectively utilize GeneNetwork to identify and analyze genes related to diabetes in humans.",
+  "question": "How can I use genenetwork to find genes related with diabetes in humans?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_8 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_8
new file mode 100644
index 0000000..9b532e2
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_8
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2011 - Prioritizing candidate disease genes by network-based boosting of genome-wide association data.pdf",
+    "2007 - Integrative analysis for finding genes and networks involved in diabetes and other complex diseases.pdf",
+    "2017 - diabetes-mellitus-in-developing-countries-and-underserved-commun-2017.pdf",
+    "2010 - Genetics of Type 1 Diabetes What\u2019s Next.pdf",
+    "2015 - Biological interpretation of genome-wide association studies using predicted gene functions.pdf",
+    "2022 - A haplotype-resolved genome assembly of the Nile rat facilitates exploration of the genetic basis of diabetes.pdf",
+    "2011 - Prioritizing candidate disease genes by network-based boosting of genome-wide association data.pdf",
+    "2020 - Insights into pancreatic islet cell dysfunction from type 2 diabetes mellitus genetics..pdf",
+    "2007 - Integrative analysis for finding genes and networks involved in diabetes and other complex diseases.pdf",
+    "2009 - Gene prioritization based on biological plausibility over genome wide association studies renders new loci associated with type 2 diabetes.pdf"
+  ],
+  "extraction_id": [
+    "dffdea93-109e-5114-8795-e0fc66d6d3ed",
+    "f13b4fee-14f4-5827-9482-3692165c8ce6",
+    "f7fe5916-4f25-5740-8737-f668f216575d",
+    "e5a38afd-cb9c-5552-9edd-3e9043d4f30d",
+    "0b09c4c7-a276-517f-a6e1-9388032fe622",
+    "afa54304-6ffc-5f81-9431-d4c19f58527b",
+    "dcb6101e-cf09-5220-a3c9-ed5106c065b2",
+    "f7013243-3e5f-509d-a414-edc4d7f27bc2",
+    "f13b4fee-14f4-5827-9482-3692165c8ce6",
+    "a6b6c2df-f79b-58b8-a67a-fcf55b18d221"
+  ],
+  "document_id": [
+    "db0aa4b3-66ec-5d51-be72-2a1289db944a",
+    "b91aeacf-6e83-52ac-beb6-034ad77cab18",
+    "8a9451b9-d7e8-5417-b6a5-5fd1b791cc4d",
+    "261cbb40-ed6b-554c-a70d-db6b9f14cf74",
+    "8f9f62fd-9423-55b3-abf9-24cde0d2e775",
+    "6a8eb0a5-807d-5ef9-a732-b1dd722c0499",
+    "db0aa4b3-66ec-5d51-be72-2a1289db944a",
+    "2a386c81-8f24-5993-8e48-0e89d7fb4fec",
+    "b91aeacf-6e83-52ac-beb6-034ad77cab18",
+    "0fd2b5c8-9bda-5cc8-adb4-231d3842d50f"
+  ],
+  "id": [
+    "chatcmpl-ADYnAWVHUhI0y6oaOnDZY8VopPust",
+    "18817608-0557-5acb-a091-9bc4d3640f7e",
+    "2e004b17-d266-50d9-be7f-33b523e59e54",
+    "e81d17bd-858c-52b7-8c02-2076e59afe20",
+    "375e0eba-87cf-5081-9f39-da1938e8be9e",
+    "b3455bcd-494e-5288-93ae-2fd761dd4157",
+    "2360c49b-412e-5e9d-b95d-87a67b82e729",
+    "5e5e6bac-7695-5405-ad90-f24f5336fa34",
+    "65941ce7-c762-5ae5-b1cd-4c62d8caddac",
+    "4f009356-41ed-5cdc-9cfa-80cbb913874f",
+    "c21d117e-a223-5293-b794-6aa60729f7f7"
+  ],
+  "contexts": [
+    "that from orthologous genes of yeast, worm, and fly. The resulting HumanNet gene network can be accessed through a web interface (http://www.functionalnet.org/humannet). Using this interface, researchers can easily search the network using a set of seedTable 1. Selected top-ranked Crohns disease and type 2 diabetes genes for which network data added support to GWAS evidence, measured as an increase in odds (prior =1.7 for each) Crohns disease",
+    "Genome Biology  2007, 8:R253Open Access2007Bergholdtet al.Volume 8, Issue 11, Article R253Research Integrative analysis for finding genes and networks involved in  diabetes and other complex diseases Regine Bergholdt*, Zenia M Strling, Kasper Lage, E Olof Karlberg,  Pll  lason, Mogens Aalund, Jrn Nerup*, Sren Brunak,  Christopher T Workman and Flemming Pociot* Addresses: *Steno Diabetes Center, Niels Steensensvej 2, DK-2820 Gentofte, Denmark. Center for Biological Sequence Analysis, Technical",
+    "9.    Ehm MG, Karnoub MC, Sakul H, Gottschalk K,  Holt DC, Weber JL, American Diabetes Association GENNID Study Group. Genetics of NIDDM, et al. Genome wide search for type 2 diabetes susceptibil-ity genes in four American populations. Am J Hum Genet. 2000;66:187181.      10.    McCarthy M, Zeggini E. Genome-wide association  studies in type 2 diabetes. Curr Diab Rep. 2009;9:16471.      11.    Hivert MF, Jablonski KA, Perreault L, Saxena R,",
+    "77. Bergholdt R, Brorsson C, Lage K, Nielsen JH, Brunak S, Pociot F. Expression proling of human genetic and protein interaction networks intype 1 diabetes. PLoS One 2009;4:e6250 78. Bergholdt R, Storling ZM, Lage K, Karlberg EO, Olason PI, Aalund M, Nerup J, Brunak S, Workman CT, Pociot F. Integrative analysis for ndinggenes and networks involved in diabetes and other complex diseases.Genome Biol 2007;8:R253 79. Oresic M, Simell S, Sysi-Aho M, Na nto -Salonen K, Seppa nen-Laakso T,",
+    "31. Saxena, R. et al. Genome-wide association analysis identies loci for type 2 diabetes and triglyceride levels. Science 316, 13311336 (2007). 32. Franke, L. et al. Reconstruction of a functional human gene network, with an application for prioritizing positional candidate genes. Am. J. Hum. Genet. 78, 10111025 (2006). 33. Su, Z., Marchini, J. & Donnelly, P. HAPGEN2: simulation of multiple disease SNPs. Bioinformatics 27,23042305 (2011).",
+    "Page 16 of 21 Tohetal. BMC Biology           (2022) 20:245  Identification ofdiabeteslinked genes bytext mining We used four techniques to derive a set of genes associ - ated with type 2 diabetes and with diet-induced diabe - tes. First, we compiled an expert-curated gene-disease  association database from standard resources, the Com - parative Toxicogenomics Database [35] and PharmGKB  [36]. The result gave 277 genes associated with type 2  diabetes, but none associated with diet-induced dia -",
+    "2 diabetes alone and in combination with HumanNet and measuring performance as AUC ( <5% FPR) for recovering the top 20 genes from a type 2 diabetes meta-analysis of 4549 cases and 5579 controls (Zeggini et al. 2008). As for Crohns disease, consideration of the network boosts performance across a wide range of parameter values. Notably, consideration of the network strongly implicates the genes CTNNB1 and BACH2 in type 2 diabetes;",
+    "twins. Diabetologia 30, 763768 (1987). 3. Neel, J. V. in The Genetics of Diabetes Mellitus    (eds W. Creutzfeldt, J. Kbberling, & J. V. Neel) 1-11 (Springer, 1976). 4. International HapMap Consortium, etal. A second generation human haplotype map of over 3.1 million  SNPs. Nature 449, 851861 (2007). 5. Sabeti, P . C. etal. Genome-wide detection and  characterization of positive selection in human  populations. Nature 449, 913918 (2007). 6. Genomes Project, C. etal. A global reference",
+    "type 1 diabetes genome scan data, and a high -confidence human protein interaction network. Resulting networks were ranked by the significance  of the enrichment of proteins from interacting regions. We identified a number of new prot ein network modules and novel candidate genes/ proteins for type 1 diabetes. We propose this type of integrative analysis as a general method for the elucidation of genes and networks involv ed in diabetes and other complex diseases. Background",
+    "gene prioritization are explained in detail in the Appendix, Supplemental Digital Content 1 , http://links.lww.com/A1049. In addition, the complete list of the training genes, including both the Gene HGNC symbol, and gene name are shown in the Appendix , Supplemental Digital Content 1 , http://links.lww.com/A1049. Moreover, from the freely available site http:// www.broad.mit.edu/ diabetes/, we downloaded the results of the GWA study in 3000 Scandinavian individuals about the genetic variants that inu-"
+  ],
+  "answer": "To use the GeneNetwork tool to find genes related to diabetes in humans, you can follow these steps:  1. **Access the HumanNet Gene Network**: The HumanNet gene network can be accessed through its web interface at [http://www.functionalnet.org/humannet](http://www.functionalnet.org/humannet) [1].  2. **Search the Network**: Once you are on the HumanNet web interface, you can search the network using a set of seed genes. This allows you to explore the network for genes that are functionally related to your genes of interest, such as those involved in diabetes [1].  3. **Integrative Analysis**: You can perform integrative analysis to find genes and networks involved in diabetes. This involves combining data from various sources, such as genome-wide association studies (GWAS) and protein interaction networks, to identify candidate genes and protein network modules associated with diabetes [2], [4], [9].  4. **Review Network Data**: Consider network data that supports GWAS evidence. For example, network data has been used to add support to GWAS evidence for type 2 diabetes genes, measured as an increase in odds [1].  By following these steps, you can effectively use the GeneNetwork tool to identify and analyze genes related to diabetes in humans.",
+  "question": "How can I use the GeneNetwork tool to find genes related with diabetes in humans?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_9 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_9
new file mode 100644
index 0000000..2474ccd
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_diabetes_9
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2015 - Association between circulating 25-hydroxyvitamin D.pdf",
+    "2008 - Inflammation, Insulin Resistance and Diabetes.pdf",
+    "2019 - Adiponectin Levels and Risk of Type 2 Diabetes.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "2001 - A Gene Conferring Susceptibility to Type 2 Diabetes.pdf",
+    "2004 - Diabetes Genes a.pdf",
+    "2010 - The Genetics of Type 2 Diabetes.pdf",
+    "2005  - Animal models of diabetes mellitus.pdf",
+    "2008 - Inflammation, Insulin Resistance and Diabetes.pdf",
+    "2006 - Analysis of 14 Candidate Genes for Diabetic Nephropathy.pdf"
+  ],
+  "extraction_id": [
+    "bea401d8-c665-5910-b3f6-0974617dc32f",
+    "5f281214-ff2f-5f01-ad99-9faa2edf5751",
+    "2b483dbc-bc68-5065-b650-811216009db5",
+    "0b7b0898-2ca5-59aa-aa53-1ab187fe3dd0",
+    "57ab3edd-7ed7-568a-9e47-b351e2cbf987",
+    "1a24f9ed-c710-589a-a7f6-7ec46b8cf6a8",
+    "609eb8ef-0827-567c-98ed-51b1945e5c9d",
+    "3e91ce60-4e0b-527d-b8c7-147d5717e17a",
+    "ea677f19-62e2-5f59-b995-babba0066f5f",
+    "dd18cc20-78a9-5589-b605-d5bf37628258"
+  ],
+  "document_id": [
+    "3affa62b-11b1-5068-b699-1799d5566ea3",
+    "71615f74-43a6-5f1c-9300-db97f10b212c",
+    "584ea813-0169-5c1e-90f2-4abfbcc0b4fa",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "8c07c8b6-4192-56cf-9332-715ff7085577",
+    "805ad0d5-6a3d-5e5f-b1ba-121e10c6daaa",
+    "a1d211d4-279e-51d7-b2b2-33bc2763d089",
+    "2fd381ac-2898-5a8c-af93-bcc86e7dec14",
+    "71615f74-43a6-5f1c-9300-db97f10b212c",
+    "6512e711-d879-50b3-96e7-2e97339ebad5"
+  ],
+  "id": [
+    "chatcmpl-ADYnFB30HdQXCs4mqZmx2zUlQo7uq",
+    "ad3e5060-4495-5a6f-9bb3-0f372359d852",
+    "c1c877eb-2dfe-586b-bde8-99b2c0ae8515",
+    "50aa0f0c-1095-5660-bde3-f6320c88506d",
+    "0dd93045-9b04-53eb-9d51-1dcc1c9c975a",
+    "a9ade0cc-6bd0-57d7-ab72-4a0b6735e455",
+    "8e3bbcea-6f25-57c4-86b8-845b1cd0fc02",
+    "ee0041e3-b9a3-5a97-9761-55208d06bb2d",
+    "e35ff76b-92a1-51fa-b28f-d9c90a81f2fd",
+    "82404153-20f8-53e5-92a9-5cf0818d3c4d",
+    "b9e891e3-eba8-5a00-86ce-55d5d144361d"
+  ],
+  "contexts": [
+    "confounding, which is plausible in observational studies of incident type 2 diabetes. Measurements of confounders (eg, physical activity) are susceptible to errors and are not adequately controlled for in epidemiological analyses. 5  Although results from clinical trials6,7 have shown no e ect of vitamin D supplementation on the incidence of  type 2 diabetes, these   ndings require cautious  interpretation because of issues with doses, combination treatment with calcium, compliance, and generalisability. 3",
+    "common (confounding factors) that are the real causes of diabetes. In this study, the researchers use Mendelian randomization to examine whether increased blood CRP causes diabetes. Some variants of CRP (the gene that encodes CRP) increase the amount of CRP in the blood. Because these variants are inherited randomly, there is no likelihood ofconfounding factors, and an association between these variants and the development of insulin resistance and diabetes indicates, therefore, that",
+    "residual confounding. As shown inTable 2, many of the included studiesadjusted for a wide range of potentialconfounders, including demographicand lifestyle factors. The strength of theadjusted RRs for adiponectin levels anddiabetes risk and the consistency of as-sociations across diverse populations re-duce the likelihood that residual con-founding by these variables can explainthe findings. Another issue is whetheradiponectin has a causal effect on dia-betes or is only a surrogate marker forother",
+    "diabetes are related to impaired glucose counterregulation and  hypoglycemia unawareness, one should also keep in mind that  hypoglycemia can be multifactorial and be the result of several unrelated  diseases. These include liver disease, malnutrition, sepsis, burns, total  parenteral nutrition, malignancy and administration of certain medications  known to reduce plasma glucose concentrations (Table 1).27  In principle, the same risk factors for hypoglycemia apply to",
+    "exists in the overall sample. In the case of type 2 diabetes,one would ideally stratify on the basis of insulin resistanceand/or severity of insulin secretion defect. However, con-founding environmental effects, including varying durationof disease, differing access to health care, heterogeneity inprescription, and variation in adherence to treatmentregimes, make inferences about insulin action in diabeticpatients problematic, especially inferences based solely onoral glucose tolerance test (OGTT) data",
+    "of diabetes remains one of the great challenges in human genetics.  Diabetes is a result of complex interactions between genetic and  non-genetic (including environmental) factors. Although diabetes and its  related traits have been shown to cluster within families, their .transmission  does not follow a Mendelian fashion, except for some rare syndromes such  as MODY. Diabetes could be the result of few common variants with a  relatively large effect, such as HLA alleles at the MHC locus and VNTR",
+    "predisposing to diabetes through effects on insulin sensitivity, however, may be more dif  cult to track down because of strong",
+    "is still unclear. Genetic studies in both animalsand humans are complex, given the many susceptibility andprotective loci that contribute to the overall risk of diabetes",
+    "adjustment for adiposity and other confounding factors [4 10]. Preventing or delaying onset of diabetes and its compli- cations is an important therapeutic aim, and there is interest in inammatory effectors including CRP as drug targets [11,12]. It is therefore highly desirable to establish which mediators in the inammatory cascade are causal for diabetes. Mendelian randomization involves comparison of pheno- type and genotype effects in observational studies [13]. If the",
+    "adjusting for sex, diabetes duration, HbA1c, and smoking, assuming either additive or dominant effects of the polymorphisms.N. VIONNET AND ASSOCIATES DIABETES, VOL. 55, NOVEMBER 2006 3169Downloaded from http://diabetesjournals.org/diabetes/article-pdf/55/11/3166/649912/zdb01106003166.pdf by Kenya Institution user on 14 July 2023"
+  ],
+  "answer": "Confounding factors in diabetes are variables that can influence both the exposure and the outcome, potentially leading to misleading associations. Some of the confounding factors in diabetes include:  1. **Physical Activity**: Errors in measuring physical activity can lead to inadequate control of this confounder in epidemiological analyses [1]. 2. **Demographic and Lifestyle Factors**: These include a wide range of variables such as age, sex, socioeconomic status, and lifestyle habits like diet and exercise, which can influence diabetes risk [3]. 3. **Environmental Effects**: These include varying duration of disease, differing access to healthcare, heterogeneity in prescription, and variation in adherence to treatment regimes [5]. 4. **Adiposity**: Adjusting for adiposity and other confounding factors is crucial in studies examining the onset and complications of diabetes [9].  These factors need to be carefully controlled for in studies to ensure accurate and reliable results.",
+  "question": "what are confounding factors in diabetes?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_1 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_1
new file mode 100644
index 0000000..886b472
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_1
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2021 - Human Molecular Genetics and Genomics.pdf",
+    "2015 - A new age in functional genomics using CRISPR Cas9 in arrayed library screening.pdf",
+    "2016 - Genome editing comes of age.pdf",
+    "2018 - Neuro-Immuno-Gene- and GenomeEditing-Therapy for Alzheimer\u2019s.pdf",
+    "2016 - Genome editing comes of age.pdf",
+    "2016 - Dissecting diabetes metabolic disease.pdf",
+    "2020 - Functional Genomics in Pancreatic \u03b2 Cells Recent Advances in Gene Deletion and Genome Editing Technologies for Diabetes Research.pdf",
+    "2021 - Human Molecular Genetics and Genomics.pdf",
+    "2018 - Neuro-Immuno-Gene- and GenomeEditing-Therapy for Alzheimer\u2019s.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf"
+  ],
+  "extraction_id": [
+    "08a2c0e6-8ca8-5a72-974c-3f1e27ba1b15",
+    "49b81415-ef6f-5cc4-bb30-71e971070ebe",
+    "190e8838-4f61-5431-8848-98564ded7140",
+    "66dbf4f0-2b37-5219-9eeb-0a560df8d888",
+    "c7b143d7-347c-5160-bfd4-82283b342d7d",
+    "ebabc771-1777-56c1-9101-c1677c5ae908",
+    "fe5bf2df-2eda-5ef0-8aad-79bbc5b898d6",
+    "08a2c0e6-8ca8-5a72-974c-3f1e27ba1b15",
+    "3dd04f3c-0dc4-5bf7-aff6-3d9282761a2e",
+    "54972d7f-0ddc-5076-9d58-890a85f71332"
+  ],
+  "document_id": [
+    "68e362a5-e580-5a4d-8d41-6a138c873ede",
+    "20df9469-e1cc-500e-ac30-fbba981d7aee",
+    "4078087a-c2a4-5c58-95b5-4ae243794800",
+    "cc0a025b-71e7-5712-bbf7-4ee1e0f769ef",
+    "4078087a-c2a4-5c58-95b5-4ae243794800",
+    "eee2f79d-e093-52fb-871a-798fd859235e",
+    "51350055-d53c-5692-ab53-337b8a8bafd6",
+    "68e362a5-e580-5a4d-8d41-6a138c873ede",
+    "cc0a025b-71e7-5712-bbf7-4ee1e0f769ef",
+    "62b635c3-040e-512a-b016-6ef295308a1e"
+  ],
+  "id": [
+    "chatcmpl-ADZ9J7DmFDWNhxuisZ7zPEZmBG259",
+    "ffbca864-26db-5f36-8ad4-3b8d24d46de6",
+    "2fbabecd-22c1-5570-8f38-bc934d463710",
+    "6dfc48be-a762-55d6-9aba-799d80e8140d",
+    "a4aa9de7-cc9f-5c3f-a9fe-c37a47faa5b7",
+    "ff2d183b-c5be-5e05-94c8-e2db379dcd96",
+    "a3d6f231-29aa-5cf6-b856-004d3d9dd9c1",
+    "6ba3cf43-be4d-561f-ad84-f79921cab37e",
+    "4474c4e9-bc07-5610-8bb2-dafe5c95774b",
+    "77b11b36-35c8-55dc-a2d1-25d5e4ca218d",
+    "176b7aa5-17ef-590d-8807-1aa7def904bb"
+  ],
+  "contexts": [
+    "neered nucleases, CRISPR-Cas9  tools have accelerated the pace of  genomic research by permitting  highly efficient knockouts or  edits of virtually any gene in cells  or model organisms. Multiple CRISPR-Cas9based clinical trials  are in progress or are expected  to begin soon. Although Cas9- engineered cells havent yet dem - onstrated efficacy at scale, early  trial results suggest that such  cells are stable and dont cause  acute adverse reactions in humans.  Long-term safety is yet to be de -",
+    "stageissetforCRISPRtomakeanenormousimpactongenomic screening and thus scientic discovery in the coming years, and recent demonstrations of this system have shown great promise (Shalem etal., 2015 ).However,a number of technical challenges must be addressed in order to maximize the benet of this technology. In this review, we will discuss current applications of CRISPR in functional genomics and provide a perspective on futuredevelopmentsinthisarea. CRISPR/Cas9 Genome Editing",
+    "heralding the age of genome editing. Furthermore, Cas9 or guide RNAs have been linked to various effector proteins to enable targeted gene regulation 12,13 and epigenome modifications14,15.  It is worth noting, however, that many of these feats had been demonstrated previously using other nucleases or DNA-binding proteins 1,16. In this Perspective, I shed light on early genome  editing platforms that laid the groundwork for the widespread use of CRISPRCas9 in research and medicine (Fig. 1 ).",
+    "CRISPR/CAS9 HOLDS SIGNIFICANT PROMISE FOR THE DEVELOPMENT OFNEW AD MODELS AND PRECISIONTARGETED AD THERAPY Clustered regularly interspaced short palindromic repeat (CRISPR)-Cas nucleases have revolutionizedthe eld of gene editing and have tremendous appli-cation in the eld of molecular medicine [98102].Despite a signicant surge in CRISPR/Cas9-mediated genome editing in various disease models,the progress in the eld of AD has lagged behindsubstantially. We believe that genome editing can sig-",
+    "81. Applications for CRISPRCas9 beyond genome editing",
+    "cline- or Tet-regulated Cas9 system. Current CRISPR/Cas systems arefrom Streptococcus pyogenes ,Streptococcus thermophilus ,Neisseria meningitides and Treponema denticola .2.5. Caveats of advanced genome editing tools Off-target effects . The DNA-binding domains of ZFNs and TALENs need to be very speci c for the target site to avoid off-target cleavage, which results in unwanted mutations and potentially cytotoxic effects [27]. CRISPR/Cas9 is also known to generate off-target alterations,",
+    "on transcriptional interfere nce (CRISPRi) and activation (CRISPRa) have also harnessed Cas9-based technologies for use in genome-wide studies ( 59,174). In addition, recent improvements in lentiviral library generation and propagation,as well as large-scale DNA and RNA synthesis, have allowedCRISPR-Cas9 technology to be exploited across multiple modelplatforms ( 59,175178). nCas9 The CRISPR-Cas9 system can tolerate certain mismatches to the DNA target since the required gRNAs are short. A disadvantage,",
+    "CRISPR-Cas9 can be used to in - duce genome edits by creating  targeted DNA breaks that trigger  site-specific DNA repair. In next- generation formats, it can also  control the transcriptional output  of genes or alter genome se - quences using a process of nu - cleotide base editing  that does not require  repair of DNA breaks.  As these technolo - gies continue to mature, it will  become increasingly possible to  alter cellular genomes efficiently  and accurately. Coming on the heels of engi -",
+    "S.P . Raikwar et al. / Alzheimers Disease: New Therapeutic Horizons 333 gene editing efciency of the CRISPR/Cas9 systems.",
+    "13. Kleinstiver BP, etal. High-fidelity CRISPRCas9 nucleases with no detectable genome-wide  off-target effects. Nature. 2016;529:4905.  14. Brane A, Tollefsbol T.Targeting telomeres and telomerase: studies in aging and disease uti- lizing CRISPR/Cas9 technology. Cells. 2019;8:186.  15. Wang H, etal. One-step generation of mice carrying mutations in multiple genes by CRISPR/ Cas-mediated genome engineering. Cell. 2013;153:9108."
+  ],
+  "answer": "### Potential Benefits of CRISPR-Cas9:  1. **Accelerated Genomic Research**: CRISPR-Cas9 tools have significantly accelerated the pace of genomic research by allowing highly efficient knockouts or edits of virtually any gene in cells or model organisms [1]. 2. **Clinical Trials and Stability**: Multiple CRISPR-Cas9-based clinical trials are in progress, and early trial results suggest that CRISPR-engineered cells are stable and do not cause acute adverse reactions in humans [1]. 3. **Functional Genomics**: CRISPR is set to make a significant impact on genomic screening and scientific discovery, with recent demonstrations showing great promise [2]. 4. **Targeted Gene Regulation and Epigenome Modifications**: CRISPR-Cas9 has been linked to various effector proteins to enable targeted gene regulation and epigenome modifications [3]. 5. **Precision Medicine**: CRISPR-Cas9 holds significant promise for the development of new disease models and precision-targeted therapies, particularly in molecular medicine [4]. 6. **Versatility in Applications**: CRISPR-Cas9 can induce genome edits by creating targeted DNA breaks that trigger site-specific DNA repair. It can also control the transcriptional output of genes or alter genome sequences using nucleotide base editing, which does not require DNA break repair [8].  ### Potential Risks of CRISPR-Cas9:  1. **Off-Target Effects**: CRISPR-Cas9 is known to generate off-target alterations, which can result in unwanted mutations and potentially cytotoxic effects [6]. 2. **Technical Challenges**: There are several technical challenges that need to be addressed to maximize the benefits of CRISPR technology [2]. 3. **Long-Term Safety**: While early trial results are promising, the long-term safety of CRISPR-engineered cells is yet to be determined [1]. 4. **Mismatch Tolerance**: The CRISPR-Cas9 system can tolerate certain mismatches to the DNA target, which could be a disadvantage as it may lead to unintended edits [7].  In summary, while CRISPR-Cas9 offers numerous benefits such as accelerating genomic research, enabling precision medicine, and providing versatile applications, it also poses risks like off-target effects, technical challenges, and concerns about long-term safety.",
+  "question": "What are the potential benefits and risk associated with gene editing technologies like CRISPRR-Cas9?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_10 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_10
new file mode 100644
index 0000000..2f7afe0
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_10
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2007 - Classification of microarray data using gene networks.pdf",
+    "2020 - Gene network a continuously updated tool for systems genetics analyses.pdf",
+    "2020 - Gene network a completely updated tool for systems genetics analyses.pdf",
+    "2013 - Integrated Enrichment Analysis of Variants.pdf",
+    "2013 - Candidate gene association studies a comprehensive guide to useful in silicotools.pdf",
+    "2017 - Integrative functional genomics for systems genetics in GeneWeaver. org.pdf",
+    "2023 - Genome-wide RNA polymerase stalling.pdf",
+    "2011 - The age of the \u201come\u201d Genome, transcriptome and proteome data set collection and analysis.pdf",
+    "2020 - Gene network a continuously updated tool for systems genetics analyses.pdf",
+    "2020 - Gene network a completely updated tool for systems genetics analyses.pdf"
+  ],
+  "extraction_id": [
+    "b509b652-ff49-53b8-9fe1-6b2340c166a6",
+    "0ddd5599-537b-581d-9775-b4ec0662cfae",
+    "b1a51c38-5376-51ec-9d6b-a02b63164eb5",
+    "39015cf6-2e14-5ef7-a5af-b1a87ef22594",
+    "e4aaaddf-ed9f-5663-b8b7-403b02631793",
+    "800a4df7-3d75-50cf-bb6c-aef53b97af0f",
+    "801887dc-6c57-5d4d-8ba3-8a7a84707a8e",
+    "87e61158-ff52-5bbc-926d-47cd018529aa",
+    "f9dee762-add3-56b8-baa6-f260e05af531",
+    "58f46b5e-7cfe-5926-ae36-d0a6d7741171"
+  ],
+  "document_id": [
+    "639e0456-a445-5e2e-adf5-8eaf987ce2d1",
+    "374fd6d3-e6c1-560c-a421-a4b393ba23b2",
+    "128224f1-3545-52c3-93cb-77c3cf4ec70a",
+    "e4b37f87-e940-563c-851c-b272fc30e394",
+    "72134204-0751-5b57-a051-a0ea2d320fa1",
+    "cbe10d1f-5271-5c0e-94e3-1479b7e39146",
+    "78812a12-8d31-5159-8367-b0d38e5bc84b",
+    "ca99ed69-ee09-5717-95ed-c26eefb5e42d",
+    "374fd6d3-e6c1-560c-a421-a4b393ba23b2",
+    "128224f1-3545-52c3-93cb-77c3cf4ec70a"
+  ],
+  "id": [
+    "chatcmpl-ADZAEvRN04c44oiql0rqsqslEqSpD",
+    "8f5c5693-f995-5ded-8498-701ff0889deb",
+    "6ab69fa3-2dcf-50dc-bd36-283407a39451",
+    "02b60e7c-25ee-5583-822d-a0a4799f4eeb",
+    "7dbba72c-bf76-5431-aa01-9c828355bed8",
+    "3539d21a-cc75-54dc-aca3-2d936893481b",
+    "0c72f387-9074-592d-a87e-7643c2f37d0c",
+    "5014c31a-1e5c-5101-9c4c-9b6b40c65435",
+    "e17e2cc4-ca55-55e8-9461-b692c3c5bf00",
+    "e9748c2d-a9f6-596e-bba2-97bf34ed86d6",
+    "96f49474-9477-5ac6-8606-81296848493a"
+  ],
+  "contexts": [
+    "[3] and KEGG [4] all allow a list of genes to be crossed with biological functions and genetic networks, including metabolic, signalling or other regulation pathways. Basic statistical analysis (e.g., [5,6]) can then determine whether a pathway is over-represented in the list, and whether it is over-activated or under-activated. However, one can argue that introducing information on the path- way at this point in the analysis process sacrifices some statistical power to the simplicity of the approach. For",
+    "Sidiropoulos, K., Viteri, G., Sevilla, C., Jupe, S., Webber, M., Orlic -Milacic, M., et al. (2017).  Reactome enhanced pathway visualization. Bioinformatics  33, 3461 3467.  doi:10.1093/bioinformatics/btx441.   Slenter, D. N., Kutmon, M., Hanspers, K., Riutta, A., Windsor, J., Nunes, N., et al. (2018).  WikiPathways: a multifaceted pathway database bri dging metabolomics to other omics  research. Nucleic Acids Res.  46, D661 D667. doi:10.1093/nar/gkx1064.",
+    "Sidiropoulos, K., Viteri, G., Sevilla, C., Jupe, S., Webber, M., Orlic -Milacic, M., et al. (2017).  Reactome enhanced pathway visualization. Bioinformatics  33, 3461 3467.  doi:10.1093/bioinformatics/btx441.   Slenter, D. N., Kutmon, M., Hanspers, K., Riutta, A., Windsor, J., Nunes, N., et al. (2018).  WikiPathways: a multifaceted pathway database bri dging metabolomics to other omics  research. Nucleic Acids Res.  46, D661 D667. doi:10.1093/nar/gkx1064.",
+    "analysis, we restrict the analysis to curated, peer-reviewedpathways based on experimental evidence, and pathways inferred via gene homology. We draw candidate pathways from the collections listed in Figure 6 (see also Supplementary Materials). KEGG [146] and HumanCyc [147] are primarily databases of metabolic pathways, and are unlikely to be relevant to someJoint Analysis of Variants and Pathways in Disease PLOS Genetics | www.plosgenetics.org 11 October 2013 | Volume 9 | Issue 10 | e1003770",
+    "textual interface, also linking out to the original articles. Analysing participating pathways is an important aspect of any gene s functional analysis strategy. In this view, REACTOME (http://www.reactome.org) [13] is a cross referenced, manually curated and peer reviewed pathway database. LitInspector (http://www.litinspector.org) [14]and NetPath (http://www.netpath.org/index.html) [15] allow one to access curated signal transduction related lit-",
+    "I, Babur O, Anwar N, Schultz N, Bader GD, Sander C (2011) Pathway Commons, a web resource for biological pathway data. Nucleic Acids Res 39(Database issue):D685D690. doi:   10.1093/nar/gkq1039           6.    Baker EJ, Jay JJ, Bubier JA, Langston MA,  Chesler EJ (2012) GeneWeaver: a web-based system for integrative functional genomics. Nucleic Acids Res 40(Database issue):D1067D1076. doi:   10.1093/nar/gkr968            7.    Bubier JA, Phillips CA, Langston MA, Baker",
+    "67. Krmer, A., Green, J., Pollard, J. Jr. & Tugendreich, S. Causal analysis approaches in ingenuity pathway analysis. Bioinformatics   30, 523530 (2014). 68. Jassal, B. et al. The reactome pathway knowledgebase. Nucleic Acids Res. 48, D498D503 (2020). 69. Okonechnikov, K., Conesa, A. & Garca-Alcalde, F. Qualimap 2: advanced multi-sample quality control for high-throughput sequencing data. Bioinformatics 32, 292294 (2016).",
+    "Biocarta    pathway  maps  www.biocarta.com BioGRID    genetic  and  protein  interaction  database  thebiogrid.org AnalysisPLINK  processing  and  QC  of  genetic  data  sets  pngu.mgh.harvard.edu/ purcell/plink Bioconductor    processing  and  QC  of  expression  data  sets  www.bioconductor.org DAVID   gene  ontology,  pathway  analysis  david.abcc.ncifcrf.gov WebGestalt    gene  ontology,  pathway  analysis  bioinfo.vanderbilt.edu/webgestalt Sage",
+    "2004; Gene Ontology Consortium, 2015; The Gene Ontology Consortium, 2019) , KEGG  pathways  (Kanehisa and Goto, 2000; Kanehisa et al., 2012) , Panther pathways  (Mi et al., 2019a,  2019b) , Reactome pathways  (Sidiropoulos et al., 2017; Jassal et al., 2020) , and Wikipathway  pathways  (Pico et al., 2008; Slenter et al., 2018)  (Figure 31). As many different annotations as  wanted can be chosen by clicking on the + icon ( Figure 31). Also note, that the user can",
+    "2004; Gene Ontology Consortium, 2015; The Gene Ontology Consortium, 2019) , KEGG  pathways  (Kanehisa and Goto, 2000; Kanehisa et al., 2012) , Panther pathways  (Mi et al., 2019a,  2019b) , Reactome pathways  (Sidiropoulos et al., 2017; Jassal et al., 2020) , and Wikipathway  pathways  (Pico et al., 2008; Slenter et al., 2018)  (Figure 31). As many different annotations as  wanted can be chosen by clicking on the + icon ( Figure 31). Also note, that the user can"
+  ],
+  "answer": "For pathway analyses, you can use the following resources:  1. **KEGG (Kyoto Encyclopedia of Genes and Genomes)**: This database allows you to cross-reference a list of genes with biological functions and genetic networks, including metabolic, signaling, or other regulation pathways [1], [4], [9], [10].  2. **Reactome**: A cross-referenced, manually curated, and peer-reviewed pathway database that provides enhanced pathway visualization [2], [3], [5], [7], [9], [10].  3. **WikiPathways**: A multifaceted pathway database that bridges metabolomics to other omics research [2], [3], [9], [10].  4. **HumanCyc**: Primarily a database of metabolic pathways [4].  5. **Pathway Commons**: A web resource for biological pathway data [6].  6. **BioCarta**: Provides pathway maps [8].  7. **BioGRID**: A genetic and protein interaction database [8].  8. **DAVID (Database for Annotation, Visualization, and Integrated Discovery)**: Offers gene ontology and pathway analysis [8].  9. **WebGestalt**: Provides gene ontology and pathway analysis [8].  These resources can help you determine whether a pathway is over-represented in your gene list and whether it is over-activated or under-activated [1].",
+  "question": "What resources can I use to do pathway analyses?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_11 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_11
new file mode 100644
index 0000000..8dcfd54
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_11
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - Opening Up the Conversation on Genetics.pdf",
+    "2008 -  Study Design and Statistical Issues.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2019 - Sexual Dimorphism in the Age of Genomics How, When, Where.pdf",
+    "2008 - Genotype-phenotype relationships and the patterning of complex traits as exemplified in the mammalian dentition.pdf",
+    "2007 - Promoting_Student_Scientific_Literacy_of_Molecular Genetics and Genomics.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2019 - The influence of paternal diet on sncRNA-mediated epigenetic.pdf"
+  ],
+  "extraction_id": [
+    "51dbd5e2-fde6-5097-aa05-fcf57d3ca6b1",
+    "06bf0605-388a-592c-96ad-3a53bb36362c",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "8a1ce8fa-b5f4-5942-b7b1-14a8a7887710",
+    "e22bb6fb-bec4-5c4c-8690-c96d0b8d13d4",
+    "5aab3e60-b8b0-52ad-b4d3-817cf012cfa5",
+    "67369433-749b-5d6a-b5ef-3f0afe78b767",
+    "206b8810-b7c1-5195-a10f-4e291864b77c",
+    "84335575-34d7-56b6-aa06-5a8ac13d637a"
+  ],
+  "document_id": [
+    "b62a8f54-c2f5-5bbb-9324-af80f7537167",
+    "c3bd9cf0-f768-55c4-be94-96590d7acc21",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "3f8c03b0-4235-5774-9d26-e43d55c1001b",
+    "f6e866b8-b233-5862-bfb8-9949d0dabb97",
+    "755f34c4-cc06-5275-a744-16d48162b012",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "dfcbd6e6-f60d-5eb7-867b-34ec78415e82"
+  ],
+  "id": [
+    "chatcmpl-ADZAJ4XRzNSAEiekxHtxfyNvHLw8G",
+    "60ad1512-b0c0-59cd-ace4-c146e2c04b52",
+    "1e151ad5-59d9-598d-97ba-90ba0e64c4cb",
+    "a66b8b00-d51c-575b-b6ac-fa445c4ca715",
+    "df4c6108-740d-5bcf-99e6-dbda74f7e41a",
+    "4472740a-d22d-5bb1-98e3-e91332cbb303",
+    "0158f264-120f-5942-ad55-ef5fde1f188a",
+    "47b9142f-98a3-5a45-8eaa-d327c9cc055d",
+    "8e3fdc2c-0962-5854-83e7-a60ab05cf6de",
+    "6c8dfaa1-a96f-5f1c-8b5a-870acfd46f5f",
+    "be93ee68-72ae-5015-a3f0-19e7bf24827a"
+  ],
+  "contexts": [
+    "the egg and the sperm. Such a process would result in genetic changes that will be copied into every cell of the future adult, including reproductive cells (Stock & Campbell, 2000), opening the door to irreversibly alter the human species. Inevitably, signifi  cant self-disclosure and discussion challenges await families",
+    "phenomena such as mutations and gene conversion events) occur in relevant meioses  leading up to the formation of the gametes (i.e., egg and sperm) which are combined  during fertilization and the formation of zygotes. Thus, individuals inherit a patch- work of chromosomal segments from maternal and paternal chromosomes.",
+    "a fertilized egg is a complicated process that relies on controlling: which genes are active; whenthese genes activate; and for how long they are active. In broad terms, there are four ways that thiscontrol can be achieved: First, inside the sperm or egg, genes can be marked with small chemical tags that flag these genes",
+    "to be activated (or remain inactive) after fertilization, depending on whether the modification wasmade by the father (in the sperm) or the mother (in the egg); this process is known as imprinting. Second, the mother can alter the gene activity in her offspring via the placenta; this process is known as maternal effect. Third, instructions encoded within the embryos DNA can directly control if, andwhen, a nearby gene becomes activated; this is known as  cis-regulation. Finally, similar instructions",
+    "(Figures 8 and 9). Two gametes (egg and sperm) ultimately  join into a single cell, the zygote, which has the full comple-ment of 23 chromosome pairs restored. If all goes well, the zygote gives rise to a live offspring. The Mendel Laws: Segregation and Independent  Assortment Both of the Mendel laws pertain directly to the process of  meiosis. The first Mendel law, the law of segregation, states  that each parent passes a randomly selected allele for a given",
+    "sex chromosome effects. (B)Soon after fertilization, male and female cells have sex-specic transcriptomes, epigenomes, and phenotypes (for example, male embryos grow faster than female embryos). At implantation, lineage determination begins and gene expression differences are reduced. Epigenetic marks, however, are less constrained and some are maintained, affecting gene expression, and phenotype later in development. Once specic lineages are established, differences in",
+    "the subset of that genetic information that is active.  But how does the differentiation process  begin?  The key insight in resolving this conundrum came from fly genetics and was the  realization that the egg is not a homogenous sack of protoplasm.  The maternally-derived genes  active in the fertilized egg are asymmetrically distributed such that at the first cell division each  daughter cell receives a different complement of factors.  Development continues as a",
+    "genes.  An altered gene may be passed on to every cell that develops from it.  The resulting features my help, harm, or have little or no effect on the offsprings success in its environment. (AAAS, pg. 109, 5B:9-12#4 ) 6. Heritable material: The information passed from parents to offspring is coded in DNA molecules (AAAS, pg 108, 5B:9-12#3) 7. Mutagens: Gene mutations can be caused by such things as radiation and chemicals.  When they occur in sex cells, the mutations can be passed onto offspring; if they",
+    "or father (sperm cell). Each gamete has a set of 23 unpaired chromosomes. Two human gametes (egg and sperm) combine  to create a cell (zygote) that contains the full human genome  of 23 paired chromosomes.Genetic Information Nondiscrimination Act (GINA) US federal legislation that makes it unlawful to discriminate  against individuals on the basis of their genetic profiles in regard to health insurance and employment. These protections  are intended to encourage Americans to take advantage of",
+    "spermatozoa: more than the sum of its parts? DNA, histones, pro - tamines and epigenetics. Reproduction 139:287301 Nilsson EE, Sadler-Riggleman I, Skinner MK (2018) Environmentally  induced epigenetic transgenerational inheritance of disease. Envi-ron Epigenet 4:dvy016Pembrey M, Saffery R, Bygren LO, Network in Epigenetic Epide-"
+  ],
+  "answer": "Once a sperm combines with an egg, the traits passed onto the resulting lifeform are determined by several genetic and epigenetic factors:  1. **Chromosomal Segments**: Individuals inherit a patchwork of chromosomal segments from maternal and paternal chromosomes, which are combined during fertilization to form the zygote [2].     2. **Mendelian Laws**: The Mendel laws of segregation and independent assortment play a crucial role. Each parent passes a randomly selected allele for a given trait to the offspring, ensuring genetic diversity [5].  3. **Gene Activation and Regulation**: The process of gene activation and regulation is complex and involves several mechanisms:    - **Imprinting**: Genes can be marked with small chemical tags that determine whether they are activated or remain inactive after fertilization, depending on whether the modification was made by the father (in the sperm) or the mother (in the egg) [3], [4].    - **Maternal Effect**: The mother can alter gene activity in her offspring via the placenta [4].    - **Cis-regulation**: Instructions encoded within the embryo's DNA can directly control if and when a nearby gene becomes activated [4].  4. **Epigenetic Marks**: Epigenetic marks, which are less constrained and can be maintained, affect gene expression and phenotype later in development [6].  These factors collectively determine how traits are passed onto the resulting lifeform after the sperm combines with the egg.",
+  "question": "Once a sperm combines with an egg, what determines how traits are passed onto the resulting lifeform?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_12 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_12
new file mode 100644
index 0000000..34fd588
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_12
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - When Family Means More (or Less) Than Genetics.pdf",
+    "2012 - Mitochondrial Genomic Analysis of Late Onset.pdf",
+    "2017 - Parental influence on human germline de novo.pdf",
+    "2020 - Mitonuclear genomics and aging.pdf",
+    "2015 - Self-reported race or ethnicity in the age of genomic.pdf",
+    "2009 - When Family Means More (or Less) Than Genetics.pdf",
+    "2017 - Parental influence on human germline de novo.pdf",
+    "1996 - IDDM2-VNTR-encoded Susceptibility to Type 1 Diabetes.pdf",
+    "2012 - Mitochondrial Genomic Analysis of Late Onset.pdf",
+    "2016 - A genetic method for dating ancient genomes provides.pdf"
+  ],
+  "extraction_id": [
+    "baf15552-4198-5701-9175-c3fd31b2068e",
+    "ed29f84f-f2c9-5cbe-bab1-f5d5d2a334b6",
+    "a3b7edd7-f50f-53f1-b875-6d6733ddfde9",
+    "472c8adc-54e7-5c27-a7b8-882b7e49cd2b",
+    "6d68e979-ad62-5f85-ab03-5e898ce1c73b",
+    "baf15552-4198-5701-9175-c3fd31b2068e",
+    "163ce027-26ce-5625-8b63-5b7a910b4462",
+    "a324397e-1525-55ff-a9e8-92dc2aafa237",
+    "ed29f84f-f2c9-5cbe-bab1-f5d5d2a334b6",
+    "fcf5296e-6be4-5789-b1e1-ac57fef15119"
+  ],
+  "document_id": [
+    "7ba44399-3765-5ef5-9fdd-119b62117f66",
+    "5404a17c-34a9-5881-8b1a-2acacdc996a8",
+    "7c8bee23-b142-5fce-be77-6910277a136f",
+    "e05fdc09-c8d8-5134-a1fd-bf07a1564981",
+    "51ff0b84-193b-525a-b686-f29a423fcef9",
+    "7ba44399-3765-5ef5-9fdd-119b62117f66",
+    "7c8bee23-b142-5fce-be77-6910277a136f",
+    "bbaa99aa-3ae9-558d-bc97-7f85b6d0cf81",
+    "5404a17c-34a9-5881-8b1a-2acacdc996a8",
+    "5a5e67ea-4830-5fe8-95c3-ccfcc8324036"
+  ],
+  "id": [
+    "chatcmpl-ADZAODsOOCY3TdcinzGlvT4IHQgnR",
+    "f19ba98e-963f-5ecf-ad88-47215a3096e1",
+    "0e3b3480-c288-53cb-ac18-1d57478f9d34",
+    "06d4d82e-6eb9-59aa-a762-64de13149041",
+    "99a2cfc1-5a54-53af-b2a4-4c274e1d5ef1",
+    "612366c9-fcdc-5081-bc6d-47cd39922eeb",
+    "2ca2ab07-78b5-5268-93f1-297d83447163",
+    "db1fe67a-3d0c-549f-a54a-74ea0fa44d11",
+    "74ef6cdc-ea40-5d10-9ee8-b4288b3a70b4",
+    "27f40683-de33-5ec1-852d-6905f2dc389c",
+    "74484e0c-c862-5091-9fb5-957453a069af"
+  ],
+  "contexts": [
+    "variation with cultural practices around lineage. In certain societies, individuals place greater importance on (and have greater knowledge about) one side of the family than another (unilineal descent). Thus, individuals in patrilineal groups trace relationships through males only so that your fathers brothers children are members of your family, but not your fathers sisters (Kottak, 2007 ). They are members of their husbands group or family. Efforts to create",
+    "maternal lineage membership with those who weredirectly genotyped. Based on these pedigree (matrilineal) relation-",
+    "in three-generation families, and read pair tracing DNMs with phased variants. In the former approach, we determined the parent of origin as in our previous  analysis4. For example, if an offspring of the proband was a carrier of the DNM  allele and had haplotype sharing to paternal chromosome of the proband, we  assigned the mutation to the father. Meanwhile, if the offspring was not a DNM  allele carrier, we would assign it to the maternal germline. We restricted the haplo -",
+    "Unlike the nuclear genome, which requires both paternal and maternal contributions, mtDNA is inherited solely from the maternal lineage. It is unclear what advantage a uniparental mtDNA transmission confers, but one possibil-ity is to minimize the number of distinct genomes to maxi-mize the efficiency of a multi-genomic system (Hill etal. 2019). In fact, humans have developed complex, redundant mechanisms to ensure uniparental inheritance of mtDNA (DeLuca and OFarrell 2012; Rojansky etal. 2016). Paternal",
+    "c) Mitochondrial DNA (maternal line testing) markers: mitochondrial DNA or mtDNA haploid is the maternally inherited mitochondrial genome (mtDNA) [ 44]. All children inherit mtDNA from their mother, with no admixture from the father. Like Y-line DNA, mtDNA is passed intact from one generation to the next but through maternal line. Mitochondrial DNA does not follow any surname. In fact, the surname changes in every generation when women marry. Polymorphisms of mtDNA",
+    "a family pedigree may be hampered if the participant is not familiar with her mothers relatives, but her mothers brothers children (her cousins) may be able to supplement her overall family history. Knowledge about the cultural system of unilineal descent avoids assuming the universality of bilateral descent. Cultural beliefs such as these also have implications in the conduct of genetic research in terms of confidentiality and autonomy (Benkendorf et al.,",
+    "225 three-generation families using haplotype sharing (Fig. 1c and  Methods), 80.4% were found to be of paternal origin (Extended Data  Fig. 1). Figure 1e shows a strong relationship between the number of  paternal DNMs and the fathers age at conception (1.47 per year, 95%  CI 1.341.59) and a weaker impact of the mothers age on the number  of maternal DNMs (0.37 per year, 95% CI 0.300.45). The parental origin of all DNMs was also assessed by read pair",
+    "sistent with a maternal imprinting effect in familiesfrom France [18], the USA[10, 18, 21] (Figure 2; Table3) and Canada [27]. However, in a large family dataset from the UK, and in smaller data sets fromDenmark and Sardinia, the transmission of VNTRsusceptibility alleles is more pronounced frommothersthanfromfathers,andnowsignicantlysoinUK families (Figure 2; Table 3). Comparison of theresults from the USAwith those from the UK suggestthat unexplained inter-population differences in thisparent-of-origin",
+    "started with the largest matrilineage and worked down the list. Theparticipants selected for mtDNA sequencing were selected inde-pendent of their cognitive or dementia status. 274 matrilineageswere represented by this dataset. As a result, the sequencedmitochondrial genomes also represent as many different majormitochondrial haplogroups and clusters as possible (Table 1).Selection was made blind to case-control status. 287 samples weresent to Family Tree DNA (www.familytreedna.com) for Sangersequencing of",
+    "genetics-based population divergence studies. Am J Phys Anthropol 128(2):415 423.22. Helgason A, Hrafnkelsson B, Gulcher JR, War d R, Stefnsson K (2003) A populationwide coalescent analysis of Icelandic matrilineal and patrilineal genealogies: Evidence for a faster evolutionary rate of mtDNA lineages than Y chromosomes. Am J Hum Genet 72(6): 1370 1388. 23. Amster G, Sella G (2015) Life history effects on the molecular clock of autosomes and sex chromosomes. Proc Natl Acad Sci USA 113(6):1588 1593."
+  ],
+  "answer": "Genetic tracing is often matrilineal rather than patrilineal primarily because mitochondrial DNA (mtDNA) is inherited solely from the maternal lineage. Unlike nuclear DNA, which is a combination of both paternal and maternal contributions, mtDNA is passed down from mothers to all their children without any admixture from the father [4], [5]. This uniparental inheritance of mtDNA simplifies the tracing of maternal lineage as it remains relatively unchanged through generations, making it a reliable marker for genetic studies [4], [5]. Additionally, humans have developed complex mechanisms to ensure this uniparental inheritance of mtDNA, which further supports its use in genetic tracing [4].",
+  "question": "Why is genetic tracing matrilineal rather than patrilineal?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_13 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_13
new file mode 100644
index 0000000..49e3211
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_13
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2005 - Numerical Algorithms for Mapping of Multiple Quantitative Trait Loci in Experimental Populations.pdf",
+    "2005 -Ljungberg- Numerical algos for Multi QTL.pdf",
+    "2011 - Regulation of ribosomal RNA gene copy number and its role.pdf",
+    "2001 - Genome maintenance mechanisms.pdf",
+    "2008 - Rutter_s child and adolescent psychiatry-Blackwell Pub (2008).pdf",
+    "2013 - Causes of Genome Instability.pdf",
+    "2009 - Basic Genetics and Genomics A Primer for Nurses.pdf",
+    "2007 - Promoting_Student_Scientific_Literacy_of_Molecular Genetics and Genomics.pdf",
+    "2013 - Causes of Genome Instability.pdf",
+    "2007 - DNA replication stress, genome instability and aging.pdf"
+  ],
+  "extraction_id": [
+    "3f482661-0759-54cf-9926-8a39abb538bf",
+    "33c27a82-4633-5f0c-9d9e-716aee665879",
+    "28addd51-38b1-5405-bed4-140f7224da0b",
+    "17bbb094-4a6f-5931-be1d-ee46abc25820",
+    "86760f12-2e7c-56c6-80d8-0d62c611843d",
+    "67e0ca38-bd7c-551f-9941-bcd6025a630d",
+    "6e7863c0-dc75-550a-b3ca-9fb0d95af788",
+    "67369433-749b-5d6a-b5ef-3f0afe78b767",
+    "e1efc4b8-b33d-513f-b6cb-9f35de5eda30",
+    "908ae89d-39b7-51ff-9bc4-c4a1de926b87"
+  ],
+  "document_id": [
+    "dd7d3ea5-b23a-514e-898f-a4259ce6f6f9",
+    "bea0655c-7ef4-5754-ba14-817b72a21be2",
+    "20147943-f329-5ac6-8343-3cea72fdc040",
+    "db0de7b5-6c1c-521c-be6d-0ea91c700277",
+    "59daba11-206e-5bbc-8833-9d1b661532b5",
+    "63f87ec0-9437-5d67-b36d-0b24059e9c9a",
+    "c37e2ace-171b-5776-8969-86eda9736481",
+    "755f34c4-cc06-5275-a744-16d48162b012",
+    "63f87ec0-9437-5d67-b36d-0b24059e9c9a",
+    "dd83ce7f-dfe1-5fa7-8509-bfdb1e27af9f"
+  ],
+  "id": [
+    "chatcmpl-ADZAStB25S6kWtJMcUZWpQvWYeOgv",
+    "4c576a55-b4cf-569f-8771-3d2a81480104",
+    "3a88ba5e-d93b-5c50-a03a-02def897390f",
+    "8049fc56-1fd2-58ba-9d5c-9529d4cc7e84",
+    "c793587e-e46f-5b48-9e49-c150637c5f5a",
+    "440ffc7a-2810-5245-bc20-9284d6861472",
+    "55f07e34-bcb4-5a68-a222-477378f6c9d0",
+    "53707c68-7cf1-51aa-9d4c-1eb4a9816182",
+    "dc2dc054-f0f9-5e78-92b0-1caa0a6239e0",
+    "f18e23f1-67a3-5d7f-831e-358fa44e7873",
+    "134f4d33-1645-591b-ac20-d8d8b298bcfc"
+  ],
+  "contexts": [
+    "the DNA, i.e. the whole genome. During replication the two strands of themother cell DNA are separated, and new nucleotides are put together to maketwo double helices identical to the original one, see Figure 2.1. TAAGACCG AT T CTGGCCCGTGGC. . . . . . .. . ATTCTGGCTAAGACCG. . . . . . . . Figure 2.1: A DNA chain consists of two strands of complementary nucleotides. When DNA is replicated, two double chains identical to the original one are created.",
+    "the DNA, i.e. the whole genome. During replication the two strands of themother cell DNA are separated, and new nucleotides are put together to maketwo double helices identical to the original one, see Figure 2.1. TAAGACCG AT T CTGGCCCGTGGC. . . . . . .. . ATTCTGGCTAAGACCG. . . . . . . . Figure 2.1: A DNA chain consists of two strands of complementary nucleotides. When DNA is replicated, two double chains identical to the original one are created.",
+    "The mechanism to maintain the rDNA copy number The gene amplication mechanism that counteracts recombination-mediated loss of rDNA copies is well studied in budding yeast [ 6,11]. During the S phase of the cell cycle, replication starts from replication origins, and isinhibited at the replication fork barrier site (RFB) by the function of the fork blocking protein, Fob1 (Fig. 3)[12]. This inhibition works as a recombinational hotspot toinduce amplication for copy number recovery as follow;",
+    "S and G2 when the DNA is replicated, providing a pristine secondcopy of the sequence (sister chromatid) for aligning the breaks. Incontrast, the less-accurate end joining is most relevant in the G1phase of the cell cycle, when a second copy is not available 14.  Finally, some single repair proteins directly revert certain injuries, such as O6-methylguanine methyltransferase, which removes  O6-methyl guanine. This highly mutagenic lesion permits base",
+    "Replication",
+    "genotoxic agents and to guarantee faithfulchromosome duplication and transmission to the offspring. In addition to DNA damage repair, cells monitor replication to minimize er-rors of DNA synthesis. In eukaryotes, cell-cycle checkpoints guarantee coordination of DNA synthesis and DNA repair with cell division.Genome instability is mainly due to sporadic replication or repair errors but can also take place in response to developmental or environ-mental signals, as occurs in meiosis, and antigen",
+    "This section will explain how cells normally divide. It will also desc ribe how an unexpected change in  the structure of DNA can sometimes cause harm to th e body. New tools to study genetic variations of  common diseases and to identify genetic variatio ns common to specific diseases will also be  presented.  Cell Division   Humans grow and develop as a result of a process called cell  division. There are two types of cell division  mitosis and meiosis.",
+    "and replicated (by a templating mechanism).  Each DNA molecule in a cell forms a single chromosome. (NRC, pg. 185, 9-12:C2#1) 4. Genes as information for building proteins: The genetic information in DNA molecules provide the instructions on assembling protein molecules. The code is virtually the same for all life forms. (AAAS, pg. 114, 5C:9-12#4 ) 5. Molecular nature of genes and mutations: Genes are segments of DNA molecules.  Inserting, deleting, or substituting DNA segments can alter genes.  An altered",
+    "When a replication fork encounters a DNA adduct, cells induce DNA damage toler-ance mechanisms that allow completion of replication. Adducts can be bypassed by postreplicative repair via translesion poly-merases (either faithful or error-prone) or via error-free template switching using the sister chromatid (64, 105). Postreplicativerepair guarantees genome stability by allowing completion of replication (albeit at the expense",
+    "genome instability in part because of the unique structureof replicating DNA molecules (Figure 2). When single-strand lesions occur in non-replicating molecules of DNA,the overall integrity of chromosomes is maintained byhydrogen bond base pairing on either side of these lesionsuntil they are repaired (Figure 2A). In contrast to non-replicating DNA, replicating DNA at replication forkscontains unwound, highly recombinogenic single-strandedtemplate DNA before this DNA is converted to double-strand DNA by"
+  ],
+  "answer": "DNA replication is a critical process that ensures the accurate copying of genetic information during cell division. Here is an explanation of the process and how it maintains accuracy:  1. **Separation of DNA Strands**: During replication, the two strands of the mother cell DNA are separated [1], [2]. This separation is essential to expose the nucleotide sequences of each strand, which will serve as templates for the synthesis of new complementary strands.  2. **Formation of New Strands**: New nucleotides are assembled to form two double helices that are identical to the original DNA molecule [1], [2]. Each original strand pairs with a newly synthesized complementary strand, resulting in two DNA molecules, each consisting of one old and one new strand. This method is known as semi-conservative replication.  3. **Initiation at Replication Origins**: Replication begins at specific locations on the DNA molecule called replication origins. During the S phase of the cell cycle, replication starts from these origins [3].  4. **Replication Fork and Enzyme Activity**: The replication process involves the formation of a replication fork, where enzymes like DNA helicase unwind the DNA double helix. DNA polymerase then adds new nucleotides to the growing DNA strand by matching them with the complementary nucleotides on the template strand.  5. **Error Minimization and Repair Mechanisms**: Cells have mechanisms to monitor replication and minimize errors. For instance, cell-cycle checkpoints ensure coordination of DNA synthesis and repair with cell division [6]. Additionally, specific repair proteins can revert certain DNA injuries, ensuring the fidelity of the replication process [4].  6. **Postreplicative Repair**: If replication encounters obstacles such as DNA adducts, cells can induce DNA damage tolerance mechanisms. These mechanisms allow the completion of replication through error-free template switching using the sister chromatid or through translesion polymerases [9].  By following these steps and employing various error-checking and repair mechanisms, DNA replication ensures that genetic information is accurately copied and transmitted to daughter cells during cell division.",
+  "question": "Explain the process of DNA replication and how it ensures accurate copying of genetic information during cell division."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_14 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_14
new file mode 100644
index 0000000..d5f4e5d
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_14
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2021 - Human Molecular Genetics and Genomics.pdf",
+    "2015 - A new age in functional genomics using CRISPR Cas9 in arrayed library screening.pdf",
+    "2016 - Genome editing comes of age.pdf",
+    "2016 - Dissecting diabetes metabolic disease.pdf",
+    "2018 - Neuro-Immuno-Gene- and GenomeEditing-Therapy for Alzheimer\u2019s.pdf",
+    "2016 - Genome editing comes of age.pdf",
+    "2021 - Human Molecular Genetics and Genomics.pdf",
+    "2020 - Functional Genomics in Pancreatic \u03b2 Cells Recent Advances in Gene Deletion and Genome Editing Technologies for Diabetes Research.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf",
+    "2020 - Clinical Genetics and Genomics of Aging.pdf"
+  ],
+  "extraction_id": [
+    "08a2c0e6-8ca8-5a72-974c-3f1e27ba1b15",
+    "49b81415-ef6f-5cc4-bb30-71e971070ebe",
+    "190e8838-4f61-5431-8848-98564ded7140",
+    "ebabc771-1777-56c1-9101-c1677c5ae908",
+    "66dbf4f0-2b37-5219-9eeb-0a560df8d888",
+    "c7b143d7-347c-5160-bfd4-82283b342d7d",
+    "08a2c0e6-8ca8-5a72-974c-3f1e27ba1b15",
+    "fe5bf2df-2eda-5ef0-8aad-79bbc5b898d6",
+    "54972d7f-0ddc-5076-9d58-890a85f71332",
+    "c072d600-8450-5842-ade1-aefd03854312"
+  ],
+  "document_id": [
+    "68e362a5-e580-5a4d-8d41-6a138c873ede",
+    "20df9469-e1cc-500e-ac30-fbba981d7aee",
+    "4078087a-c2a4-5c58-95b5-4ae243794800",
+    "eee2f79d-e093-52fb-871a-798fd859235e",
+    "cc0a025b-71e7-5712-bbf7-4ee1e0f769ef",
+    "4078087a-c2a4-5c58-95b5-4ae243794800",
+    "68e362a5-e580-5a4d-8d41-6a138c873ede",
+    "51350055-d53c-5692-ab53-337b8a8bafd6",
+    "62b635c3-040e-512a-b016-6ef295308a1e",
+    "62b635c3-040e-512a-b016-6ef295308a1e"
+  ],
+  "id": [
+    "chatcmpl-ADZAaZ3EtxdQEfdMEkkYN8ccIeUPg",
+    "ffbca864-26db-5f36-8ad4-3b8d24d46de6",
+    "2fbabecd-22c1-5570-8f38-bc934d463710",
+    "6dfc48be-a762-55d6-9aba-799d80e8140d",
+    "a3d6f231-29aa-5cf6-b856-004d3d9dd9c1",
+    "a4aa9de7-cc9f-5c3f-a9fe-c37a47faa5b7",
+    "ff2d183b-c5be-5e05-94c8-e2db379dcd96",
+    "4474c4e9-bc07-5610-8bb2-dafe5c95774b",
+    "6ba3cf43-be4d-561f-ad84-f79921cab37e",
+    "176b7aa5-17ef-590d-8807-1aa7def904bb",
+    "e5eef445-772e-5721-bb5f-24566a61e4e3"
+  ],
+  "contexts": [
+    "neered nucleases, CRISPR-Cas9  tools have accelerated the pace of  genomic research by permitting  highly efficient knockouts or  edits of virtually any gene in cells  or model organisms. Multiple CRISPR-Cas9based clinical trials  are in progress or are expected  to begin soon. Although Cas9- engineered cells havent yet dem - onstrated efficacy at scale, early  trial results suggest that such  cells are stable and dont cause  acute adverse reactions in humans.  Long-term safety is yet to be de -",
+    "stageissetforCRISPRtomakeanenormousimpactongenomic screening and thus scientic discovery in the coming years, and recent demonstrations of this system have shown great promise (Shalem etal., 2015 ).However,a number of technical challenges must be addressed in order to maximize the benet of this technology. In this review, we will discuss current applications of CRISPR in functional genomics and provide a perspective on futuredevelopmentsinthisarea. CRISPR/Cas9 Genome Editing",
+    "heralding the age of genome editing. Furthermore, Cas9 or guide RNAs have been linked to various effector proteins to enable targeted gene regulation 12,13 and epigenome modifications14,15.  It is worth noting, however, that many of these feats had been demonstrated previously using other nucleases or DNA-binding proteins 1,16. In this Perspective, I shed light on early genome  editing platforms that laid the groundwork for the widespread use of CRISPRCas9 in research and medicine (Fig. 1 ).",
+    "cline- or Tet-regulated Cas9 system. Current CRISPR/Cas systems arefrom Streptococcus pyogenes ,Streptococcus thermophilus ,Neisseria meningitides and Treponema denticola .2.5. Caveats of advanced genome editing tools Off-target effects . The DNA-binding domains of ZFNs and TALENs need to be very speci c for the target site to avoid off-target cleavage, which results in unwanted mutations and potentially cytotoxic effects [27]. CRISPR/Cas9 is also known to generate off-target alterations,",
+    "CRISPR/CAS9 HOLDS SIGNIFICANT PROMISE FOR THE DEVELOPMENT OFNEW AD MODELS AND PRECISIONTARGETED AD THERAPY Clustered regularly interspaced short palindromic repeat (CRISPR)-Cas nucleases have revolutionizedthe eld of gene editing and have tremendous appli-cation in the eld of molecular medicine [98102].Despite a signicant surge in CRISPR/Cas9-mediated genome editing in various disease models,the progress in the eld of AD has lagged behindsubstantially. We believe that genome editing can sig-",
+    "81. Applications for CRISPRCas9 beyond genome editing",
+    "CRISPR-Cas9 can be used to in - duce genome edits by creating  targeted DNA breaks that trigger  site-specific DNA repair. In next- generation formats, it can also  control the transcriptional output  of genes or alter genome se - quences using a process of nu - cleotide base editing  that does not require  repair of DNA breaks.  As these technolo - gies continue to mature, it will  become increasingly possible to  alter cellular genomes efficiently  and accurately. Coming on the heels of engi -",
+    "on transcriptional interfere nce (CRISPRi) and activation (CRISPRa) have also harnessed Cas9-based technologies for use in genome-wide studies ( 59,174). In addition, recent improvements in lentiviral library generation and propagation,as well as large-scale DNA and RNA synthesis, have allowedCRISPR-Cas9 technology to be exploited across multiple modelplatforms ( 59,175178). nCas9 The CRISPR-Cas9 system can tolerate certain mismatches to the DNA target since the required gRNAs are short. A disadvantage,",
+    "13. Kleinstiver BP, etal. High-fidelity CRISPRCas9 nucleases with no detectable genome-wide  off-target effects. Nature. 2016;529:4905.  14. Brane A, Tollefsbol T.Targeting telomeres and telomerase: studies in aging and disease uti- lizing CRISPR/Cas9 technology. Cells. 2019;8:186.  15. Wang H, etal. One-step generation of mice carrying mutations in multiple genes by CRISPR/ Cas-mediated genome engineering. Cell. 2013;153:9108.",
+    "Since its discovery, CRISPR-Cas technology has ignited a biological revolu- tion by providing a highly versatile platform that allows fast and efficient genome  editing in an ever-growing list of organisms. In this chapter we will first describe  the most recent advances in the development and application of the CRISPR-Cas  platform in biomedical research. Then we will discuss the most recent and notable  basic research applications of this technology in the study of the molecular causes"
+  ],
+  "answer": "### Potential Benefits of CRISPR-Cas9:  1. **Accelerated Genomic Research**: CRISPR-Cas9 tools have significantly accelerated the pace of genomic research by enabling highly efficient knockouts or edits of virtually any gene in cells or model organisms [1].     2. **Clinical Trials and Stability**: Multiple CRISPR-Cas9-based clinical trials are in progress, and early trial results suggest that CRISPR-engineered cells are stable and do not cause acute adverse reactions in humans [1].  3. **Functional Genomics**: CRISPR-Cas9 is set to make a significant impact on genomic screening and scientific discovery, with recent demonstrations showing great promise [2].  4. **Targeted Gene Regulation and Epigenome Modifications**: CRISPR-Cas9 has been linked to various effector proteins to enable targeted gene regulation and epigenome modifications [3].  5. **Precision Medicine**: CRISPR-Cas9 holds significant promise for the development of new models and precision-targeted therapies for diseases such as Alzheimer's Disease (AD) [5].  6. **Versatility and Efficiency**: CRISPR-Cas9 provides a highly versatile platform that allows fast and efficient genome editing in an ever-growing list of organisms [10].  ### Potential Risks of CRISPR-Cas9:  1. **Off-Target Effects**: CRISPR-Cas9 is known to generate off-target alterations, which can result in unwanted mutations and potentially cytotoxic effects [4].  2. **Technical Challenges**: There are several technical challenges that need to be addressed to maximize the benefits of CRISPR-Cas9 technology [2].  3. **Long-Term Safety**: While early trial results are promising, the long-term safety of CRISPR-Cas9-engineered cells is yet to be determined [1].  4. **Mismatch Tolerance**: The CRISPR-Cas9 system can tolerate certain mismatches to the DNA target, which could potentially lead to unintended edits [8].  By considering these benefits and risks, researchers and clinicians can better navigate the development and application of CRISPR-Cas9 technologies.",
+  "question": "What are the potential benefits and risks associated with gene editing technologies like CRISPR-Cas9?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_15 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_15
new file mode 100644
index 0000000..17dff0e
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_15
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - eQTL analysis in mice and rats.pdf",
+    "2015 - Genetic Control of Survival and Weight Loss during Pneumonic Burk.pdf",
+    "2015 -Emery- Genetic Control of Survival and Weight Loss during Pneumonic Burk.pdf",
+    "2006 - From_gene_to_behavior_and_back_again_new.pdf",
+    "2005 - quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2005 -Broadkin- quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2009 - Experimental_Evolution.pdf",
+    "2009 - Garland_and_Rose_Experimental_Evolution.pdf",
+    "2005 - quantitative-trait-analysis-in-the-investigation-of-function-and.pdf",
+    "2016 - Social interactions and indirect genetic effects on complex juvenile and adult traits.pdf"
+  ],
+  "extraction_id": [
+    "71981bfb-284e-50ad-854e-2055c07f77a7",
+    "615ee0cd-5960-57e5-b4e6-56e4b8020a1b",
+    "268a23e8-f528-5b59-89f2-188331e0a03c",
+    "64c0287d-aeea-52eb-a074-e9591c5593ae",
+    "9de93371-6239-53c2-b42c-71f615a0614b",
+    "0a5c759e-8dab-55f1-ac59-e8211ec683b8",
+    "8ee78018-b998-590c-99ab-788a447ede81",
+    "cbce50ea-be78-5d54-beb1-849222c5bfdd",
+    "0a895880-91c0-5079-b258-73926b38430f",
+    "0b91ce42-1ba4-530c-8d77-6ddbdc0e759d"
+  ],
+  "document_id": [
+    "8d67ea90-f7b1-5bb8-937c-4a9eceddff43",
+    "ae1025b0-1410-51ae-9be2-26fa2e9d5808",
+    "a9aceace-bf48-5472-b54c-59a458a84c62",
+    "7a088b36-11b7-5379-bfe5-ce571e11de07",
+    "0dc730ba-4ff4-52aa-a988-71075113c416",
+    "e6027e7f-aec0-5e76-8aff-96b36389e701",
+    "34821353-1b74-5ee2-ac39-66dd46f145bf",
+    "496faa7f-9623-5ab7-9816-7c3755abb3aa",
+    "dac1c73c-0b5f-5a54-bb12-7e8b654009c0",
+    "06e126d3-b75d-57db-8edb-09de6ae13b24"
+  ],
+  "id": [
+    "chatcmpl-ADZAiadgDkXNlnQWQS4n6DqZI6elc",
+    "73540700-b5cf-5838-852b-b281ca086140",
+    "374c456a-d1db-5b4a-8713-97abe4162d77",
+    "b9d52798-0235-5018-bccd-560565d16cc3",
+    "fef212bc-631b-591d-b8e3-d1523da0507d",
+    "c8f17022-aeae-5242-9082-d6d1eee4c4bf",
+    "1b2de424-be9f-572d-bd62-dc2ecd92192b",
+    "f72795a1-66c3-5a98-84bc-b085e8008073",
+    "31a32dc5-81ac-52ba-a463-c61e293f21e5",
+    "b660d882-1cb0-5150-ae76-8eb3ccb88a58",
+    "985378d7-e164-581b-ac1c-97bbcda9c06f"
+  ],
+  "contexts": [
+    "While most of the Y chromosome does not undergo recombination, the recombination rate of the X chromosomeis slower than that of the autosomes. This has important consequences on the detection of significant QTLs. For a comprehensive view of these issues, see(43). 9.Probe hybridization artifacts When several probes are available for the same gene, it is not uncommon to observe a difference in the mapping results",
+    "8 QTL Mapping     Allelic variation exists among natural populations and inbred strains, and this is  reflective of the segregation of quantitative tr ait loci (QTLs) [96]. QTLs are stretches of  DNA that are closely linked to genes that underlie a phenotype of interest. QTL analysis has been proven to be an invaluable tool to  help unravel heritable traits, by enabling  researchers to map different quantitative traits back to the genomic location involved in the regulation of these phenotypes.",
+    "8 QTL Mapping     Allelic variation exists among natural populations and inbred strains, and this is  reflective of the segregation of quantitative tr ait loci (QTLs) [96]. QTLs are stretches of  DNA that are closely linked to genes that underlie a phenotype of interest. QTL analysis has been proven to be an invaluable tool to  help unravel heritable traits, by enabling  researchers to map different quantitative traits back to the genomic location involved in the regulation of these phenotypes.",
+    "The basic  pr emise  of QTL  an alysis  is simple  (Ph illips  and Belknap, 2002 ) . First,  one must  meas  ure a speci  c phen  otype  within  a popul  ation. Next, the population must be genotyped at a hundred or more marker loci186 Boehm II et al.",
+    "genes underlying QTLs in animals and plants (see for example Shirley et al 2004,Korstanje & Paigen 2002, Fridman et al 2004). I should also point out, though, that even in a single QTL region isolated in a congenic strain, it is possible that there is more than one allele that aects the phenotype. So, you have a fair pointabout the challenges and complexities of QTL analysis. Koolhaas: There are dierent questions underlying both approaches. The QTL",
+    "genes underlying QTLs in animals and plants (see for example Shirley et al 2004,Korstanje & Paigen 2002, Fridman et al 2004). I should also point out, though, that even in a single QTL region isolated in a congenic strain, it is possible that there is more than one allele that aects the phenotype. So, you have a fair pointabout the challenges and complexities of QTL analysis. Koolhaas: There are dierent questions underlying both approaches. The QTL",
+    "through analysis of line crosses, quantitative trait loci (QTL) mapping, and verification of candidate genes with quantitative complementation tests or genetic engineering (e.g.,McGuire and Tully 1987; Chandra et al. 2001; Dierick and Greenspan 2006; Edwardset al. 2006). They can also be used to study the underlying physiological, neural, andmolecular mechanisms of the differences in behavior between selected and controllines, or between divergently selected lines.",
+    "through analysis of line crosses, quantitative trait loci (QTL) mapping, and verification of candidate genes with quantitative complementation tests or genetic engineering (e.g.,McGuire and Tully 1987; Chandra et al. 2001; Dierick and Greenspan 2006; Edwardset al. 2006). They can also be used to study the underlying physiological, neural, andmolecular mechanisms of the differences in behavior between selected and controllines, or between divergently selected lines.",
+    "genetic background. Gene identification of QTL should be distinguished from identification of the quanti- tative trait nucleotide (QTN). The latter is a daunting task, since SNPs are so frequent. Final proof for a QTN in mice would require placing a genomic segment containing theputative QTN from a donor mouse strain on the background of another strain using homologous recombination and reproducing the phenotype of the donor strain.",
+    "measuring correlations between genetic markers and phenotypic  traits in a population. Individuals are scored for their  phenotype      for  a particular trait, and their genotype at a marker. If there is a differ- ence in mean phenotype between those individuals with one geno- type at a particular locus compared with the other, than we can infer  that there is a QTL linked to that marker [ 40 ,  153 ]. 2.3  Analysis and QTL  MappingDavid G. Ashbrook and Reinmar Hager"
+  ],
+  "answer": "To differentiate between X and Y DNA in the context of DNA tracing and determining quantitative trait loci (QTLs), one must consider the unique characteristics of each chromosome:  1. **Recombination Rates**: The Y chromosome largely does not undergo recombination, whereas the X chromosome has a slower recombination rate compared to autosomes. This difference in recombination rates has significant implications for the detection of QTLs [1].  2. **QTL Mapping**: QTLs are stretches of DNA closely linked to genes that influence specific phenotypes. The process of QTL mapping involves measuring phenotypes within a population and genotyping at numerous marker loci [2], [4]. Given the unique recombination characteristics of the X and Y chromosomes, the strategies for mapping QTLs on these chromosomes may differ.  3. **Genetic Markers and Phenotypic Traits**: When measuring correlations between genetic markers and phenotypic traits, the lack of recombination on the Y chromosome and the slower recombination on the X chromosome must be taken into account. This affects how QTLs linked to these chromosomes are identified and analyzed [10].  In summary, the primary differences between X and Y DNA in the context of DNA tracing and QTL determination lie in their recombination rates and the subsequent impact on QTL detection and mapping strategies.",
+  "question": "How does one tell the difference between X and Y DNA, with repsect to DNA tracing and determining QTLs?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_16 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_16
new file mode 100644
index 0000000..83f0761
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_16
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf",
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "2010 - Teaching Bioinformatics and Neuroinformatics by using Free Web-Based Tools.pdf",
+    "2012 - Biological Databases for Behavioral Neurobiology.pdf",
+    "2008 - (Infectious Disease) Karl A. Western (auth.), Vassil St. Georgiev PhD, Karl A. Western MD, John J. McGowan PhD (eds.) - National Institute of Allergy and Infectious Diseases, NIH_ Frontiers in Researc (3).pdf",
+    "2008 - Biotools for Determining the Genetics of Susceptibility to Infectious Diseases.pdf"
+  ],
+  "extraction_id": [
+    "49a42e3c-e1f9-5433-9643-192a592454d4",
+    "a1588a50-8f88-5d50-9232-706bdc46ec88",
+    "11762190-61cd-585d-96c1-7aa6717d9d47",
+    "af1c63bf-772d-554e-be88-bd62daee49ee",
+    "025c4afb-d749-54a7-a183-9a7b8b1332c7",
+    "c0098aa5-5eba-5b6a-97f2-661388daeb82",
+    "0017cf22-e712-5a41-9bb7-ea3632bb825a",
+    "9a38ba60-52da-5f67-9c75-db1ac7a7b7ec",
+    "16c8fbb0-ab2a-563f-a6b2-e0d8733b69fb",
+    "fe6eb7f0-9f09-50f8-a7a1-c71e507226d5"
+  ],
+  "document_id": [
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0",
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "c802c4ea-e99d-501a-ba20-1cd1b369dfc3",
+    "99d3a5c1-8511-5fba-92ce-1ceef2b6c402",
+    "4db8c752-c8e2-5f6d-a091-dc4f1d0c48bc",
+    "fcbbb3ce-6524-50e3-9f8d-c191dc551231"
+  ],
+  "id": [
+    "chatcmpl-ADZArO5xKdt382Vj4oDYlsZwv0rng",
+    "fcbb83a7-84f2-55cd-b26d-80883a022c52",
+    "7f1ea794-1c26-5a90-abe3-f60f338f5985",
+    "8473f1fc-d615-54de-92de-a3faf5e2045e",
+    "4741caf7-1306-52e8-874c-fa200f067978",
+    "f74f9aa9-3464-58e8-a0f6-e3e38efa3c40",
+    "4121b591-0dda-5347-9833-23e3d9c6d8fe",
+    "f337b34c-de96-5b8c-ac3e-80417634b5c1",
+    "df5e9619-d45e-5958-a88d-d33ecc59387d",
+    "71eac758-37cb-5fec-8380-7d9f4d4c2845",
+    "e4180707-bb0f-5b00-8de7-f6937bc38e07"
+  ],
+  "contexts": [
+    "for people to exchange data easily over the Web. Two other notable developments are BioMart and GBrowse. The BioMart project (http://www.biomart.org/), originally a spin-off from Ensembl, offers a generic data management system that allows complex searches of biological data such as sequence annotation. The GBrowse project (Stein et al. , 2002; http://www.gmod.org/) has produced a generic genome browser that can be customized to organize, display and query a new genome scale data set. These",
+    "for people to exchange data easily over the Web. Two other notable developments are BioMart and GBrowse. The BioMart project (http://www.biomart.org/), originally a spin-off from Ensembl, offers a generic data management system that allows complex searches of biological data such as sequence annotation. The GBrowse project (Stein et al. , 2002; http://www.gmod.org/) has produced a generic genome browser that can be customized to organize, display and query a new genome scale data set. These",
+    "for people to exchange data easily over the Web. Two other notable developments are BioMart and GBrowse. The BioMart project (http://www.biomart.org/), originally a spin-off from Ensembl, offers a generic data management system that allows complex searches of biological data such as sequence annotation. The GBrowse project (Stein et al. , 2002; http://www.gmod.org/) has produced a generic genome browser that can be customized to organize, display and query a new genome scale data set. These",
+    "(http://ensembl.org/ ) and the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/ ) all provide portals to the most current, and archived public assemblies. These sites also provide means of searching the assem- blies, such as BLAST (Altschul et al. , 1997), BLAT (Kent, 2002) and SSAHA (Ning et al. , 2001) as well as precomputed annotation for the genome assemblies that can be readily incorporated into comparative genomic analyses.",
+    "(http://ensembl.org/ ) and the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/ ) all provide portals to the most current, and archived public assemblies. These sites also provide means of searching the assem- blies, such as BLAST (Altschul et al. , 1997), BLAT (Kent, 2002) and SSAHA (Ning et al. , 2001) as well as precomputed annotation for the genome assemblies that can be readily incorporated into comparative genomic analyses.",
+    "(http://ensembl.org/ ) and the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/ ) all provide portals to the most current, and archived public assemblies. These sites also provide means of searching the assem- blies, such as BLAST (Altschul et al. , 1997), BLAT (Kent, 2002) and SSAHA (Ning et al. , 2001) as well as precomputed annotation for the genome assemblies that can be readily incorporated into comparative genomic analyses.",
+    "resources. We present an easy-to-adopt module that weaves together several important bioin-formatic tools so students can grasp how these tools are used in answering research questions.Students integrate information gathered from websites dealing with anatomy (Mouse BrainLibrary), quantitative trait locus analysis (WebQTL from GeneNetwork), bioinformatics and geneexpression analyses (University of California, Santa Cruz Genome Browser, National Center forBiotechnology Informations Entrez Gene, and the",
+    "References Altman RB. Building successful biological databases. Briefings in Bioinformatics. 2004; 5:45.  [PubMed: 15153301]  Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: Tool for the  unification of biology. The Gene Ontology Consortium. Nature Genetics. 2000; 25:2529.  [PubMed: 10802651]  Ashish N, Ambite JL, Muslea M, Turner JA. Neuroscience data integration through mediation: an",
+    "Sequences, Protein Structures, Complete Genomes, Tax- onomy, Medical Genetics resources (see later), and others (see   http://www.ncbi.nlm.nih.gov/Database/index.html     for a complete listing of databases).  Entrez   PubMed provides  access to full-text articles at journal websites and other related web resources, some of which are free to the  public. This site also provides links to other molecular biology resources.      The National Center for Biotechnology Information (  http://",
+    "Sequences, Protein Structures, Complete Genomes, Tax- onomy, Medical Genetics resources (see later), and others (see   http://www.ncbi.nlm.nih.gov/Database/index.html     for a complete listing of databases).  Entrez   PubMed provides  access to full-text articles at journal websites and other related web resources, some of which are free to the  public. This site also provides links to other molecular biology resources.      The National Center for Biotechnology Information (  http://"
+  ],
+  "answer": "For text and biological resources, it seems you are referring to adding web resources such as Ensembl to your system. This is evident from the context which mentions various web-based biological data management systems and genome browsers like BioMart, GBrowse, and Ensembl [1], [4]. These resources provide portals to current and archived public assemblies, as well as tools for searching and annotating genome assemblies [4], [6]. Therefore, it appears you are more focused on integrating web resources rather than books.",
+  "question": "For text and biological resources, do you mean add some books (on biology stuff) or/and web resources (as ensembl) on your system?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_17 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_17
new file mode 100644
index 0000000..42b1aeb
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_17
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf"
+  ],
+  "extraction_id": [
+    "4ef84d22-b428-5386-bbc0-39dbd364d3d7",
+    "7524bdfa-63f8-57c1-b5fe-1edcf11c275e",
+    "c8e9c4b7-19c6-5426-83a2-6f8628b68ceb",
+    "a3ae6875-b0fc-5a4e-866f-4fee99c7d2a2",
+    "bb247bfe-333b-553a-94e6-2dc1b13b4723",
+    "9c89683f-aca5-57f9-b28d-62e9eb64377b",
+    "23bb58ad-7835-58f4-862f-dd17e1ec5140",
+    "52fc5fdb-48b5-5c1e-a8d2-1e67d7702c9f",
+    "20d0c226-76aa-5c1e-85cc-9d5c1bcce2f2",
+    "bb247bfe-333b-553a-94e6-2dc1b13b4723"
+  ],
+  "document_id": [
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0"
+  ],
+  "id": [
+    "chatcmpl-ADZAv1loJK3Vz78QdzvVvEogO2ngd",
+    "c36215f6-2230-58ef-b3eb-44d1799ba5c2",
+    "89a578c7-5961-5b88-9a6d-f338216702c3",
+    "81e589eb-aa51-5f2a-966f-31928fb31943",
+    "1bf9bb72-ebaa-51d1-82ce-aae2f16dd92b",
+    "f0c00edb-f07d-5975-a16b-16a072d0f2d4",
+    "e2e526cb-0ac3-51ff-a1c5-43ff032b5558",
+    "66294988-1566-5bec-8f63-658ea9011e26",
+    "b19972d1-7ec6-5f66-ac2f-518e69c5f22b",
+    "69fdc34c-c187-5c7a-973c-a629045841a7",
+    "7e15e9b2-c731-5ab0-85c0-b6b432623220"
+  ],
+  "contexts": [
+    "supported by a signicant BLAST match to one or more expressed sequences or proteins. Ensembl also identies the positions of known human genes from public sequence database entries, usually using GENEWISE to predict their exon structures. The total set of Ensembl genes should therefore be a much more accurate reection of reality than ab initio predictions alone, but it is clear that some novel genes are missed (Hogenesch et al. , 2001). Of the many novel genes that are detected, some are",
+    "supported by a signicant BLAST match to one or more expressed sequences or proteins. Ensembl also identies the positions of known human genes from public sequence database entries, usually using GENEWISE to predict their exon structures. The total set of Ensembl genes should therefore be a much more accurate reection of reality than ab initio predictions alone, but it is clear that some novel genes are missed (Hogenesch et al. , 2001). Of the many novel genes that are detected, some are",
+    "supported by a signicant BLAST match to one or more expressed sequences or proteins. Ensembl also identies the positions of known human genes from public sequence database entries, usually using GENEWISE to predict their exon structures. The total set of Ensembl genes should therefore be a much more accurate reection of reality than ab initio predictions alone, but it is clear that some novel genes are missed (Hogenesch et al. , 2001). Of the many novel genes that are detected, some are",
+    "Ostell/Spidey/ SSAHA at Sanger Institute http://www.sanger.ac.uk/Software/analysis/SSAHA/ human and mouse genomes, where there are large full-length cDNA collections to guide the hunt for genes, Ensembl should be very reliable. From the beginning, many genomic features other than predicted genes were included in Ensembl: different repeat classes, cytological bands, CpG island predic- tions, tRNA gene predictions, expressed sequence clusters from the UniGene database",
+    "Ostell/Spidey/ SSAHA at Sanger Institute http://www.sanger.ac.uk/Software/analysis/SSAHA/ human and mouse genomes, where there are large full-length cDNA collections to guide the hunt for genes, Ensembl should be very reliable. From the beginning, many genomic features other than predicted genes were included in Ensembl: different repeat classes, cytological bands, CpG island predic- tions, tRNA gene predictions, expressed sequence clusters from the UniGene database",
+    "Ostell/Spidey/ SSAHA at Sanger Institute http://www.sanger.ac.uk/Software/analysis/SSAHA/ human and mouse genomes, where there are large full-length cDNA collections to guide the hunt for genes, Ensembl should be very reliable. From the beginning, many genomic features other than predicted genes were included in Ensembl: different repeat classes, cytological bands, CpG island predic- tions, tRNA gene predictions, expressed sequence clusters from the UniGene database",
+    "database, which aims to compile a non-redundant, curated data set representing current knowledge of known genes (Wheeler et al. , 2002; http://www.ncbi.nlm.nih. gov/entrez/query.fcgi?db=gene). Like the Ensembl protocol, many Acembly- predicted structures (the NCBI estimate 42 per cent) are incomplete. These struc- tures can be displayed alongside ab initio gene models, Ensembl-predicted genes, and matching UniGene clusters to allow users to make their own conclusions about the likeliest gene structure.",
+    "database, which aims to compile a non-redundant, curated data set representing current knowledge of known genes (Wheeler et al. , 2002; http://www.ncbi.nlm.nih. gov/entrez/query.fcgi?db=gene). Like the Ensembl protocol, many Acembly- predicted structures (the NCBI estimate 42 per cent) are incomplete. These struc- tures can be displayed alongside ab initio gene models, Ensembl-predicted genes, and matching UniGene clusters to allow users to make their own conclusions about the likeliest gene structure.",
+    "database, which aims to compile a non-redundant, curated data set representing current knowledge of known genes (Wheeler et al. , 2002; http://www.ncbi.nlm.nih. gov/entrez/query.fcgi?db=gene). Like the Ensembl protocol, many Acembly- predicted structures (the NCBI estimate 42 per cent) are incomplete. These struc- tures can be displayed alongside ab initio gene models, Ensembl-predicted genes, and matching UniGene clusters to allow users to make their own conclusions about the likeliest gene structure.",
+    "populations as Ensembl reects the progress of the International Haplotype Map Project (Thorisson et al. , 2005). More speculative data, such as GENSCAN-predicted exons that have not been incorporated into Ensembl-conrmed genes, may also be viewed. This means that the display can be used as a workbench for the user to develop personalized an- notation. For example, one may discover novel exons by nding GENSCAN exon predictions which coincide with good matches to a fragment of the draft mouse"
+  ],
+  "answer": "Ensembl is a genomic database that identifies the positions of known human genes from public sequence database entries, usually using tools like GENEWISE to predict their exon structures. It aims to provide a more accurate reflection of reality compared to ab initio predictions alone [1]. Ensembl also includes many genomic features other than predicted genes, such as different repeat classes, cytological bands, CpG island predictions, tRNA gene predictions, and expressed sequence clusters from the UniGene database [4].",
+  "question": "what is ensembl?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_18 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_18
new file mode 100644
index 0000000..f8b9214
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_18
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2012 - Identifying Gene Networks Underlying the Neurobiology of Ethanol and Alcoholism.pdf",
+    "2010 - Systems genetics, bioinformatics and eQTL mapping.pdf",
+    "2011 - Genetical genomics approaches for systems genetics.pdf",
+    "2012 - Functional genomics research in aquaculture principles and general approaches.pdf",
+    "2020 - A Multi-Omics Perspective of Quantitative Trait Loci in Precision Medicine.pdf",
+    "2014 - Identification of a QTL in Mus musculus for Alcohol Preference, Withdrawal, and Ap3m2 Expression Using Integrative Functional Genomics and Precision Genetics.pdf",
+    "2012 - Functional genomics research in aquaculture principles and general approaches.pdf",
+    "2020 - A platform for experimental precision medicine The extended BXD mouse family.pdf",
+    "2014 - Genetics of Gene Expression in CNS.pdf",
+    "2012 - Functional genomics research in aquaculture principles and general approaches.pdf"
+  ],
+  "extraction_id": [
+    "4253d9a7-5ade-5ac3-b37d-c27ed5a71ef6",
+    "298ee1f5-58a9-567c-86ba-8ac5967e1718",
+    "19febe84-f1fa-599d-84b4-95329b3d7f3f",
+    "a261cf24-3fe0-5cf1-ba6d-adf91794be38",
+    "65a5b8cd-6cf6-5c37-95b2-8677516d01e8",
+    "08ca6342-74ea-5196-b5b4-b46c9ec46713",
+    "0c3d0cb3-d4b0-5655-8b04-285a87710636",
+    "a797ba45-1fd5-58c5-af8f-e81341ecb7b2",
+    "7cea62b5-dbd8-5447-8126-9f2bcfe8b9eb",
+    "0c3d0cb3-d4b0-5655-8b04-285a87710636"
+  ],
+  "document_id": [
+    "c02542c0-eff8-5ec7-8f73-78f5d28d4226",
+    "27c922c6-e449-5f83-868a-3ad7284facc8",
+    "de78a01d-8d03-5afb-af5b-ce2ed2167766",
+    "a39b4cc1-8661-578b-a61b-b9962e45fc33",
+    "8503b166-b917-5efb-a356-5ba371504cc1",
+    "22ac294c-736f-5adb-8a0d-bd7166b578e8",
+    "a39b4cc1-8661-578b-a61b-b9962e45fc33",
+    "dd4994b9-9546-59c0-bc71-60e2617b6bcd",
+    "51a824c3-34c2-5be0-87a1-9f9f08b06e4a",
+    "a39b4cc1-8661-578b-a61b-b9962e45fc33"
+  ],
+  "id": [
+    "chatcmpl-ADZAyuRH6G3gbrSvS025ayW0rNJ7F",
+    "33fc773a-2ee3-572d-8629-2ec1e359aca1",
+    "68a13597-c223-54d9-9664-604d69b97c50",
+    "02a79024-51ee-5bdc-9a5b-ac26a6f3e40b",
+    "63febc09-7871-5cfc-9f7d-1f05eed65f41",
+    "1ad460e0-0a45-5f5e-9d37-d40bc8c65054",
+    "ee54bf38-d7c6-5e1f-bc75-8951d5ae917e",
+    "fa1981fe-6730-59a1-b331-c6c7250b0f2c",
+    "7924dd91-82c7-50d5-b663-0f5390f43065",
+    "ec03e315-1d99-5149-945c-5c4c0f4afed9",
+    "a5e77dc2-9d73-5d24-a446-0df546e34d85"
+  ],
+  "contexts": [
+    "traditional QTL mapping and GWASsapproaches can benefit from systems-biological approaches by filling in criticalinformation about the molecular phenotypes that stand between DNAvariation and complex disease (figure5). The incorporation of data fromhigh-throughput molecular profilingtechnologies, such as gene expressionmicroarrays, can better define a diseaseby identifying groups of genes thatrespond to or covary with disease-associated traits. Network analysis ofdisease-associated genes allows",
+    "knowledge of the true QTL location (Doss et al. 2005 ), which can be used to empirically estimate the power of aGWAS performed at a similar scale (Hao et al. 2008 ; Schadt et al. 2008 ). A GWAS on its own does little more than establish correlations between changes in DNA at agiven locus and changes in a disease trait of interest, with respect to populations of interest. Further, these studies on",
+    "genotypes. Since association studies allow for a mu ch finer mapping of the QTL  than that obtained with linkage analysis, there is a trade-off to consider between  power and resolution when choosing the mapping stra tegy. Genome-wide associa-  tion studies (GWAS) have naturally been used to per form genetical genomics  studies in humans [18, 24-27] and are emerging in m odel organisms studies using  outbred populations [28].   8.2.2  Combining studies",
+    "genetically also mapped to the same genomic location. In order to locate the positions of genes that are responsible for a certain trait, GWAS can be conducted. GWAS is a quan- titative approach to analyze the association of whole genome DNA polymorphisms and a phe- notypic trait, thereby localizing the genes un- derlining the trait. Genome-Wide Association Studies (GWAS) GWAS is a holistic whole-genome approach to robustly determine the association of DNA polymorphisms with correlated phenotypic",
+    "(PHMs) use principles of MR embedded within a Bayesian hierarchical model to detect interac-tions between regulatory elements [ 98]. Furthermore, GWAS is often integrated with the QTL analysis despite the fact that many GWAS loci are not strong eQTL loci [ 56]. GWAS-eQTL colocalization methods, including RTC [ 145], QTLMacth [ 158], Sherlock [ 159], and coloc [ 160], are based on the concept that disease-",
+    "association studies (GWAS) or linkage studies (Enoch 2013). QTL mapping studies historically had very low resolution,and many have been performed using populations for whichlimited genetic data exist. Publications of gene expressionstudies typically highlight a few interesting gene centered results, but the bulk of information is rejected due to concern",
+    "pairs that include many genes within the seg- ment. On the other hand, GWAS may point to several or even many genomic locations for the trait of interest, complicating further functional analysis. Analysis of Quantitative Trait Loci (QTL) QTL analysis reveals statistically signicant linkage between phenotypes and genotypes, thereby providing explanation for the genetic basis of variation in complex traits (Falconer and Mackay, 1996; Lynch and Walsh, 1998). In a sense, QTL analysis can be viewed as incom-",
+    "QTL mapping QTL mapping using GeneNetwork has been described in detail elsewhere ( Mulligan et al., 2017 ). However, in brief, quantitative trait loci (QTLs) are segments of the genome affecting a particular phenotype ( Falconer and Mackay, 1996 ). QTL mapping, identifying",
+    "3. Genetic Mapping Methods Several statistical approaches have been developed for genome-wide linkage analysis of  traditional phenotypes. The same approaches can be used to map eQTLs. These approaches  range from single marker tests ( t-test, ANOVA, and simple regression analysis) to multiple  locus mapping methods. The only major difference is that eQTL studies involve tens of  thousands of expression traits and require fast algorithms. Since an eQTL study tests for",
+    "plete GWAS analysis with limited number of markers that does not cover the entire genome. As such, if one or few QTLs are found, there may be more QTLs in the genome to be dis- covered. More importantly, in the absence of closely linked markers in the genomic regions containing signicant QTLs for the trait, the most signicant genes responsible for the trait can be missed. However, because of historical reasons such as the lack of genome-wide mark- ers, or the lack of funding, QTL analysis is still"
+  ],
+  "answer": "The primary difference between QTL mapping and GWAS lies in their approaches and resolutions for identifying genetic associations with traits.  1. **QTL Mapping**:    - **Approach**: QTL mapping involves identifying segments of the genome (quantitative trait loci) that are associated with a particular phenotype. It typically uses linkage analysis to find statistically significant linkages between phenotypes and genotypes [7], [8].    - **Resolution**: Historically, QTL mapping has had low resolution, meaning it identifies larger genomic regions associated with traits rather than pinpointing specific genetic variants [6].  2. **GWAS (Genome-Wide Association Studies)**:    - **Approach**: GWAS is a holistic, whole-genome approach that analyzes the association between DNA polymorphisms and phenotypic traits across the entire genome. It establishes correlations between changes in DNA at specific loci and changes in disease traits within populations [2], [4].    - **Resolution**: GWAS allows for much finer mapping of QTLs compared to traditional linkage analysis, providing higher resolution in identifying specific genetic variants associated with traits [3].  In summary, while QTL mapping identifies broader genomic regions linked to traits with lower resolution, GWAS provides a finer, more detailed mapping of genetic associations across the entire genome [2], [3], [4], [6], [7], [8].",
+  "question": "What is the difference between QTL mapping and GWAS?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_19 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_19
new file mode 100644
index 0000000..56cbe4f
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_19
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2005 -Dipetrillo- Bioinformatics toolbox QTL.pdf",
+    "2005 - Bioinformatics toolbox for narrowing rodent quantitative trait loci .pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2016 - Systems proteomics of liver mitochondria function.pdf",
+    "2020 - Gene network a completely updated tool for systems genetics analyses.pdf",
+    "2005 - quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2012 - Systems genetic analysis of the effects of iron deficiency in mouse brain.pdf",
+    "2005 -Broadkin- quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2020 - A Multi-Omics Perspective of Quantitative Trait Loci in Precision Medicine.pdf",
+    "2009 - Multiscale Genomic Analysis of the Corticolimbic System_ Uncoveri (1).pdf"
+  ],
+  "extraction_id": [
+    "e3adaae7-b5c1-5d35-9ba8-e082ccbb6fee",
+    "8311a931-a1b1-5228-bd9d-e9fcdd803ae9",
+    "6f6a41a6-61ef-5d73-8bce-5de9a9cc4798",
+    "1a46d28d-fc4a-5154-b887-3956d64959ef",
+    "2134720b-01d9-5e45-96bf-d1ff449d406d",
+    "c9fe8c31-86f9-5c59-8af8-8e81157cb99f",
+    "6ab990b0-4f9c-5be3-ab79-9ca6835271fa",
+    "cf8c1f06-37d2-5b82-bc22-6bc129759445",
+    "81b38987-de8d-5e87-8c09-72efc4cf781e",
+    "3296b30e-7dd3-576d-a2df-442406caa472"
+  ],
+  "document_id": [
+    "9b089457-5804-594a-99ea-e716b65c216c",
+    "5d87aefe-dee5-5f25-8b46-d87b24907dcc",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "9843107b-2d0c-5cbe-b55f-74e8090a575d",
+    "128224f1-3545-52c3-93cb-77c3cf4ec70a",
+    "0dc730ba-4ff4-52aa-a988-71075113c416",
+    "99fc80f0-f3c3-5766-a604-921552bb3298",
+    "e6027e7f-aec0-5e76-8aff-96b36389e701",
+    "8503b166-b917-5efb-a356-5ba371504cc1",
+    "3d0df5a3-7d7c-5edc-b94d-cae582f59c12"
+  ],
+  "id": [
+    "chatcmpl-ADZB337KFj8E0Gaou64p4qZLCzlv7",
+    "3aa13846-0d67-5481-8684-e414fdf712aa",
+    "a917decd-be31-53bd-9b1d-03eb6ba1e082",
+    "11bb5512-c251-5348-a53e-88ebaa9bcbbb",
+    "e4d4bbfc-d6b5-51cc-bf94-10bfc7099bf8",
+    "7d6a48a0-e046-520c-8434-7544e20b7a6c",
+    "38ba1c61-a6cc-504c-a6a3-da3c1252a17d",
+    "7ef9df1d-b21a-597a-9e74-6eace5d0c33c",
+    "fed778e8-dda5-5766-886d-287c24ff3758",
+    "c3d5e48d-ebf6-54b9-811e-6dbe2a171b77",
+    "bcb46fb1-dc0a-5f38-85e4-8c6d99809d74"
+  ],
+  "contexts": [
+    "candidate genes. These candidate genes must then betested for a causal link to the phenotype. A good starting point would be sequencing the cDNA of strong candidate genes to identify amino acid polymorphisms and testingfor mRNA and protein expression differences in target tissues of the original strains used to detect the QTL. Sequencing and expression studies will rene the list ofcandidate genes that can then be tested rigorously for proof of cause and effect. The nal proof of a causal gene",
+    "candidate genes. These candidate genes must then betested for a causal link to the phenotype. A good starting point would be sequencing the cDNA of strong candidate genes to identify amino acid polymorphisms and testingfor mRNA and protein expression differences in target tissues of the original strains used to detect the QTL. Sequencing and expression studies will rene the list ofcandidate genes that can then be tested rigorously for proof of cause and effect. The nal proof of a causal gene",
+    "do you identify the responsible gene within a QTL that you have identified?  Generally, one starts by performing a strain survey to find two parental inbred  strains that have a markedly different trait. One can now look up many different  traits of inbred mice online at the Mouse Phenome Database ( http://phenome. jax.org/pub-cgi/phenome/mpdcgi?rtn=docs/home ). However, the trait you may  want to study may not be present in wild type mice, so you may want to cross",
+    "used to test the hypothesis at locus-specific sig-nificance (LRS 12). In doing so, an additional 7 cQTLs are observed as consistent in both diets(Fig. 2I, red number). Solving QTLs: Finding the quantitative trait gene For cis-QTLs, the causal factors can be quickly identified: With few exceptions, they will be driv-en by variants within the gene itself or imme-diately adjacent. For trans-QTLs, mQTLs, and cQTLs, the identification of the causal quanti-",
+    "data is to find a quantitative trait locus, or QTL. A QTL  (http://gn1.genenetwork.org/glossary.html#Q ) is an area on a chromosome that can contain  one or many genes, that is linked to a change in phenotype. After a  QTL that is responsible for  the apparent variation  in phenotype  has been identified , one can  start stu dying the  genes  within that locus  to identify  the likely causal gene .     Once the data  is normalized appropriately  (in our case, no normalization was required) , the QTL",
+    "candidate genes that are expressed in tissues likely to inuence the traits of interest(Su et al 2004). These candidate genes are then sequenced in the two parental inbred strains looking for sequence dierences in coding or regulatory regions. After ne mapping the QTL interval and shortening the list of plausible candidate polymorphisms, the major challenge remains /C246 proving denitively which nucleotide polymorphism underlies the QTL. The most direct proof",
+    "because these strains have been genotyped at more than 14,000 markers, including single nucleotide polymorphisms (SNP). Hundreds of genes may lie within a QTL interval, so identifying the underlying genes requires complementary methods. One method is to use BXD gene expression data (a public resource at www.genenetwork.org) to screen for genes within the QTL interval whose expression correlates with the trait of interest [23].",
+    "candidate genes that are expressed in tissues likely to inuence the traits of interest(Su et al 2004). These candidate genes are then sequenced in the two parental inbred strains looking for sequence dierences in coding or regulatory regions. After ne mapping the QTL interval and shortening the list of plausible candidate polymorphisms, the major challenge remains /C246 proving denitively which nucleotide polymorphism underlies the QTL. The most direct proof",
+    "curate approaches to identify various types of QTL according to their molecular features, in par- ticular to control various confounding factors, such as dietary habit and population structure. Fine Mapping of Causal Variants and Causal Genes Despite the identi cation of large numbers of QTLs, it remains challenging to establish causal",
+    "to date, only a small handful of genes have been definitively identified for complex traits.  Our own efforts to identify a causal gene were stymied by the compound nature of QTLs  and the high gene density  in Qrr1 , and in Vol8a . Furthermore, it is now becoming clear  that in addition to the canonical candidate genes, there are multiple spliced variants,  microRNAs, and epigenetic factors to be considered.     With what appears to be an increasingly complex genom ic landscape, it is now all"
+  ],
+  "answer": "To determine which gene in your QTL is causal for the trait, you can follow these steps:  1. **Identify Candidate Genes**: Start by identifying candidate genes within the QTL interval. This can be done by looking at genes that are expressed in tissues likely to influence the traits of interest [6].  2. **Sequence Candidate Genes**: Sequence the cDNA of strong candidate genes to identify amino acid polymorphisms. This helps in detecting sequence differences in coding or regulatory regions between the two parental inbred strains [1], [6].  3. **Expression Studies**: Test for mRNA and protein expression differences in target tissues of the original strains used to detect the QTL. This will help refine the list of candidate genes [1].  4. **Fine Mapping**: Perform fine mapping of the QTL interval to shorten the list of plausible candidate polymorphisms [6].  5. **Correlation with Trait**: Use gene expression data to screen for genes within the QTL interval whose expression correlates with the trait of interest [7].  6. **Proof of Causality**: After refining the list of candidate genes, test them rigorously for proof of cause and effect. This involves proving definitively which nucleotide polymorphism underlies the QTL [1], [6].  By following these steps, you can systematically narrow down and identify the causal gene within your QTL.",
+  "question": "How do I determine which gene in my QTL is causal for the trait?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_2 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_2
new file mode 100644
index 0000000..46bb673
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_2
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2020 - Shared (epi)genomic background connecting neurodegenerative diseases.pdf",
+    "2018 - DNA methylation in the APOE genomic.pdf",
+    "2016 - Te-Mata-Ira-Genome-Research-Guidelines.pdf",
+    "2015 - Maternal diabetes, gestational diabetes and the role of epigenetics in their long term effects on offspring.pdf",
+    "2011 - Molecular Genomic Research Designs.pdf",
+    "2022 - Genomics and Epigenomics of Gestational Diabetes Mellitus Understanding the Molecular Pathways of the Disease Pathogenesis.pdf",
+    "2012 - Systems Biology Approaches to Nutrition.pdf",
+    "2012 - Aging, Rejuvenation, and Epigenetic.pdf",
+    "2008 - Genetic Effects on Environmental Vulnerability to Disease Novartis Foundation Symposium 293.pdf",
+    "2011 - EXPLOITING NATURAL AND INDUCED GENETIC VARIATION TO STUDY HEMATOPOIESIS.pdf"
+  ],
+  "extraction_id": [
+    "8963fcd1-8685-5518-9dd4-cb6d7075fe56",
+    "f8846e53-c9c0-5feb-8616-f2adcbf139eb",
+    "05ecf103-b037-5216-93f5-329714fc422c",
+    "746af210-6a0f-5814-80b6-8a3147246af2",
+    "66dfdd26-c34d-58b7-bc9b-fddd291c80c4",
+    "0072a2f8-0a81-5327-bfc9-24ed9886ef28",
+    "2f188d05-2160-5e55-b7b7-e18adebcfb12",
+    "9c1c1db0-57cf-5fae-bedd-f7fc61e8e6cb",
+    "eb19a2ea-02e9-5b7b-b493-2ed13c25a0e2",
+    "83da0679-fd33-562c-a3a7-1d7d4c5b79ed"
+  ],
+  "document_id": [
+    "3a7a3370-8de6-5d16-aac8-ba62336c7397",
+    "34b623d2-af48-5fc7-8e9f-e83b5f7a799a",
+    "86047c9b-e1f6-5c2d-b1d2-5becf4cb0957",
+    "3e92bd8e-fbf7-5bc4-9395-0a6dd0b0934e",
+    "ced08e27-8655-59a4-bf63-0ba746f139b7",
+    "f2353e3e-a250-5543-9906-d7d675c10eca",
+    "6955478b-950d-5d29-b24c-3a5ca656f3ae",
+    "bde26feb-f423-51b0-89ec-6f079bfc8b17",
+    "5d65e407-34e5-5c1c-b394-989b7a09b57d",
+    "6f250b15-61b3-57ed-8900-5aa4a173fa8c"
+  ],
+  "id": [
+    "chatcmpl-ADZ9QxtSzyI2BzaSwoHdkiSzca6zm",
+    "bc59df3b-f204-5bf4-8915-9d172cdc040f",
+    "bb94a5a9-2c25-5952-940d-05e102f2f8e5",
+    "9b4ddd27-ffbd-5c10-beae-e808c75e7fa5",
+    "8530798b-380a-5511-a61c-bcb75004a2f1",
+    "de68ac40-3950-53e5-b13e-7459026f02a9",
+    "d96d8aca-6024-5f5b-80bc-e1e018a8ceed",
+    "4e952f12-2c91-54fd-9662-4200ed92cad8",
+    "6030ef44-f93f-5637-8f09-2ab6cd06d180",
+    "1aacc908-4ed2-54ee-bb8f-5f8e000d4ae3",
+    "e3cf7319-1be5-5c01-b462-559ef450d72c"
+  ],
+  "contexts": [
+    "to regulate lifetime and aging processes. In fact, epigenetics modulate gene expression without altering the DNA sequence. This is possible by means of different kinds of epigenetic modifications, including DNA methylation and histone modifications (which might affect gene transcription), and noncoding (nc)RNAs (which might change gene expression at the post-transcriptional level)[59]. Given the crucial role of epigenetics in the modulation of gene expression, its alteration can contribute to",
+    "can regulate gene expression while the underlying DNA sequence remains the same. The epigenome is influenced both by underlying genetic variants as well as by environ- mental factors including the social environment, health behaviors, and environmental pollutants [ 11]. Methylation of CpG dinucleotides, the best understood epigenetic mechanism, is also dynamic over the life course. It is well established that epigenomic patterns of DNA methylation change with age [ 12]. A recent study in lymphocytes",
+    "Epigenetics Changes arising from alterations in gene expression  levels that are caused by reversible chemical  modification of DNA, but not changes to the DNA  sequence passed on from parents to offspring.",
+    "Epigenetic changes refer to heritable changes in gene expression which do not involve changes in DNA sequences. Several epigenetic mechanisms have been found to regulate gene expression. Whilst the most studied mechanism relates to DNA methylation, other changes, including histone modi cations and non-coding RNAs, also play an important role, and can be transmitted from one generation to the next. DNA methylation involves the addition of methyl groups to DNA, mainly at CpG sites, which converts cytosine",
+    "EPIGENETIC STUDIES   An epigenetic mechanism is a biochemical alteration to the DNA molecule that  does not change the sequence of the DNA but does in  uence gene expression.   Epigenetics   is often de  ned as the study of mitotically and/or meiotically heri- table changes in gene function that cannot be explained by changes in DNA sequence (Russo, Martienssen, & Riggs, 1996, p. 1).   The epigenetic/epigenomic approach shares many advantages and disad-",
+    "ity and expression of genes without changing their DNA sequence [ 4]. These modications are: DNA methylation, histone modications, and ncRNAs including miRNA [4]. The en- vironment and lifestyle can induce epigenetic changes, such as pollution, tobacco smoking, obesity, lack of physical activity, and alcohol consumption [ 108]. Furthermore, exposure to such environmental factors can have a buttery effect: epigenetic modications may",
+    "epigenetics is the study of mitotically heritable alterations in gene expression potential that are not caused by changes in DNA sequence (Jaenisch and Bird,  2003 ). Hence, rather than encompassing all of developmental biology, modern epigenetics is focused on understanding the spe-ci c molecular mechanisms that convey cellular memory.   Within the nucleus, the mammalian genome is wrapped",
+    "gene expression can also occur by  trans-epigenetics ( Bonasio et al., 2010 ), in which proteins and RNAs inuence gene expres-sion and repression. Stable transcription factor networks are anexample of trans -epigenetics ( Young, 2011 ). Clearly, enzymes that modify DNA and histones (methyltransferases, demethy-lases, acetyltransferases, deacetylases) are central epigeneticregulatory mechanisms ( Rando and Chang, 2009 ). The essence of epigenetics is not only the establishment, but",
+    "pay attention to epigenetic effects on gene expressionmeaning changes that are heritable but that do not involve any change in DNA sequence (see Rutter 2006). Three key points are relevant. First, genes only have effects when they are expressed. Many genes are expressed in only some body tissues and only at certain phases in development. Second, there are multiple inherited DNA elements that do not code for proteins but yet which have important effects through their in  uence on gene expression. We need to",
+    "genetics of gene expression (i.e. regular genetical genomics) and the genetics of epigenetics could be studied simultaneously, thus revealing genes that directly or indirectly affect epigenetic gene states. An additional issue that could be addressed by such an approach is to estimate the percentage of variation in gene expression that can be explained by different epigenetic conformations."
+  ],
+  "answer": "Epigenetics influences gene expression without changing the underlying DNA sequence through several mechanisms. These include:  1. **DNA Methylation**: This involves the addition of methyl groups to DNA, primarily at CpG sites, which can affect gene transcription by converting cytosine to 5-methylcytosine, thereby repressing gene activity [1], [4]. 2. **Histone Modifications**: Chemical modifications to histone proteins, such as methylation, acetylation, phosphorylation, and ubiquitination, can alter the chromatin structure, making genes more or less accessible for transcription [1], [6]. 3. **Non-coding RNAs (ncRNAs)**: These RNAs, including microRNAs (miRNAs), can regulate gene expression at the post-transcriptional level by binding to messenger RNAs (mRNAs) and preventing their translation into proteins [1], [6].  These epigenetic modifications are influenced by both genetic variants and environmental factors such as pollution, tobacco smoking, obesity, lack of physical activity, and alcohol consumption [2], [6]. These changes are heritable and can be passed from one generation to the next without altering the DNA sequence itself [4], [5].",
+  "question": "How does epigenetics inluence gene expression without changing the underlying DNA sequence?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_20 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_20
new file mode 100644
index 0000000..7055f47
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_20
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2007 - Quantitative genetics of age-related retinal degeneration a second F1 intercross between the AJ and C57BL6 strains.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2007 - Prenatal nicotine exposure alters gene expression in a sexually dimorphic manner.pdf",
+    "2009 - Experimental_Evolution.pdf",
+    "2009 - Garland_and_Rose_Experimental_Evolution.pdf",
+    "2009 - Experimental_Evolution.pdf",
+    "2009 - Garland_and_Rose_Experimental_Evolution.pdf",
+    "2022 - Genetic and genomic architecture in eight strains of the laboratory opossum.pdf",
+    "2012 - Needs Analysis of Genetics and Genomics in Communication Sciences and Disorders.pdf",
+    "2017 - Primer in Genetics and Genomics, Article 1 DNA, Genes, and Chromosomes.pdf"
+  ],
+  "extraction_id": [
+    "749877a1-0114-5bcd-8a5b-3b944012f5c9",
+    "34fa36d0-0b64-5c70-8645-ba3576d9262c",
+    "061d1490-4ce6-5f60-bdf8-15e8d863baf6",
+    "29e674a2-7ec9-5e00-9db3-308b112e439f",
+    "2f77d356-4cca-595c-912a-099efcc8b797",
+    "29e674a2-7ec9-5e00-9db3-308b112e439f",
+    "2f77d356-4cca-595c-912a-099efcc8b797",
+    "5afcc18d-5385-5d5e-8683-dd38f86131e7",
+    "10a507d1-60ca-5dae-9e49-4a6bace53668",
+    "89acea57-5c8a-55a6-90cf-ad11e5d527b6"
+  ],
+  "document_id": [
+    "f41cf6ad-273a-571a-866e-46b3dd407731",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "036efa18-a4b0-51bf-99d6-7c65193ccfed",
+    "34821353-1b74-5ee2-ac39-66dd46f145bf",
+    "496faa7f-9623-5ab7-9816-7c3755abb3aa",
+    "34821353-1b74-5ee2-ac39-66dd46f145bf",
+    "496faa7f-9623-5ab7-9816-7c3755abb3aa",
+    "f09eaa22-afb8-5bf7-90d3-4703056c18c5",
+    "c8a76cb1-506d-57e4-a18e-548e777898e2",
+    "b30c111b-1ca2-5f0a-93f3-862aa733fcad"
+  ],
+  "id": [
+    "chatcmpl-ADZBAENbLHFzwNSyDkvHF2ndPXSYM",
+    "45fd59f1-baa6-54b9-bfd6-9ba7ad122b86",
+    "e761426e-5f1d-5add-be86-bd6060d75ca7",
+    "748b07c1-c80f-5a4f-b295-9726493a698f",
+    "4e99669a-96cc-5269-a463-ff13337c56c3",
+    "9c00e371-7349-5ff0-8469-ffd95dd58e57",
+    "3cf13ae8-6c1c-5ddb-a719-81340d1c8ef6",
+    "27608ea2-c234-56f5-ad58-01fb67362130",
+    "c171e03f-4baf-5a0c-b961-401be867d691",
+    "e625cca4-7b62-5adf-b94e-1fdecc8e143c",
+    "03b1323c-d449-55fe-966e-d4925246b013"
+  ],
+  "contexts": [
+    "that accounts for the significant difference. One explanationis a contribution of the Y chromosome from the B strain. Sincethe cross was non-reciprocal all F2 mice carried the B strain Ychromosome. Thus, males carrying Chr X B QTL alleles andthe B Y chromosome differ in two ways from females carry-ing Chr X A alleles (or AB but B alleles are recessive) and noY chromosome, but in only one way from males carrying ChrX A/J QTL alleles because they share the B Y chromosome.However, pursuit of the identity of",
+    "women comprises 2 X chromosomes and in men 1 X and 1 Y chromosome (Figure 2). For each chromosome pair, 1 chro- mosome was inherited from the mother and 1 from the father. The full set of chromosomes is collectively called the genome.  The human genome is largely contained within the nucleus  of each cell, where it is separated from the rest of the cell functions. However, a small amount of DNA exists outside  the nucleus in the mitochondria and is considered to be part of  the human genome.",
+    "betweenmalesandfemalesisthesexchromosomes.MaleshaveanXYgenotypeand femaleshaveanXXgenotype.TheXisamuchlargerchromosome,165.5x106bpsvs. 16.0x106bps,withapproximately30timesmoregenesthantheYchromosome.To compensateforthelargernumberofgenes,andtoensurefemalesdonothaveover expressionofgenesresidingontheXchromosome,oneoftheXchromosomesis inactivated(7).TheXinactivationoccursearlyindevelopmentandisarandomprocess. Onlyasmallportionoftheinactivatedchromosomeretainstranscriptionalability.This",
+    "mammals. Instead of a dominant gene for maleness on the Y chromosome, it is the ratioof X chromosomes to autosomes that determines gender. The 2:2 ratio of XX femalesand the 1:2 ratio in XY males produce different ratios of regulatory proteins encoded byX-linked and autosomal genes. Those regulatory genes in turn cause transcripts of theregulatory Sex-lethal (Sxl) gene to be spliced differently in males and females, which be-",
+    "mammals. Instead of a dominant gene for maleness on the Y chromosome, it is the ratioof X chromosomes to autosomes that determines gender. The 2:2 ratio of XX femalesand the 1:2 ratio in XY males produce different ratios of regulatory proteins encoded byX-linked and autosomal genes. Those regulatory genes in turn cause transcripts of theregulatory Sex-lethal (Sxl) gene to be spliced differently in males and females, which be-",
+    "gins the process of sexual differentiation. A fly with two X chromosomes can thereforecarry a Y and still be a fertile female, leading to a paradoxical sex chromosome system inwhich males inherit X chromosomes from their fathers (figure 16.13).  Rice and Chippindale (2001) used a combination of these genetic techniques to test",
+    "gins the process of sexual differentiation. A fly with two X chromosomes can thereforecarry a Y and still be a fertile female, leading to a paradoxical sex chromosome system inwhich males inherit X chromosomes from their fathers (figure 16.13).  Rice and Chippindale (2001) used a combination of these genetic techniques to test",
+    "ity on the X chromosome compared to the other five strains(Figure 2B ). Compared to females, males had a deficiency of heterozygous X-linked SNP loci ( Supplementary Figure S2 ), which was expected because males are hemizygous. The resid-ual X-linked heterozygous SNPs in males could be due to mis-assembled autosomal contigs on the X chromosome, multiplecopies on the X, or homology between X and autosomalsequences. Chromosome XAutosomesProportion of SNP lociHomozygous SNPs Heterozygous SNPs",
+    "sex chromosome   Y chromosome:  One of the two sex chromosomes, X and Y. See also; X chromosome,  sex chromosome",
+    "one Y chromosome. Human chromosomes are typically displayed pictorially in a karyotype, as shown in Figure 9, arranged according to length and position of the centromere (i.e., the most con-stricted area of a chromosome). The ends of the chromosomesare called telomeres. Most human karyotypes look identicalbecause they are constructed from cells arrested in the phaseof the cell cycle when chromosomes are most condensed. During this phase of the cell cyc le, allelic differences cannot be detected."
+  ],
+  "answer": "The query appears to be based on a misunderstanding. Males do not have two Y chromosomes; they have one X and one Y chromosome, while females have two X chromosomes. This is clearly explained in the provided context:  - Males have an XY genotype, meaning they have one X chromosome and one Y chromosome [3]. - Females have an XX genotype, meaning they have two X chromosomes [3].  Additionally, it is noted that for each chromosome pair, one chromosome is inherited from the mother and one from the father [2]. Therefore, males inherit their Y chromosome from their father and their X chromosome from their mother, while females inherit one X chromosome from each parent.",
+  "question": "Why do males have two Y chromosomes and females only one?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_21 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_21
new file mode 100644
index 0000000..2e96cf2
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_21
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - eQTL analysis in mice and rats.pdf",
+    "2015 - Genetic Control of Survival and Weight Loss during Pneumonic Burk.pdf",
+    "2015 -Emery- Genetic Control of Survival and Weight Loss during Pneumonic Burk.pdf",
+    "2005 - quantitative-trait-analysis-in-the-investigation-of-function-and.pdf",
+    "2006 - From_gene_to_behavior_and_back_again_new.pdf",
+    "2005 - Gene Expression Differences in Mice.pdf",
+    "2008 - Using gene expression databases for classical trait QTL candidate gene discovery in the BXD recombinant inbred genetic reference population Mouse forebrain weight.pdf",
+    "2005 - quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2005 -Broadkin- quantitative-trait-locus-analysis-of-aggressive-behaviours-in-mi.pdf",
+    "2005 -Knott- Regression based QTL mapping.pdf"
+  ],
+  "extraction_id": [
+    "71981bfb-284e-50ad-854e-2055c07f77a7",
+    "615ee0cd-5960-57e5-b4e6-56e4b8020a1b",
+    "268a23e8-f528-5b59-89f2-188331e0a03c",
+    "0a895880-91c0-5079-b258-73926b38430f",
+    "64c0287d-aeea-52eb-a074-e9591c5593ae",
+    "2ee9945a-e33c-5303-84f6-6bb4fec529ea",
+    "dbf6a85f-6ae5-54da-87e4-8c2c70c2b37d",
+    "9de93371-6239-53c2-b42c-71f615a0614b",
+    "0a5c759e-8dab-55f1-ac59-e8211ec683b8",
+    "a4a2e963-3b9b-576e-885a-d5e757a6ce8c"
+  ],
+  "document_id": [
+    "8d67ea90-f7b1-5bb8-937c-4a9eceddff43",
+    "ae1025b0-1410-51ae-9be2-26fa2e9d5808",
+    "a9aceace-bf48-5472-b54c-59a458a84c62",
+    "dac1c73c-0b5f-5a54-bb12-7e8b654009c0",
+    "7a088b36-11b7-5379-bfe5-ce571e11de07",
+    "47abbcce-503c-552f-a02e-bf2f31fd1d8a",
+    "d2dc6644-2feb-5d2b-8ec7-436fc9e449b6",
+    "0dc730ba-4ff4-52aa-a988-71075113c416",
+    "e6027e7f-aec0-5e76-8aff-96b36389e701",
+    "cd41c63b-e5c2-5040-bbc5-ab20925b7d17"
+  ],
+  "id": [
+    "chatcmpl-ADZBER3gC3GniJPKr4d0S0Jc8x850",
+    "73540700-b5cf-5838-852b-b281ca086140",
+    "374c456a-d1db-5b4a-8713-97abe4162d77",
+    "b9d52798-0235-5018-bccd-560565d16cc3",
+    "b660d882-1cb0-5150-ae76-8eb3ccb88a58",
+    "fef212bc-631b-591d-b8e3-d1523da0507d",
+    "60643722-3d4e-571c-97e9-3b5c67670ca0",
+    "e9424ae3-c15b-5b96-aa5f-fe0865f4b2fd",
+    "c8f17022-aeae-5242-9082-d6d1eee4c4bf",
+    "1b2de424-be9f-572d-bd62-dc2ecd92192b",
+    "1c584e4b-db8b-5f00-ad8b-d43702b65f22"
+  ],
+  "contexts": [
+    "While most of the Y chromosome does not undergo recombination, the recombination rate of the X chromosomeis slower than that of the autosomes. This has important consequences on the detection of significant QTLs. For a comprehensive view of these issues, see(43). 9.Probe hybridization artifacts When several probes are available for the same gene, it is not uncommon to observe a difference in the mapping results",
+    "8 QTL Mapping     Allelic variation exists among natural populations and inbred strains, and this is  reflective of the segregation of quantitative tr ait loci (QTLs) [96]. QTLs are stretches of  DNA that are closely linked to genes that underlie a phenotype of interest. QTL analysis has been proven to be an invaluable tool to  help unravel heritable traits, by enabling  researchers to map different quantitative traits back to the genomic location involved in the regulation of these phenotypes.",
+    "8 QTL Mapping     Allelic variation exists among natural populations and inbred strains, and this is  reflective of the segregation of quantitative tr ait loci (QTLs) [96]. QTLs are stretches of  DNA that are closely linked to genes that underlie a phenotype of interest. QTL analysis has been proven to be an invaluable tool to  help unravel heritable traits, by enabling  researchers to map different quantitative traits back to the genomic location involved in the regulation of these phenotypes.",
+    "genetic background. Gene identification of QTL should be distinguished from identification of the quanti- tative trait nucleotide (QTN). The latter is a daunting task, since SNPs are so frequent. Final proof for a QTN in mice would require placing a genomic segment containing theputative QTN from a donor mouse strain on the background of another strain using homologous recombination and reproducing the phenotype of the donor strain.",
+    "The basic  pr emise  of QTL  an alysis  is simple  (Ph illips  and Belknap, 2002 ) . First,  one must  meas  ure a speci  c phen  otype  within  a popul  ation. Next, the population must be genotyped at a hundred or more marker loci186 Boehm II et al.",
+    "verify the difference, and the data were then ana-lyzed by the QTL detection method of Belknap et al.(1997) based on allele frequency differences betweenthe two lines. When a difference was confirmed,individual genotypes and individual behavioral re-sponses to MA were used to estimate the position ofthe bQTL using the interval mapping methods as implemented in R/qtl (Broman et al. 2003). The lat-",
+    "X axis depicts 19 autosomes and X chromoso me. The Y axis is the likelihood ratio statistic from a single QTL model. Two  QTLs, on chromosomes 1 and 11, are significant at a mu ltiple test corrected permut ation threshold as shown. Chromosome 1 and 11 likeli hood ratio statistic plots Figure 2 Chromosome 1 and 11 likelih ood ratio statistic plots . Interval mapping plots of chromosomes 1 and 11, showing more  detail of Figure 1. 2 LOD support inte rvals are shown in Mb on the X axis.",
+    "genes underlying QTLs in animals and plants (see for example Shirley et al 2004,Korstanje & Paigen 2002, Fridman et al 2004). I should also point out, though, that even in a single QTL region isolated in a congenic strain, it is possible that there is more than one allele that aects the phenotype. So, you have a fair pointabout the challenges and complexities of QTL analysis. Koolhaas: There are dierent questions underlying both approaches. The QTL",
+    "genes underlying QTLs in animals and plants (see for example Shirley et al 2004,Korstanje & Paigen 2002, Fridman et al 2004). I should also point out, though, that even in a single QTL region isolated in a congenic strain, it is possible that there is more than one allele that aects the phenotype. So, you have a fair pointabout the challenges and complexities of QTL analysis. Koolhaas: There are dierent questions underlying both approaches. The QTL",
+    "model at the QTL assumes that the original lines arexed for different alleles although genes can besegregating elsewhere. Hence, it is possible to combineinformation about the QTL across families. The assumption of xation at the QTL can be tested by"
+  ],
+  "answer": "To differentiate between X and Y DNA with respect to DNA tracing and determining quantitative trait loci (QTLs), consider the following points:  1. **Recombination Rates**:    - The Y chromosome largely does not undergo recombination, which means it is passed down relatively unchanged from father to son. This makes it useful for tracing paternal lineage [1].    - The X chromosome, on the other hand, does undergo recombination, but at a slower rate compared to autosomes. This slower recombination rate has significant implications for the detection of QTLs on the X chromosome [1].  2. **QTL Mapping**:    - QTLs are stretches of DNA closely linked to genes that underlie specific phenotypes. The process of QTL mapping involves identifying these regions to understand the genetic basis of heritable traits [2].    - The detection of QTLs on the X chromosome may be influenced by its slower recombination rate, which can affect the resolution and accuracy of QTL mapping compared to autosomes [1].  In summary, the primary differences between X and Y DNA in the context of DNA tracing and QTL determination lie in their recombination behaviors and how these behaviors impact the detection and mapping of QTLs. The Y chromosome's lack of recombination makes it a stable marker for paternal lineage tracing, while the X chromosome's slower recombination rate affects the detection of significant QTLs [1], [2].",
+  "question": "How does one tell the difference between X and Y DNA, with respect to DNA tracing and determining QTLs"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_22 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_22
new file mode 100644
index 0000000..abd1498
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_22
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2008 -  Study Design and Statistical Issues.pdf",
+    "2009 - Opening Up the Conversation on Genetics.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2008 - Genotype-phenotype relationships and the patterning of complex traits as exemplified in the mammalian dentition.pdf",
+    "2019 - Sexual Dimorphism in the Age of Genomics How, When, Where.pdf",
+    "2007 - Promoting_Student_Scientific_Literacy_of_Molecular Genetics and Genomics.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2019 - The influence of paternal diet on sncRNA-mediated epigenetic.pdf"
+  ],
+  "extraction_id": [
+    "06bf0605-388a-592c-96ad-3a53bb36362c",
+    "51dbd5e2-fde6-5097-aa05-fcf57d3ca6b1",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "8a1ce8fa-b5f4-5942-b7b1-14a8a7887710",
+    "5aab3e60-b8b0-52ad-b4d3-817cf012cfa5",
+    "e22bb6fb-bec4-5c4c-8690-c96d0b8d13d4",
+    "67369433-749b-5d6a-b5ef-3f0afe78b767",
+    "206b8810-b7c1-5195-a10f-4e291864b77c",
+    "84335575-34d7-56b6-aa06-5a8ac13d637a"
+  ],
+  "document_id": [
+    "c3bd9cf0-f768-55c4-be94-96590d7acc21",
+    "b62a8f54-c2f5-5bbb-9324-af80f7537167",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "f6e866b8-b233-5862-bfb8-9949d0dabb97",
+    "3f8c03b0-4235-5774-9d26-e43d55c1001b",
+    "755f34c4-cc06-5275-a744-16d48162b012",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "dfcbd6e6-f60d-5eb7-867b-34ec78415e82"
+  ],
+  "id": [
+    "chatcmpl-ADZBKlm7c8ZE1GCPmwcK7X0HEcljf",
+    "1e151ad5-59d9-598d-97ba-90ba0e64c4cb",
+    "60ad1512-b0c0-59cd-ace4-c146e2c04b52",
+    "a66b8b00-d51c-575b-b6ac-fa445c4ca715",
+    "df4c6108-740d-5bcf-99e6-dbda74f7e41a",
+    "4472740a-d22d-5bb1-98e3-e91332cbb303",
+    "47b9142f-98a3-5a45-8eaa-d327c9cc055d",
+    "0158f264-120f-5942-ad55-ef5fde1f188a",
+    "8e3fdc2c-0962-5854-83e7-a60ab05cf6de",
+    "6c8dfaa1-a96f-5f1c-8b5a-870acfd46f5f",
+    "be93ee68-72ae-5015-a3f0-19e7bf24827a"
+  ],
+  "contexts": [
+    "phenomena such as mutations and gene conversion events) occur in relevant meioses  leading up to the formation of the gametes (i.e., egg and sperm) which are combined  during fertilization and the formation of zygotes. Thus, individuals inherit a patch- work of chromosomal segments from maternal and paternal chromosomes.",
+    "the egg and the sperm. Such a process would result in genetic changes that will be copied into every cell of the future adult, including reproductive cells (Stock & Campbell, 2000), opening the door to irreversibly alter the human species. Inevitably, signifi  cant self-disclosure and discussion challenges await families",
+    "a fertilized egg is a complicated process that relies on controlling: which genes are active; whenthese genes activate; and for how long they are active. In broad terms, there are four ways that thiscontrol can be achieved: First, inside the sperm or egg, genes can be marked with small chemical tags that flag these genes",
+    "to be activated (or remain inactive) after fertilization, depending on whether the modification wasmade by the father (in the sperm) or the mother (in the egg); this process is known as imprinting. Second, the mother can alter the gene activity in her offspring via the placenta; this process is known as maternal effect. Third, instructions encoded within the embryos DNA can directly control if, andwhen, a nearby gene becomes activated; this is known as  cis-regulation. Finally, similar instructions",
+    "(Figures 8 and 9). Two gametes (egg and sperm) ultimately  join into a single cell, the zygote, which has the full comple-ment of 23 chromosome pairs restored. If all goes well, the zygote gives rise to a live offspring. The Mendel Laws: Segregation and Independent  Assortment Both of the Mendel laws pertain directly to the process of  meiosis. The first Mendel law, the law of segregation, states  that each parent passes a randomly selected allele for a given",
+    "the subset of that genetic information that is active.  But how does the differentiation process  begin?  The key insight in resolving this conundrum came from fly genetics and was the  realization that the egg is not a homogenous sack of protoplasm.  The maternally-derived genes  active in the fertilized egg are asymmetrically distributed such that at the first cell division each  daughter cell receives a different complement of factors.  Development continues as a",
+    "sex chromosome effects. (B)Soon after fertilization, male and female cells have sex-specic transcriptomes, epigenomes, and phenotypes (for example, male embryos grow faster than female embryos). At implantation, lineage determination begins and gene expression differences are reduced. Epigenetic marks, however, are less constrained and some are maintained, affecting gene expression, and phenotype later in development. Once specic lineages are established, differences in",
+    "genes.  An altered gene may be passed on to every cell that develops from it.  The resulting features my help, harm, or have little or no effect on the offsprings success in its environment. (AAAS, pg. 109, 5B:9-12#4 ) 6. Heritable material: The information passed from parents to offspring is coded in DNA molecules (AAAS, pg 108, 5B:9-12#3) 7. Mutagens: Gene mutations can be caused by such things as radiation and chemicals.  When they occur in sex cells, the mutations can be passed onto offspring; if they",
+    "or father (sperm cell). Each gamete has a set of 23 unpaired chromosomes. Two human gametes (egg and sperm) combine  to create a cell (zygote) that contains the full human genome  of 23 paired chromosomes.Genetic Information Nondiscrimination Act (GINA) US federal legislation that makes it unlawful to discriminate  against individuals on the basis of their genetic profiles in regard to health insurance and employment. These protections  are intended to encourage Americans to take advantage of",
+    "spermatozoa: more than the sum of its parts? DNA, histones, pro - tamines and epigenetics. Reproduction 139:287301 Nilsson EE, Sadler-Riggleman I, Skinner MK (2018) Environmentally  induced epigenetic transgenerational inheritance of disease. Envi-ron Epigenet 4:dvy016Pembrey M, Saffery R, Bygren LO, Network in Epigenetic Epide-"
+  ],
+  "answer": "Once a sperm combines with an egg, the traits passed onto the resulting lifeform are determined by several factors:  1. **Chromosomal Segments from Parents**: Individuals inherit a patchwork of chromosomal segments from both maternal and paternal chromosomes [1]. 2. **Mendelian Laws**: The Mendel laws of segregation and independent assortment play a crucial role. The law of segregation states that each parent passes a randomly selected allele for a given trait to their offspring [5]. 3. **Gene Activation and Regulation**: The process involves controlling which genes are active, when they activate, and for how long. This control can be achieved through mechanisms such as imprinting, maternal effect, cis-regulation, and other instructions encoded within the embryo's DNA [3], [4]. 4. **Epigenetic Marks**: Epigenetic marks, which are chemical modifications to DNA or histones, can affect gene expression and phenotype later in development [7].  These combined genetic and epigenetic factors determine how traits are passed onto and expressed in the resulting lifeform.",
+  "question": "Once a sperm combines with an egg, what determines how traits are  passed onto the resulting lifeform?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_23 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_23
new file mode 100644
index 0000000..f1c323c
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_23
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2020 - Gene network a completely updated tool for systems genetics analyses.pdf",
+    "2012 - Using Genome-Wide Expression Profiling to Define Gene Networks Relevant to the Study of Complex Traits From RNA Integrity to Network Topology.pdf",
+    "2020 - Gene network a completely updated tool for systems genetics analyses.pdf",
+    "2020 - Phylogenetic tree building.pdf",
+    "2011 - Peroxisomal L-bifunctional enzyme (Ehhadh) is essential for the production of medium-chain dicarboxylic acids.pdf",
+    "2018 - Invited review Genetic and genomic_ xmltexbreak_ mouse models for livestock research.pdf",
+    "2013 - Pathogenesis and reversal of liver fibrosis Effects of genes and environment.pdf",
+    "2022 - Systems genetics in the rat HXBBXH family identifies Tti2 as a pleiotropic quantitative trait gene for adult hippocampal neurogenesis and serum glucose.pdf",
+    "2022 -Senko- System Genetics in the Rat HXB\uf022BXH Family.pdf",
+    "2022 -Senko- Hippocampal neurogenesis serum glucose.pdf"
+  ],
+  "extraction_id": [
+    "858f630f-9443-5f13-ac40-8e16eadd9ba1",
+    "3e0c2a06-e6de-5888-a360-a2c483d9f744",
+    "f7e3761d-1baa-573a-9cbd-4070a400c42e",
+    "e697c9f2-c175-5e85-9a7a-03bf5ef921b7",
+    "a8b40857-7ae8-512a-9817-bea1ae3345ba",
+    "6983f2dd-b440-5696-92a1-84f4c332834b",
+    "71f1aefb-6e32-5add-804e-6fbaa39ca720",
+    "c6be3dd3-f076-54bd-b1fb-04678962a817",
+    "065807db-909e-5654-8dd7-5652f07be29c",
+    "7fe8fbaa-b950-5553-940b-317f306efb4b"
+  ],
+  "document_id": [
+    "128224f1-3545-52c3-93cb-77c3cf4ec70a",
+    "1eb6f5b7-a3bc-5455-91f0-6f2eb37be861",
+    "128224f1-3545-52c3-93cb-77c3cf4ec70a",
+    "7310a5bc-2fc1-5fa0-ad1d-f3411830062b",
+    "ac61753e-bcb2-55c3-804b-e821e3d1a4ad",
+    "5b167564-85a2-5886-b800-37932c3143a9",
+    "73724c99-98df-53b2-a378-29c8b4faa171",
+    "e6323aba-6fec-500b-99e3-a41c2e7f17ff",
+    "c67a6829-954a-5202-85fb-7524b03fab28",
+    "bac2ab98-4317-59ed-99ef-deda8c22786d"
+  ],
+  "id": [
+    "chatcmpl-ADZBPsS9vmK32TYjsc1YUUJCtSSmn",
+    "c63cfaee-749e-547b-9c0a-086266f10670",
+    "312eae52-ede7-5c13-8974-fce0126426cf",
+    "2ae780e5-9549-50c0-a260-d7ef774f7956",
+    "6a443d81-33ed-524c-9f11-318f1013a214",
+    "8b8a24da-a175-5cb8-91bd-8966fca5d344",
+    "d5c42ccf-569f-5a50-bd49-6b45097a3d00",
+    "f5c218f0-1280-55f8-912b-b32b833e93a3",
+    "ac4f8148-e6da-5d16-9e61-3a1aff2f9c81",
+    "69a916ea-abe7-5637-81fe-ee5e38f9c68b",
+    "03d0618c-8ed8-5984-a4eb-e743daf4f1a7"
+  ],
+  "contexts": [
+    "GeneNetwork have reinvigorated it, including the addition  of data from  10 species, multi -omics  analysis, updated code, and new tools. The new GeneNetwork is now an exciting resource for  predictive medicine and systems genetics, which is constantly being maintained and improved.    Here, we give a brief overview of the process  for carrying out some of the most common  functions on GeneNetwork, as a gateway to deeper analyses , demonstrating how a small",
+    "addition to this, GeneNetwork can be used to study correlations between traits and to perform data mining in genomic regions containing candidates for quantitative trait genes (Hoffman et al., 2011). All datasets in GeneNetwork are linked to a materials and methods information page that summarizes experimental details relating to the dataset. Databases within GeneNetwork include the transcriptome database, the BXD published",
+    "publication, and links to the dataset database and to the published paper (4C). There is also an option  to add this trait to your collection by pressing the Add button (4D), or to view this trait in an ear lier  version of GeneNetwork, GN1 (4E).",
+    "Bayesian inference of species networks from multilocus sequence data. Mol. Biol. Evol. 35, 504517 (2018). 167. Flouri, T ., Jiao, X., Rannala, B. & Yang, Z. A Bayesian  implementation of the multispecies coalescent   model with introgression for phylogenomic analysis. Mol. Biol. Evol. 37, 12111223 (2020). 168. Kubatko, L. in Handbook of Statistical Genomics    (eds Balding, D., Moltke, I. & Marioni, J.) 219245  (Wiley, 2019). 169. Rannala, B., Edwards, S., Leach, A. D. & Yang, Z.",
+    "subnetworks   GeneNetwork (www.genenetwork.org) is a depository of data- sets and tools for use in complex systems biology approaches in  order to generate or predict higher order gene function ( 23, 24 ).",
+    "on different cross types, such as F 2crosses (B6BTBRF2, B6D2F2, BH/HB F2, CastB6/B6Cast F2, B6JxB6N F2), butalso on more complex outbred crosses such as the HS, the CC, and the Hybrid Mouse Diversity Panel. Recently, data from other species has also been integrated into GeneNet- work (human, rat, monkey, fruit ies, and others) to facilitate the translational research of results into other species. To this end, GeneNetwork provides many tools for the analysis of",
+    "GeneNetwork (www.genenetwork.org). The web -based software further allows extraction of sets of",
+    "Phenotypes Database attheGeneNetwork (www.",
+    "Phenotypes Database attheGeneNetwork (www.",
+    "Phenotypes Database attheGeneNetwork (www."
+  ],
+  "answer": "The provided context does not include specific instructions on how to add a new species to the GeneNetwork database. However, it does mention that data from multiple species have been integrated into GeneNetwork [1], [6]. This suggests that the process is possible, but the exact steps are not detailed in the given excerpts.  To add a new species to the GeneNetwork database, you may need to refer to the official GeneNetwork website (www.genenetwork.org) [5], [7], or contact their support team for detailed guidance. Additionally, reviewing the materials and methods information page linked to datasets in GeneNetwork [2] might provide further insights into the process.",
+  "question": "How can I add a new species to the GeneNetwork database?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_24 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_24
new file mode 100644
index 0000000..e6e78ff
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_24
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2014 - The genetic basis of obesity-associated type 2 diabetes (diabesity) in polygenic mouse models.pdf",
+    "2006 - Quantitative Trait Loci on Chromosome 8q24.pdf",
+    "2017 - Genomic regulation of type 2 diabetes endophenotypes Contribution.pdf",
+    "2015 - A Chromosome 13 locus is associated with male-specific mortality in mice.pdf",
+    "2008 - Meta-Analysis Approach identifies Candidate Genes and associated Molecular Networks for Type-2 Diabetes Mellitus.pdf",
+    "2016 - The genetic architecture of type 2 diabetes.pdf",
+    "1998 - Genetic dissection of ``OLETF_, a rat model for non-insulin-dependent diabetes mellitus.pdf",
+    "2015 - Transcript Expression Data from Human.pdf",
+    "2004 - Interaction and Association Analysis of a Type 1 Diabetes Susceptibility Locus.pdf",
+    "2001 - Genetic Analysis of a New Mouse Model for Non-InsulinDependent Diabetes.pdf"
+  ],
+  "extraction_id": [
+    "1ab308e3-565f-5d14-86bc-2909dd9a1de0",
+    "d35d2e8c-0e2f-5be4-a902-18d5c857746d",
+    "9dfc060c-bf5e-5958-b446-cfc12a4f85c5",
+    "cc39ccbe-150c-5d7e-8b6b-f6c98738cb95",
+    "309adb8f-fa42-5806-9e50-95742ba90857",
+    "8b8b572d-68f5-5470-b5ed-ec5c6219dd5e",
+    "c29fe565-1167-5821-8715-559cb48f2090",
+    "b9d039d0-8982-52c6-ba45-be2e2eeda7d5",
+    "b7586c99-af71-5f11-8fed-fd8395c783b6",
+    "4cc0bd43-c6a8-55fb-8300-d2228636c89d"
+  ],
+  "document_id": [
+    "1459a93f-3052-5cea-ba83-caf266ef9b86",
+    "8c5ffeac-5108-5b03-acd0-57aa09469af5",
+    "fef1ae33-b3af-50ea-909c-f1b57f7fe981",
+    "ad8f2626-87fb-520e-8cef-ee9a9cc3ab0b",
+    "4060609b-1464-55fa-93cd-fefaf2cac900",
+    "d7e2a9de-46f1-5191-9cb0-dd68eb9f365a",
+    "0f04bb9f-6d45-5511-a05c-a09f8ee9a5e9",
+    "2b30d4f3-9ec3-574f-9a36-709b0e09c3f2",
+    "4246f8d0-69e8-56cf-9674-d379467dfb61",
+    "c6086f32-0a3a-5a92-9e5b-4d2fa7fbbc93"
+  ],
+  "id": [
+    "chatcmpl-ADZBUB9zekyDDARKA9rzHsVGglrzJ",
+    "313e590c-40a4-5adb-a2d8-18577f465b30",
+    "f5220a71-d1bc-50ae-933a-2b92bab0c4ae",
+    "569bb9be-0b57-535a-ab0c-206d85f1dd4a",
+    "f3de711d-7dff-5b13-89c1-720bb6be9e12",
+    "6299defb-19e0-5f6d-aaea-44b36cdece6e",
+    "807bf364-408f-50c9-bacd-b9da438a1703",
+    "410c1b39-1d2a-5954-ac2c-9bf4ad38aa58",
+    "0ea7a0f3-5fdd-5d9f-8f53-4620492867f7",
+    "517a8a37-697b-500b-a5e8-7eff80fc0f79",
+    "af834bd3-8462-5159-99e8-59a2fc1f09c9"
+  ],
+  "contexts": [
+    "genes that are responsible for obesity-associated diabetes. By the generation of subcongenic lines of a QTL, if pos- sible starting with chromosome substitution strains, thensmall critical regions that harbor the gene(s) in question can be identied with certainty. Sequence analysis and mRNA proling together with gene targeting in-vitro andin-vivo may lead to a solid chain of evidence linking sequence differences with altered molecular, cellular, and",
+    "tensive nondiabetic families, the QTLs on chromosomes 8q24 and 7q11, which are located in regions previouslyidentied as harboring type 2 diabetesassociated genes,may govern insulin sensitivity and insulin secretion in thepresence of insulin resistance before development of overttype 2 diabetes. Follow-up ne-scale mapping aroundthese loci and well-designed candidate gene studies, inparticular, are strongly encouraged. ACKNOWLEDGMENTS",
+    "studies used the QTL approach for statistical analysis of genotypes and phenotypes measured in the crosses. The concept of genetic dissection of diabetes into quantitative endophenotypes was introduced and resulted in the detection of genetic loci responsible for the control of fasting glycemia [39,42] , fasting insulinemia [39,43] , glucose tolerance [39,41,42] , insulin secretion induced by glucose or arginine [39], body weight [39,41,44] , adiposity [39], b-",
+    "indicating that risk factors exist on both genetic back- grounds [ 29]. QTL mapping studies indicate that these murine metabolic traits have a complex genetic architec- ture that is not dominated by any single allele [ 2931], much like humans [ 32,33]. Prior work identied candidate genes on Chr 13 that might underlie diabetes-related traits, including RASA1, Nnt, andPSK1. RASA1 show strong sequence differences between B6 and D2 strains [ 34]. Rasche et al. [ 35] reported that",
+    "genetic background [4]. Linkage analyses have shown that several quantitative trait loci interact with each other and with the environment to elicit obesity syndromes that are potentially diabetic. Several recent genome-wide associa- tion studies have identified novel candidate genes for T2DM but the effect of these variants on disease suscepti- bility is generally low, with odds ratios mostly around 1.5 [5-11]. Multiple studies on the transcriptome level have been per-",
+    "(2011). 7. Steinthorsdottir, V. et al. Identification of low-frequency and rare sequence variants associated with elevated or reduced risk of type 2 diabetes. Nat. Genet.   46, 294298 (2014).8. Ma, R. C. et al. Genome-wide association study in a Chinese population  identifies a susceptibility locus for type 2 diabetes at 7q32 near PAX4.  Diabetologia 56, 12911305 (2013). 9. Huyghe, J. R. et al. Exome array analysis identifies new loci and low-frequency",
+    "nificant QTL, strongly associated with body weight (Galli et al.1996; Gauguier et al. 1996). Moreover, Gauguier and colleagues(1996) mapped a QTL linked to postprandial insulin secretion intheregionofChr4wherewedetectedasuggestiveQTL.DifferentNIDDM models (obese OLETF rats and lean GK rats) may carryalleles conferring NIDDM susceptibility in the same genes. Thecombined results imply the possibility of common genetic factorsunderlyingNIDDMinhumans,notwithstandingthehighdegreeofgenetic heterogeneity in human",
+    "data indicates that variants regulating islet gene transcription influence type 2 diabetes(T2D) predisposition and glucose homeostasis. However, the specific genes through whichthese regulatory variants act remain poorly characterized. We generated expression quanti-tative trait locus (eQTL) data in 118 human islet samples using RNA-sequencing and high-density genotyping. We identified fourteen loci at which cis-exon-eQTL signals overlapped",
+    "linkage analysis assists in the identication of possiblegene-gene interactions and that 5q11-q13 and 7q32together constitute a signicant susceptibility factorfor type 1 diabetes. Diabetes 53:15841591, 2004Type 1 diabetes is a common multifactorial dis- ease characterized by autoimmune destructionof the insulin-producing /H9252-cells in the endocrine pancreas, resulting in deranged metabolic ho-",
+    "model for common forms of NIDDM in humans associ-ated with obesity. This study identies the location of amajor QTL and additional independent QTLs contrib-uting to development of hyperglycemia in TH malemice. We have also elucidated genegene interactionsbetween QTLs in the development of NIDDM, detect-ing new QTLs that reveal their signicant effects onlywhen they interact with other QTLs. This complexinheritance pattern associated with genegene inter-actions may be of prime importance in"
+  ],
+  "answer": "In QTL (Quantitative Trait Loci) analyses, several genes have been identified as being associated with diabetes. Specifically:  1. **RASA1, Nnt, and PSK1** on chromosome 13 have been identified as candidate genes underlying diabetes-related traits [4]. 2. **PAX4** near the locus 7q32 has been identified in a genome-wide association study in a Chinese population as a susceptibility locus for type 2 diabetes [6]. 3. **Genes on chromosomes 8q24 and 7q11** have been associated with insulin sensitivity and insulin secretion, which are critical factors in the development of type 2 diabetes [2]. 4. **Genes on 5q11-q13 and 7q32** have been identified as significant susceptibility factors for type 1 diabetes [9].  These genes and loci are typically associated with diabetes in QTL analyses, highlighting the complex genetic architecture of the disease.",
+  "question": "which genes are typically associated with diabetes in QTL analyses?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_25 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_25
new file mode 100644
index 0000000..e1d1cea
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_25
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2010 - Candidate Gene and Genome-Wide Association Studies in Behavioral Medicine.pdf",
+    "2009 - Pathomechanisms of Type 2 Diabetes Genes.pdf",
+    "2014  - Dorothy Hodgkin Lecture 2014 Understanding genes identified by genome\u2010wide association.pdf",
+    "2015 - Type 2 Diabetes Mellitus and the Association of Candidate Genes.pdf",
+    "2012 - Type 2 Diabetes Genetics Beyond GWAS.pdf",
+    "2015 - Diabetes mellitus The epidemic of the century.pdf",
+    "2013 - TCF7L2 gene polymorphisms and type 2 diabetes association with diabetic retinopathy and cardiovascular autonomic neuropathy.pdf",
+    "2007 - A German genome-wide linkage scan for type 2 diabetes supports the existence of a metabolic syndrome locus on chromosome 1p36.13 and a type 2 diabetes locus on chromosome 16p12.pdf",
+    "2015 - Diabetes mellitus The epidemic of the century.pdf",
+    "2013 - TCF7L2 gene polymorphisms and type 2 diabetes association with diabetic retinopathy and cardiovascular autonomic neuropathy.pdf"
+  ],
+  "extraction_id": [
+    "f7bbf3dd-a50c-59b5-9d67-721d4d14e744",
+    "eff1d167-9689-5c26-9a12-c66714696d86",
+    "86253f12-bb43-5236-bfb1-df5dff759f6d",
+    "5ffb710d-ca19-5415-bbb6-34b3f85bf47f",
+    "a3a875fa-e55b-52d0-b9bf-72b96330c393",
+    "36f9d4f2-293e-53e3-8b4b-12571af6669a",
+    "cdb798b2-60ec-5821-8fe5-3de463595e4d",
+    "198c5f2d-fc43-5744-9cd8-4222c8fa8ab8",
+    "7cfe9f29-a0ee-56d3-be3b-1b238a43bc07",
+    "b1e38adc-ccd8-5607-a43c-f3e4a4f5eafc"
+  ],
+  "document_id": [
+    "17637a6f-804e-50e4-9cf5-37318e17f15c",
+    "cf8ec75c-8ffe-5baa-830d-ac7a4a5964bd",
+    "11d0cb98-a00f-53f1-92e3-e1be17002c02",
+    "3236fdee-e304-5b88-921f-52e52dc5afa3",
+    "d59a38d7-889b-51b5-b896-c305c82a2169",
+    "e114dd28-fd39-56df-bdeb-8806474a6c10",
+    "89080fac-5530-5601-9822-9e1ba6defb29",
+    "f8a85060-6303-5228-ba89-7ee8701bda9f",
+    "e114dd28-fd39-56df-bdeb-8806474a6c10",
+    "89080fac-5530-5601-9822-9e1ba6defb29"
+  ],
+  "id": [
+    "chatcmpl-ADZBZNwu6nCnZNX1afjr9dVRs9c7K",
+    "0cc14865-601a-5668-9dfa-b5131c05ce91",
+    "3e678e3c-ad2c-5884-9c88-7f4d54c914bd",
+    "5d936c2c-faf7-5b0f-92e1-c3f8f43b3011",
+    "263f6b22-d314-5653-bbef-3f0e3e09839b",
+    "1eb3a215-002b-528b-a954-bb9e2419ea6f",
+    "6617e15c-ab52-596c-b628-60ec5a7001e7",
+    "b3f5855d-c61f-5bd7-bc12-508198ec48bd",
+    "05e76af5-c67b-50ca-a06a-a603d6d4b35e",
+    "ed5d8e9e-859e-5256-a7b5-468c1f7837a2",
+    "506f1dde-50cc-5f4d-b090-842a48d96f04"
+  ],
+  "contexts": [
+    "T. I., de Bakker, P . I. et al (2006). TCF7L2",
+    "single nucleotide polymorphisms in TCF7L2 are reproduc-ibly associated with type 2 diabetes and reduce the insulinresponse to glucose in nondiabetic individuals. Diabetes55:28902895 135. Cauchi S, Meyre D, Dina C, Choquet H, Samson C, Gallina S, Balkau B, Charpentier G, Pattou F, StetsyukV, Scharfmann R, Staels B, Fru  hbeck G, Froguel P 2006 Transcription factor TCF7L2 genetic study in the Frenchpopulation: expression in human /H9252-cells and adipose tissue",
+    "rs7903146 and rs12255372 in intron 3 of the TCF7L2 gene [20], associated with a ~45% increase in Type 2 diabetes risk per allele. As such, the TCF7L2 locus presently repre- sents the strongest known genetic determinant of Type 2diabetes. Risk allele carriers show impaired insulin produc-tion [21] and b-cell dysfunction in vitro [22]. TCF7L2 (previously referred to as TCF-4) is a high-mobility group box-containing transcription factor involved in Wingless-type MMTV integration site (Wnt)",
+    "et al. Variant of transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2  diabetes. Nat Genet . 2006;38:320-23.   Sladek R, Rocheleau G, Rung J, Dina C, Shen L, Serre D, et al. A genome- [9] wide association study identifies novel risk loci for type 2 diabetes. Nature .  2007;445:881-85.  Kirchhoff K, Machicao F, Haupt A, Schafer SA, Tschritter O, Staiger H, et al. [10] Polymorphisms in the TCF7L2, CDKAL1 and SLC30A8 genes are associated",
+    "transcription factor 7-like 2 ( TCF7L2 ) gene confers risk of type 2 diabetes. Nat Genet. 2006; 38:320323. [PubMed: 16415884] 172. Gloyn AL, Noordam K, Willemsen MA, Ellard S, Lam WW, et al. Insights into the biochemical and genetic basis of glucokinase activation from naturally occurring hypoglycemia mutations. Diabetes. 2003; 52:24332440. [PubMed: 12941786] 173. Pearson ER, Donnelly LA, Kimber C, Whitley A, Doney AS, et al. Variation in TCF7L2",
+    "L. Mechanisms by which common variants in the TCF7L2 gene  increase risk of type 2 diabetes. J Clin Invest  2007; 117: 2155-2163  [PMID: 17671651 DOI: 10.1172/JCI30706] 164 Gloyn AL , Braun M, Rorsman P. Type 2 diabetes susceptibility  gene TCF7L2 and its role in beta-cell function. Diabetes  2009; 58:  800-802 [PMID: 19336690 DOI: 10.2337/db09-0099] 165 da Silva Xavier G , Loder MK, McDonald A, Tarasov AI, Carzaniga  R, Kronenberger K, Barg S, Rutter GA. TCF7L2 regulates late",
+    "tion. Although the disease progression results from aninterplay of environmental factors and genetic predisposi- tion, in recent years TCF7L2 gene has been considered the strongest genetic determinant for the risk of developingT2DM [ 24,19,20]. The gene encodes a transcription factor of the canonical Wnt signaling pathway, expressed in several tissues, known to have developmental roles indetermining cell fate, survival, proliferation and movement [9]. Wnt signaling plays an important role also in B-cell",
+    "transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2diabetes. Nat Genet 38:320 3231422 Diabetologia (2007) 50:1418 1422",
+    "genes which also play a significant role in the risk and  pathogenesis of the disease[158,159]. The association  of TCF7L2  gene variants with type 2 diabetes and  its mechanism of action received special attention  by several investigators[161,162]. Over expression of the protein was shown to decrease the sensitivity of  beta islet cells to secrete insulin[163,164] and was more  precisely involved in the regulation of secretary granule  fusion that constitute a late event in insulin secretion",
+    "Muggeo M, Stoico V, Negri C, Pignatti PF, Bonora E, Bonadonna RC (2011) Variants and haplotypes of TCF7L2 are associatedwithb-cell function in patients with newly diagnosed type 2 diabetes: the Verona Newly Diagnosed Type 2 Diabetes Study (VNDS) 1. J Clin Endocrinol Metab 96(2):E389E393 13. Grundy SM, Cleeman JI, Merz CN, Brewer HB Jr, Clark LT, Hunninghake DB, Pasternak RC, Smith SC Jr, Stone NJ, National Heart, Lung, and Blood Institute, American College of Cardiol-"
+  ],
+  "answer": "The gene TCF7L2 is primarily involved in Type 2 diabetes. This association is well-documented across multiple sources:  1. Single nucleotide polymorphisms in TCF7L2 are reproducibly associated with Type 2 diabetes and reduce the insulin response to glucose in nondiabetic individuals [2]. 2. The TCF7L2 locus represents the strongest known genetic determinant of Type 2 diabetes, with risk allele carriers showing impaired insulin production and beta-cell dysfunction [3]. 3. Variants of the TCF7L2 gene confer risk of Type 2 diabetes [4], [5]. 4. TCF7L2 is considered the strongest genetic determinant for the risk of developing Type 2 diabetes [7]. 5. Overexpression of TCF7L2 protein decreases the sensitivity of beta islet cells to secrete insulin, which is involved in the regulation of insulin secretion [9].  Thus, the gene TCF7L2 is significantly involved in Type 2 diabetes.",
+  "question": "In which diseases is the gene TCF7L2 involved?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_26 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_26
new file mode 100644
index 0000000..1ac619c
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_26
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2008 -  Study Design and Statistical Issues.pdf",
+    "2009 - Opening Up the Conversation on Genetics.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2008 - Genotype-phenotype relationships and the patterning of complex traits as exemplified in the mammalian dentition.pdf",
+    "2007 - Promoting_Student_Scientific_Literacy_of_Molecular Genetics and Genomics.pdf",
+    "2019 - Sexual Dimorphism in the Age of Genomics How, When, Where.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2008 - Genotype-phenotype relationships and the patterning of complex traits as exemplified in the mammalian dentition.pdf"
+  ],
+  "extraction_id": [
+    "06bf0605-388a-592c-96ad-3a53bb36362c",
+    "51dbd5e2-fde6-5097-aa05-fcf57d3ca6b1",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "8a1ce8fa-b5f4-5942-b7b1-14a8a7887710",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "5aab3e60-b8b0-52ad-b4d3-817cf012cfa5",
+    "67369433-749b-5d6a-b5ef-3f0afe78b767",
+    "e22bb6fb-bec4-5c4c-8690-c96d0b8d13d4",
+    "206b8810-b7c1-5195-a10f-4e291864b77c",
+    "5aab3e60-b8b0-52ad-b4d3-817cf012cfa5"
+  ],
+  "document_id": [
+    "c3bd9cf0-f768-55c4-be94-96590d7acc21",
+    "b62a8f54-c2f5-5bbb-9324-af80f7537167",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "f6e866b8-b233-5862-bfb8-9949d0dabb97",
+    "755f34c4-cc06-5275-a744-16d48162b012",
+    "3f8c03b0-4235-5774-9d26-e43d55c1001b",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "f6e866b8-b233-5862-bfb8-9949d0dabb97"
+  ],
+  "id": [
+    "chatcmpl-ADZBd1UUBp1PO5G8iYcYCk8wkpCF7",
+    "1e151ad5-59d9-598d-97ba-90ba0e64c4cb",
+    "60ad1512-b0c0-59cd-ace4-c146e2c04b52",
+    "a66b8b00-d51c-575b-b6ac-fa445c4ca715",
+    "4472740a-d22d-5bb1-98e3-e91332cbb303",
+    "df4c6108-740d-5bcf-99e6-dbda74f7e41a",
+    "47b9142f-98a3-5a45-8eaa-d327c9cc055d",
+    "8e3fdc2c-0962-5854-83e7-a60ab05cf6de",
+    "0158f264-120f-5942-ad55-ef5fde1f188a",
+    "6c8dfaa1-a96f-5f1c-8b5a-870acfd46f5f",
+    "c3562719-52f3-50de-8e92-82f64e5c5b05"
+  ],
+  "contexts": [
+    "phenomena such as mutations and gene conversion events) occur in relevant meioses  leading up to the formation of the gametes (i.e., egg and sperm) which are combined  during fertilization and the formation of zygotes. Thus, individuals inherit a patch- work of chromosomal segments from maternal and paternal chromosomes.",
+    "the egg and the sperm. Such a process would result in genetic changes that will be copied into every cell of the future adult, including reproductive cells (Stock & Campbell, 2000), opening the door to irreversibly alter the human species. Inevitably, signifi  cant self-disclosure and discussion challenges await families",
+    "a fertilized egg is a complicated process that relies on controlling: which genes are active; whenthese genes activate; and for how long they are active. In broad terms, there are four ways that thiscontrol can be achieved: First, inside the sperm or egg, genes can be marked with small chemical tags that flag these genes",
+    "(Figures 8 and 9). Two gametes (egg and sperm) ultimately  join into a single cell, the zygote, which has the full comple-ment of 23 chromosome pairs restored. If all goes well, the zygote gives rise to a live offspring. The Mendel Laws: Segregation and Independent  Assortment Both of the Mendel laws pertain directly to the process of  meiosis. The first Mendel law, the law of segregation, states  that each parent passes a randomly selected allele for a given",
+    "to be activated (or remain inactive) after fertilization, depending on whether the modification wasmade by the father (in the sperm) or the mother (in the egg); this process is known as imprinting. Second, the mother can alter the gene activity in her offspring via the placenta; this process is known as maternal effect. Third, instructions encoded within the embryos DNA can directly control if, andwhen, a nearby gene becomes activated; this is known as  cis-regulation. Finally, similar instructions",
+    "the subset of that genetic information that is active.  But how does the differentiation process  begin?  The key insight in resolving this conundrum came from fly genetics and was the  realization that the egg is not a homogenous sack of protoplasm.  The maternally-derived genes  active in the fertilized egg are asymmetrically distributed such that at the first cell division each  daughter cell receives a different complement of factors.  Development continues as a",
+    "genes.  An altered gene may be passed on to every cell that develops from it.  The resulting features my help, harm, or have little or no effect on the offsprings success in its environment. (AAAS, pg. 109, 5B:9-12#4 ) 6. Heritable material: The information passed from parents to offspring is coded in DNA molecules (AAAS, pg 108, 5B:9-12#3) 7. Mutagens: Gene mutations can be caused by such things as radiation and chemicals.  When they occur in sex cells, the mutations can be passed onto offspring; if they",
+    "sex chromosome effects. (B)Soon after fertilization, male and female cells have sex-specic transcriptomes, epigenomes, and phenotypes (for example, male embryos grow faster than female embryos). At implantation, lineage determination begins and gene expression differences are reduced. Epigenetic marks, however, are less constrained and some are maintained, affecting gene expression, and phenotype later in development. Once specic lineages are established, differences in",
+    "or father (sperm cell). Each gamete has a set of 23 unpaired chromosomes. Two human gametes (egg and sperm) combine  to create a cell (zygote) that contains the full human genome  of 23 paired chromosomes.Genetic Information Nondiscrimination Act (GINA) US federal legislation that makes it unlawful to discriminate  against individuals on the basis of their genetic profiles in regard to health insurance and employment. These protections  are intended to encourage Americans to take advantage of",
+    "Proponents of the evo-devo view rightly point out that evolution occurs through changes in the  development of traits, which may or may not have changes in DNA as their root cause.  The  processes that produce traits occur during development and involve more than just genes.    All animals begin life as a fertilized egg, a single cell containing mitochondria and other  organelles, and enough maternally derived RNA and proteins to kick start development and"
+  ],
+  "answer": "Once a sperm combines with an egg, the traits passed onto the resulting lifeform are determined by several factors:  1. **Chromosomal Segments from Parents**: Individuals inherit a patchwork of chromosomal segments from both maternal and paternal chromosomes [1].     2. **Mendelian Laws**: The Mendel laws of segregation and independent assortment play a crucial role. The law of segregation states that each parent passes a randomly selected allele for a given trait to their offspring [4].  3. **Gene Activation and Regulation**: The process involves controlling which genes are active, when they activate, and for how long. This control can be achieved through mechanisms such as imprinting, maternal effect, cis-regulation, and other instructions encoded within the embryo's DNA [3], [5].  4. **Epigenetic Marks**: Soon after fertilization, epigenetic marks can affect gene expression and phenotype later in development [8].  5. **Heritable Material**: The information passed from parents to offspring is coded in DNA molecules [7].  These factors collectively determine how traits are inherited and expressed in the resulting lifeform.",
+  "question": "Once a sperm combines with an egg, what determines how traits are   passed onto the resulting lifeform?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_27 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_27
new file mode 100644
index 0000000..a0c741b
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_27
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2015 -Ghorbani- Biochemical Pathways and System Bio Analysis.pdf",
+    "2015 - Bioinformatics Methods for Biochemical Pathways and System Biology Analysis_.pdf",
+    "2015_GN_Diabets_notheses.pdf",
+    "2011 - A Role for the MS Analysis of Nucleic Acids.pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2012 - Genome-Wide Analysis of Yeast Aging.pdf",
+    "2009 - Next generation synthetic gene networks.pdf",
+    "2017 - Mutation and catastrophe in the aging genome.pdf",
+    "2008 - Gene Expression Profiling.pdf",
+    "2009 - Next generation synthetic gene networks.pdf"
+  ],
+  "extraction_id": [
+    "583e1b6c-5a64-5b10-aee0-9f25132cb1af",
+    "9166f54b-f72e-5028-a048-0a8c45e1d27e",
+    "c9653e9d-2bb8-5bfe-9c06-25286ee85b1b",
+    "24143fb1-ad8f-5ddc-a935-3ed4677445c9",
+    "02d772b3-ae4b-502c-a77e-1a4fdb261c30",
+    "168583fa-6b49-52c4-9a65-7463c95456a6",
+    "38e443bd-610e-5a1d-9f32-082e808d016a",
+    "b2844c41-6a10-5249-947a-811cbbe61bef",
+    "d78c044d-7554-54c2-a2b7-ae37e2e9b2cd",
+    "38e443bd-610e-5a1d-9f32-082e808d016a"
+  ],
+  "document_id": [
+    "222be20a-d320-540b-8b4f-888665acb158",
+    "9aa0b70d-ee1f-544b-bdac-f6a40c11e890",
+    "1744d9bf-29f9-52e3-a7c9-62a916999cda",
+    "2fa837a9-c3fb-5ee7-8e3d-58e4d8ec9ed9",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "cf24db9a-e013-5780-8b0f-369c56143f29",
+    "0d620c5e-a9ae-5b19-851b-37e40292ab8d",
+    "7ae205a2-e002-5e8b-bbf1-ea96ab599b37",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c",
+    "0d620c5e-a9ae-5b19-851b-37e40292ab8d"
+  ],
+  "id": [
+    "chatcmpl-ADZBi7XfGRRcyH1pOOCv70ZYOQgEu",
+    "b5e2c368-1437-591e-b886-e7aa879d8b1a",
+    "64fff1dc-a484-5138-9158-390bc27843ec",
+    "9f9d3ff1-8e26-53ab-b154-b72ca90e22ab",
+    "4b66fb22-87ff-5df3-99b4-6bdaea507bb5",
+    "870151f4-373d-50a7-8511-3a9a64f78514",
+    "de5d6db7-46da-5a27-bee4-48867524092d",
+    "296fc75a-e72d-5e72-a96f-8dd5fedbd709",
+    "b2b5baf2-4a44-5ecf-8c27-4789a878039f",
+    "98cdcaa8-1595-5269-9019-69f381738c58",
+    "d414d811-e98b-54e7-ad01-b4f185511dd9"
+  ],
+  "contexts": [
+    "promoters ,regulatory proteins and their binding sites,  ribosomal binding sites terminators ,et. RegulonDB contains  both documentation and prediction objects. In  addition it is linked with Swiss -prot, with microarray  databases for analysis and visualization of microarray  experiments.[5]     WIT   The WIT (What Is There) (http://wit.mcs.anl.gov/WIT2/)  is a comparable computational system for analysis of  sequenced genomes and generation of metabolic",
+    "promoters ,regulatory proteins and their binding sites,  ribosomal binding sites terminators ,et. RegulonDB contains  both documentation and prediction objects. In  addition it is linked with Swiss -prot, with microarray  databases for analysis and visualization of microarray  experiments.[5]     WIT   The WIT (What Is There) (http://wit.mcs.anl.gov/WIT2/)  is a comparable computational system for analysis of  sequenced genomes and generation of metabolic",
+    "promoters ,regulatory proteins and their binding sites,  ribosomal binding sites terminators ,et. RegulonDB contains  both documentation and prediction objects. In  addition it is linked with Swiss -prot, with microarray  databases for analysis and visualization of microarray  experiments.[5]     WIT   The WIT (What Is There) (http://wit.mcs.anl.gov/WIT2/)  is a comparable computational system for analysis of  sequenced genomes and generation of metabolic",
+    "173. Griffey, R. H.; Greig, M. J.; Haoyun, A.; Sasmor, H.; Manalili, S. Targeted Site-Specific Gas-Phase Cleavage of Oligoribonucleotides. Application in Mass Spectrometry-Based Identification of Ligand Binding Sites. J. Am. Chem. Soc. 1999, 121, 474475. 174. Hanson, C. L.; Fucini, P.; Ilag, L. L.; Nierhaus, K. H.; Robinson, C. V. Dissociation of Intact Escherichia coli Ribosomes in a Mass Spectrome- terEvidence for Conformational Change in a Ribosome Elongation",
+    "or chloramphenicol Immobilized targetDissociation of ribosome and release of mRNA5Poly(AAA)3 mRNA Isolation of mRNART-PCRdsDNA Mutagenesis by error-prone PCR Fig. 35.5.  Schematic presentation of a ribosome display round. The gene of interest is  transcribed from dsDNA into mRNA and translated into proteins by in vitro techniques.  The ribosomes remain tethered to the mRNA by either cold shock or chloramphenicol.  This step ensures that the genotype remains coupled to the phenotype. The proteins are",
+    "270 G.L. Sutphin e t a l. gene (Hinneb usch 2005 ). The m echanism of re gulation i s t hought to in v o lv e r el- ati v e a v a ilability of the l ar ge and small r ibosome s ub units. Specically , w hen 60Sribosomal sub unit l e v els a re lo w , ternary comple x e s containing initiation f actors and 40S ribosomal sub units are p roposed to more frequently scan through the",
+    "then used to develop synthetic gene networks with defined outputs, without significant post-hoc adjustments 22,4751. Alternatively, syn- thetic ribosome binding site (RBS) sequences can be used to optimize protein expression levels. Recently, Salis et al. 52 have developed a  thermodynamic model for predicting the relative translational ini -",
+    "Philips, R.M., 2017 How Many Ribosomes Are in a Cell? [WWW Document]. URL http:// book.bionumbers.org/how-many-ribosomes-are-in-a-cell/ ((accessed 7.24.16) n.d.). R Core Team, 2014. R: a Language and Environment for Statistical Computing. R Founda- tion for Statistical Computing, Vienna, Austria. Sigurdson, A.J., Ha, M., Hauptmann, M., Bhatti, P., Sram, R.J., Beskid, O., Tawn, E.J.,",
+    "structure, and to find sites that are likely to be cleaved or modified; interac- tion or catalytic mechanisms can be simulated. Bioinformatic resources on  the WWW range from the determination of the molecular weight to complex  threading and three-dimensional (3D) prediction algorithms. A huge list of  tools can be found on the ExPASy proteomic tools homepage (65). Because  of the great variety of programs available, several of these single tools have",
+    "tiation rates for a protein with different upstream RBS sequences, a model that can also be used to rationally forward-engineer RBS sequences to give desired protein expression. In addition, protein degradation can be controlled by tagging proteins with degradation-targeting peptides that impart different degradation dynamics 53. By automating the construction and characterization of biomo-"
+  ],
+  "answer": "A ribosomal binding site (RBS) is a specific sequence of nucleotides in messenger RNA (mRNA) that is crucial for the initiation of protein synthesis. At a high level, you can think of it as a \"landing pad\" for the ribosome, which is the molecular machine that reads the mRNA and translates it into a protein.  Here's a simple analogy: Imagine the ribosome as a train and the mRNA as a track. The ribosomal binding site is like a train station where the train (ribosome) stops to pick up passengers (the instructions for making a protein). Without this station, the train wouldn't know where to start its journey.  In more technical terms, the RBS helps the ribosome to correctly position itself on the mRNA so that it can start translating the genetic code into a protein at the right spot. This process is essential for the proper expression of genes and the production of proteins that perform various functions in the cell [1], [2], [3].",
+  "question": "Can you explain what a ribosomal binding site at a high level and make it accessable to a non-expert?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_28 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_28
new file mode 100644
index 0000000..89b7324
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_28
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2008 -  Study Design and Statistical Issues.pdf",
+    "2009 - Opening Up the Conversation on Genetics.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2015 - Constraint and divergence of global gene expression in the mammalian embryo.pdf",
+    "2008 - Genotype-phenotype relationships and the patterning of complex traits as exemplified in the mammalian dentition.pdf",
+    "2007 - Promoting_Student_Scientific_Literacy_of_Molecular Genetics and Genomics.pdf",
+    "2019 - Sexual Dimorphism in the Age of Genomics How, When, Where.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2008 - Genotype-phenotype relationships and the patterning of complex traits as exemplified in the mammalian dentition.pdf"
+  ],
+  "extraction_id": [
+    "06bf0605-388a-592c-96ad-3a53bb36362c",
+    "51dbd5e2-fde6-5097-aa05-fcf57d3ca6b1",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "8a1ce8fa-b5f4-5942-b7b1-14a8a7887710",
+    "261c4af7-f63d-51ac-b164-0d9e7a64bff9",
+    "5aab3e60-b8b0-52ad-b4d3-817cf012cfa5",
+    "67369433-749b-5d6a-b5ef-3f0afe78b767",
+    "e22bb6fb-bec4-5c4c-8690-c96d0b8d13d4",
+    "206b8810-b7c1-5195-a10f-4e291864b77c",
+    "5aab3e60-b8b0-52ad-b4d3-817cf012cfa5"
+  ],
+  "document_id": [
+    "c3bd9cf0-f768-55c4-be94-96590d7acc21",
+    "b62a8f54-c2f5-5bbb-9324-af80f7537167",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "3d9005f1-8f71-5d39-8749-4ebeab962cab",
+    "f6e866b8-b233-5862-bfb8-9949d0dabb97",
+    "755f34c4-cc06-5275-a744-16d48162b012",
+    "3f8c03b0-4235-5774-9d26-e43d55c1001b",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "f6e866b8-b233-5862-bfb8-9949d0dabb97"
+  ],
+  "id": [
+    "chatcmpl-ADZBnwLDpOz1ruhxgc8K7GIp36KrG",
+    "1e151ad5-59d9-598d-97ba-90ba0e64c4cb",
+    "60ad1512-b0c0-59cd-ace4-c146e2c04b52",
+    "a66b8b00-d51c-575b-b6ac-fa445c4ca715",
+    "4472740a-d22d-5bb1-98e3-e91332cbb303",
+    "df4c6108-740d-5bcf-99e6-dbda74f7e41a",
+    "47b9142f-98a3-5a45-8eaa-d327c9cc055d",
+    "8e3fdc2c-0962-5854-83e7-a60ab05cf6de",
+    "0158f264-120f-5942-ad55-ef5fde1f188a",
+    "6c8dfaa1-a96f-5f1c-8b5a-870acfd46f5f",
+    "c3562719-52f3-50de-8e92-82f64e5c5b05"
+  ],
+  "contexts": [
+    "phenomena such as mutations and gene conversion events) occur in relevant meioses  leading up to the formation of the gametes (i.e., egg and sperm) which are combined  during fertilization and the formation of zygotes. Thus, individuals inherit a patch- work of chromosomal segments from maternal and paternal chromosomes.",
+    "the egg and the sperm. Such a process would result in genetic changes that will be copied into every cell of the future adult, including reproductive cells (Stock & Campbell, 2000), opening the door to irreversibly alter the human species. Inevitably, signifi  cant self-disclosure and discussion challenges await families",
+    "a fertilized egg is a complicated process that relies on controlling: which genes are active; whenthese genes activate; and for how long they are active. In broad terms, there are four ways that thiscontrol can be achieved: First, inside the sperm or egg, genes can be marked with small chemical tags that flag these genes",
+    "(Figures 8 and 9). Two gametes (egg and sperm) ultimately  join into a single cell, the zygote, which has the full comple-ment of 23 chromosome pairs restored. If all goes well, the zygote gives rise to a live offspring. The Mendel Laws: Segregation and Independent  Assortment Both of the Mendel laws pertain directly to the process of  meiosis. The first Mendel law, the law of segregation, states  that each parent passes a randomly selected allele for a given",
+    "to be activated (or remain inactive) after fertilization, depending on whether the modification wasmade by the father (in the sperm) or the mother (in the egg); this process is known as imprinting. Second, the mother can alter the gene activity in her offspring via the placenta; this process is known as maternal effect. Third, instructions encoded within the embryos DNA can directly control if, andwhen, a nearby gene becomes activated; this is known as  cis-regulation. Finally, similar instructions",
+    "the subset of that genetic information that is active.  But how does the differentiation process  begin?  The key insight in resolving this conundrum came from fly genetics and was the  realization that the egg is not a homogenous sack of protoplasm.  The maternally-derived genes  active in the fertilized egg are asymmetrically distributed such that at the first cell division each  daughter cell receives a different complement of factors.  Development continues as a",
+    "genes.  An altered gene may be passed on to every cell that develops from it.  The resulting features my help, harm, or have little or no effect on the offsprings success in its environment. (AAAS, pg. 109, 5B:9-12#4 ) 6. Heritable material: The information passed from parents to offspring is coded in DNA molecules (AAAS, pg 108, 5B:9-12#3) 7. Mutagens: Gene mutations can be caused by such things as radiation and chemicals.  When they occur in sex cells, the mutations can be passed onto offspring; if they",
+    "sex chromosome effects. (B)Soon after fertilization, male and female cells have sex-specic transcriptomes, epigenomes, and phenotypes (for example, male embryos grow faster than female embryos). At implantation, lineage determination begins and gene expression differences are reduced. Epigenetic marks, however, are less constrained and some are maintained, affecting gene expression, and phenotype later in development. Once specic lineages are established, differences in",
+    "or father (sperm cell). Each gamete has a set of 23 unpaired chromosomes. Two human gametes (egg and sperm) combine  to create a cell (zygote) that contains the full human genome  of 23 paired chromosomes.Genetic Information Nondiscrimination Act (GINA) US federal legislation that makes it unlawful to discriminate  against individuals on the basis of their genetic profiles in regard to health insurance and employment. These protections  are intended to encourage Americans to take advantage of",
+    "Proponents of the evo-devo view rightly point out that evolution occurs through changes in the  development of traits, which may or may not have changes in DNA as their root cause.  The  processes that produce traits occur during development and involve more than just genes.    All animals begin life as a fertilized egg, a single cell containing mitochondria and other  organelles, and enough maternally derived RNA and proteins to kick start development and"
+  ],
+  "answer": "Once a sperm combines with an egg, the traits passed onto the resulting lifeform are determined by several factors:  1. **Chromosomal Segments**: Individuals inherit a patchwork of chromosomal segments from maternal and paternal chromosomes due to phenomena such as mutations and gene conversion events during meiosis [1].  2. **Mendelian Laws**: The Mendel laws of segregation and independent assortment play a crucial role. The law of segregation states that each parent passes a randomly selected allele for a given trait to their offspring [4].  3. **Gene Activation and Regulation**: The process of trait determination also involves controlling which genes are active, when they activate, and for how long. This control can be achieved through mechanisms such as imprinting, maternal effect, and cis-regulation [3], [5].  4. **Epigenetic Marks**: Epigenetic marks, which are chemical modifications to DNA that do not change the DNA sequence but affect gene expression, also play a role. These marks can be inherited and influence gene expression and phenotype later in development [8].  5. **Heritable Material**: The information passed from parents to offspring is coded in DNA molecules, and any mutations in these genes can be inherited by the offspring [7].  In summary, the combination of chromosomal inheritance, Mendelian genetic laws, gene activation and regulation mechanisms, epigenetic marks, and the heritable DNA material all contribute to how traits are passed onto the resulting lifeform after fertilization.",
+  "question": "Once a sperm combines with an egg, what determines how traits are   passed onto the resulting lifeform?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_29 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_29
new file mode 100644
index 0000000..9216e36
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_29
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2019 - Genomic Analysis in the Age.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2018 - Effects of Genetic Background on Susceptibility and the Acceleration of Hearing Loss in Mice.pdf",
+    "2012 - Functional genomics research in aquaculture principles and general approaches.pdf"
+  ],
+  "extraction_id": [
+    "c91e328e-4a01-5952-85b8-d7b5b47237c5",
+    "c91e328e-4a01-5952-85b8-d7b5b47237c5",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "7fb68eb5-75cc-5db7-a182-d0ea055d49fe",
+    "06285eb9-37a8-5f76-a6d4-69cab398f2c0",
+    "97796d0a-1595-5cc9-a0db-c4186788ad07",
+    "97796d0a-1595-5cc9-a0db-c4186788ad07",
+    "97796d0a-1595-5cc9-a0db-c4186788ad07",
+    "a97b6b0b-d841-5cd3-a79f-f6d283b8337c",
+    "e5aa10c8-8b26-517f-9725-cb809cb4a37a"
+  ],
+  "document_id": [
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "f50c4d62-acab-5024-8ec7-526fffbfbf25",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "0567de5c-e886-5660-82de-8b80d2b580a9",
+    "a39b4cc1-8661-578b-a61b-b9962e45fc33"
+  ],
+  "id": [
+    "chatcmpl-ADZBtpMuJOymoi8ODiNQwPGHnYpg6",
+    "a2d9c614-903d-513a-ad88-5a40f3534988",
+    "aa1d9f58-486c-522f-9981-5ce7e943b87f",
+    "47703589-35f9-5cff-8e62-ed299caa3356",
+    "5bd5b104-1b21-536e-90b2-2179bd152858",
+    "f49954d4-5769-5b9d-b06c-9f0050ab9e81",
+    "aec521d0-0c70-59bc-b457-6d801e8a7ab7",
+    "7445eff9-43fa-5328-84b7-5db7f16197e2",
+    "76137d35-eb92-5512-bbff-fa90de8e445c",
+    "63009249-a23b-5b5f-b9aa-34dc63c88218",
+    "c80d766b-4629-5a42-b3c2-877aa3f5af7c"
+  ],
+  "contexts": [
+    "for sequencing on existing short-read instrumentation,  after which data are split by barcode and reassembled  with the knowledge that fragments sharing barcodes Barcodes A series of known bases  addedto a template molecule  either through ligation or  amplification. After  sequencing, these barcodes  can be used to identify which  sample a particular read is  derived from. Figure 5 | Real-time and synthetic long-read sequencing approaches.",
+    "sequence 2D read. Synthetic long-reads.  Unlike true sequencing platforms,  synthetic long-read technology relies on a system of  barcoding to associate fragments that are sequenced on  existing short-read sequencers61. These approaches par - tition large DNA fragments into either microtitre wells  or an emulsion such that very few molecules exist in  each partition. Within each partition the template frag - ments are sheared and barcoded. This approach allows",
+    "sequencing. This platform is used by the Illumina  suite of platforms. 36. Dohm,J.C., Lottaz,C., Borodina,T . &  Himmelbauer,H. Substantial biases in ultra-short read  data sets from high-throughput DNA sequencing.  Nucleic Acids Res. 36, e105 (2008). 37. Nakamura,K. etal.  Sequence-specific error profile  ofIllumina sequencers. Nucleic Acids Res. 39, e90  (2011). 38. Minoche,A.E., Dohm,J.C. & Himmelbauer,H.  Evaluation of genomic high-throughput sequencing  data generated on Illumina HiSeq and genome",
+    "Comparison of short-read platforms.  Individual short- read sequencing platforms vary with respect to through - put, cost, error profile and read structure (TABLE1 ).  Despite the existence of several NGS technology pro - viders, NGS research is increasingly being conducted  within the Illumina suite of instruments21. Although  this implies high confidence in their data, it also raises  concerns about systemic biases derived from using a  single sequencing approach2628. As a consequence, new",
+    "short-read sequencing. arXiv, arXiv:1203.3907v2, https://arxiv.org/abs/ 12073907 . Garrison, E., Sire n, J., Novak, A.M., Hickey, G., Eizenga, J.M., Dawson, E.T., Jones, W., Garg, S., Markello, C., Lin, M.F., et al. (2018). Variation graph toolkit improves read mapping by representing genetic variation in the reference. Nat. Biotechnol. 36, 875879 . Giambartolomei, C., Vukcevic, D., Schadt, E.E., Franke, L., Hingorani, A.D.,",
+    "or  transcriptomic structure53. Long-read sequencing Overview.  It has become apparent that genomes are  highly complex with many long repetitive elements,  copy number alterations and structural variations that  are relevant to evolution, adaptation and disease5456.  However, many of these complex elements are so long  that short-read paired-end technologies are insufficient  to resolve them. Long-read sequencing delivers reads in  excess of several kilobases, allowing for the resolution of",
+    "these large structural features. Such long reads can span  complex or repetitive regions with a single continuous  read, thus eliminating ambiguity in the positions or size  of genomic elements. Long reads can also be useful for  transcriptomic research, as they are capable of span - ning entire mRNA transcripts, allowing researchers to  identify the precise connectivity of exons and discern  geneisoforms. Currently, there are two main types of long-read tech -",
+    "nologies: single-molecule real-time sequencing  approaches  and synthetic approaches that rely on existing short- read technologies to construct long reads  insilico . The  single-molecule approaches differ from short-read  approaches in that they do not rely on a clonal popula - tion of amplified DNA fragments to generate detectable Figure 2 | Sequencing by ligation methods. a | SOLiD sequencing. Following cluster  generation or bead deposition onto a slide, fragments are sequenced by ligation, in",
+    "Tools for alignment-free analyses of sequencing data The vast majority of next-generation sequencing experiments in mouse have read alignment  to a reference genome as their first step. However, the primary data from any sequencing  experiment are the reads themselves. Recognition that the raw reads are information-rich has  led to the development of alignment-free algorithms for error correction (among many  others, Chaisson and Pevzner 2008 ), abundance estimation ( Patro et al. 2014 ), and de novo",
+    "(right). Sequencing adaptors (depicted by short red bars and short purple bars) are subsequently ligated to each cDNA fragment (green lines) and short sequence reads (single end or paired ends) from each cDNA are generated using high-throughput sequencing technology. The resulting sequence reads [short lines beneath the genome sequence with three genes shown (fat blue bars)] are aligned with the reference genome to"
+  ],
+  "answer": "Sure! Let's break down the difference between short-read and long-read sequencing in a way that's easy to understand.  ### Short-Read Sequencing: 1. **Read Length**: Short-read sequencing generates small fragments of DNA, typically around 50-300 base pairs long [4]. 2. **Technology**: This method is commonly used by platforms like Illumina [3]. 3. **Advantages**:     - **High Throughput**: It can produce a large amount of data quickly and is relatively cost-effective [4].    - **Accuracy**: Generally, short-read sequencing has a lower error rate per base compared to long-read sequencing [4]. 4. **Limitations**:    - **Complex Regions**: It struggles to accurately sequence complex regions of the genome, such as those with long repetitive elements or structural variations [6].    - **Assembly**: Because the reads are short, assembling them into a complete genome can be challenging and may leave gaps or errors [6].  ### Long-Read Sequencing: 1. **Read Length**: Long-read sequencing produces much longer fragments, often several kilobases (thousands of base pairs) in length [6]. 2. **Technology**: There are two main types:    - **Single-Molecule Real-Time Sequencing**: This method reads long DNA fragments directly without needing to amplify them [8].    - **Synthetic Long-Reads**: This method uses short-read technology to create long reads in a computational process [2]. 3. **Advantages**:    - **Complex Regions**: It can span complex or repetitive regions in a single read, making it easier to resolve these areas accurately [7].    - **Transcriptomics**: Long reads can cover entire mRNA transcripts, helping researchers understand gene structures and variations better [7]. 4. **Limitations**:    - **Cost and Throughput**: Long-read sequencing is generally more expensive and produces less data per run compared to short-read sequencing [4].    - **Error Rate**: The error rate per base can be higher than that of short-read sequencing, although this is improving with new technologies [4].  In summary, short-read sequencing is like reading a book by looking at many small snippets of text, which is fast and accurate but can be tricky if the text is very repetitive or complex. Long-read sequencing, on the other hand, is like reading longer passages at a time, which helps to understand the context better but might be slower and more expensive.",
+  "question": "Can you explain the difference between sequencing with short reads vs long reads? Please make you answer accessible to a non-expert."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_3 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_3
new file mode 100644
index 0000000..9e65677
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_3
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "1998 - Neurodegeneration and Aging Role.pdf",
+    "1998 - Neurodegeneration and Aging Role.pdf",
+    "1998 - Neurodegeneration and Aging Role.pdf",
+    "2015 - Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke Clinicians.pdf",
+    "2020 - Mitonuclear genomics and aging.pdf",
+    "2020 - Mitonuclear genomics and aging.pdf",
+    "2002 - Genomic Medicine - A Primer.pdf",
+    "2001 - Mitochondrial genome instability in human cancers.pdf",
+    "2005 - The mitochondrial genome in human adaptive radiation and disease.pdf",
+    "2015 - Altered Levels of Mitochondrial DNA.pdf"
+  ],
+  "extraction_id": [
+    "ceaf66e9-9822-5f7e-84b7-c687982f63e1",
+    "ceaf66e9-9822-5f7e-84b7-c687982f63e1",
+    "ceaf66e9-9822-5f7e-84b7-c687982f63e1",
+    "fa1dc2c0-8cc1-53e1-ad3e-8037506ec897",
+    "9c7f0bf0-7180-587e-a852-1187f18c2aea",
+    "472c8adc-54e7-5c27-a7b8-882b7e49cd2b",
+    "8e4ad64b-5541-52aa-bcd0-d61a8add4662",
+    "e79c57f4-e46b-5d8a-b9f3-2ee45c27349f",
+    "2f3c7ffe-45b9-5437-89cf-5fb7bbadc3d5",
+    "f85318e1-8e21-512e-8805-ade90e5232d0"
+  ],
+  "document_id": [
+    "694c44f3-6025-5a2c-9c72-d9c5f16c8b85",
+    "694c44f3-6025-5a2c-9c72-d9c5f16c8b85",
+    "694c44f3-6025-5a2c-9c72-d9c5f16c8b85",
+    "8610e699-218a-50e6-8d1d-ef689623266f",
+    "e05fdc09-c8d8-5134-a1fd-bf07a1564981",
+    "e05fdc09-c8d8-5134-a1fd-bf07a1564981",
+    "a55c57bb-56e9-5218-ae80-66ee7aa93ae5",
+    "1a8d11d4-c469-5c5c-80a6-3800615dac43",
+    "c157cd40-1e5f-5ddb-b9cf-7927d8f2ee7b",
+    "ee28bbbd-66e4-5d9e-abcb-41cd89fcae22"
+  ],
+  "id": [
+    "chatcmpl-ADZ9W7mSuiYhzAdBeFHuooWOpJA78",
+    "bdba9fea-2d0e-534e-9f32-688e499ec4dc",
+    "bbf1b1b5-873d-51cb-ada1-000a02458e32",
+    "b452f309-162e-5611-87c2-6907fd854ebf",
+    "a2eaf128-38b1-592d-b340-79ff1333c36e",
+    "e134edd5-f5f0-54d7-bc05-991f3d930de4",
+    "99a2cfc1-5a54-53af-b2a4-4c274e1d5ef1",
+    "a7a9acf4-c7c3-5b14-b40f-6667f32f74ab",
+    "6baf04ee-578e-5c53-9751-46241804e085",
+    "2570a50b-982d-5a2d-b5c1-c10e2fbcaed2",
+    "9c8c7e6b-7701-5a5e-a3e9-0b22eb29c099"
+  ],
+  "contexts": [
+    "drial DNA sequence variation seems impossible withoutan understanding of some important differences betweennuclear and mitochondrial genetics (Table I). Mitochon-drial DNA replicates autonomously and is inherited viathe cytoplasm of the parent cell with the individualmitochondrion being the segregating unit (Attardi et al.,1995). Thus, in the case of mitochondrial mutations bothmutated as well as normal mitochondria may be presentwithin the same cell. This situation has been termedheteroplasmy and can",
+    "cMitochondria are semiautonomous organelles; possess their own replication-, transcription- and translation system cExclusively maternal inheritance of mitochondrial DNA cMitotic segregation of mitochondrial DNAcan lead to hetero- plasmy, i.e., the proportion of genetically different populations ofmitochondria differs between generations of mitotically activecells cApproximately tenfold higher mutation rate compared with nuclear",
+    "DIFFERENCES BETWEEN MITOCHONDRIAL AND NUCLEAR GENETICS Arealisticassessmentoftherelevanceofmitochon-",
+    "In the fifth mode of inheritance, the disease mutation lies not on a chromosome in the nucleus but rather in mitochondrial DNA outside the nucleus. Mitochondria are inherited exclu- sively from an offsprings mother; because of this phenome- non, the mutation and thus the disease can be passed only from a mother to her offspring. This is maternal inheritance, also known as extranuclear inheritance (Figure 11). Representative disorders include various mitochondrial myopathies.",
+    "The regulation of the mitochondrial genome also reflects  its prokaryotic ancestry. While nuclear DNA undergoes replication during cell division, mtDNA replication occurs independently of cell cycle. The majority of the compo-nents for mtDNA replication are imported nuclear-encoded proteins, including the catalytic subunit of mtDNA poly -",
+    "Unlike the nuclear genome, which requires both paternal and maternal contributions, mtDNA is inherited solely from the maternal lineage. It is unclear what advantage a uniparental mtDNA transmission confers, but one possibil-ity is to minimize the number of distinct genomes to maxi-mize the efficiency of a multi-genomic system (Hill etal. 2019). In fact, humans have developed complex, redundant mechanisms to ensure uniparental inheritance of mtDNA (DeLuca and OFarrell 2012; Rojansky etal. 2016). Paternal",
+    "mitochondria and sperm are not, mitochondrial DNAis usually inherited from the mother. Therefore, mito-chondrial genes  and diseases due to DNA-sequencevariants in them  are transmitted in a matrilineal pat-tern that is distinctly different from the pattern of in-heritance of nuclear genes.  MONOGENIC CONDITIONS  Over the course of the 20th century, a combination",
+    "2. Mitochondrial DNA structure and properties Mitochondrial genomes (mt-genomes) are short circular molecules that, with the exception of viruses,represent the most economically packed forms ofDNA in the whole biosphere. The human mt-genomeis only 16,569bp long [9]; within this extension, wend the coding sequences for seven subunits of theNADH-ubiquitone reductase (respiratory complex I),the apocytochrome bof the ubiquitone cytochrome creductase (respiratory complex III), three subunits",
+    "Abstract The human mitochondrial genome consists of approximately 1500 genes, 37 encoded by the maternally inherited mitochondrial DNA (mtDNA) and the remainder encoded in the nuclear DNA (nDNA). The mtDNA is present in thousands of copies per cell and encodes",
+    "(mtDNA). MtDNA carries important genetic information concerning cellular metabolismand the generation of energy. It has been suggested that mitochondria and mtDNA could be of significance during early embryo development. Our work confirms this hypothesis. Specif- ically, our findings implicate mitochondria and their genome in female reproductive agingand the generation of embryonic chromosome abnormalities. Importantly, we describe a di-"
+  ],
+  "answer": "Mitochondrial DNA (mtDNA) plays a unique role in heredity and differs significantly from nuclear DNA (nDNA) in several ways:  1. **Inheritance Pattern**: Mitochondrial DNA is inherited exclusively from the mother, a phenomenon known as maternal or extranuclear inheritance [4], [6], [7]. This contrasts with nuclear DNA, which is inherited from both parents.  2. **Replication and Segregation**: mtDNA replicates autonomously and independently of the cell cycle, unlike nuclear DNA which replicates during cell division [1], [5]. Mitochondria are semiautonomous organelles with their own replication, transcription, and translation systems [2].  3. **Genetic Variation and Mutation Rate**: Mitochondrial DNA has a higher mutation rate compared to nuclear DNA, approximately tenfold higher [2]. This can lead to a condition called heteroplasmy, where both mutated and normal mitochondria coexist within the same cell [1], [2].  4. **Structure and Size**: The mitochondrial genome is a short circular molecule, significantly smaller than the nuclear genome. The human mitochondrial genome is only 16,569 base pairs long and encodes 37 genes, whereas the nuclear genome consists of approximately 1500 genes related to mitochondrial function [8], [9].  5. **Functional Role**: mtDNA carries essential genetic information for cellular metabolism and energy generation, which is crucial for the functioning of the cell [10]. This is distinct from the broader range of functions encoded by nuclear DNA.  In summary, mitochondrial DNA is maternally inherited, replicates independently, has a higher mutation rate, and is structurally distinct from nuclear DNA. These differences underscore the unique role of mtDNA in heredity and cellular function.",
+  "question": "Describe the role of mitochondrial DNA in heredity and how it differs from nuclear DNA."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_30 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_30
new file mode 100644
index 0000000..f277081
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_30
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2021 - Plant Pan-Genomics.pdf",
+    "2018 - Effects of Genetic Background on Susceptibility and the Acceleration of Hearing Loss in Mice.pdf",
+    "2019 - Genomic Analysis in the Age.pdf",
+    "2023 - Clinical, technical, and environmental biases.pdf",
+    "2021 - Plant Pan-Genomics.pdf",
+    "2011 - The Reference Human Genome High Risk of Type 1 Diabetes and Other Disorder.pdf",
+    "2015 - Informatics resources for the Collaborative Cross and related mouse populations.pdf",
+    "2021 - Human Molecular Genetics and Genomics.pdf",
+    "2009 - Detection and interpretation of expression quantitative trait loci (eQTL).pdf",
+    "2017 - Post-genomic behavioral genetics From revolution to routine.pdf"
+  ],
+  "extraction_id": [
+    "b75d8a8c-6c3a-5fce-92ee-46ae61aceb95",
+    "bcae5dd7-f775-5634-801b-76a71c99b2f4",
+    "70f829cc-2b89-593f-9995-f3e1d369acd4",
+    "7b399dda-fb0e-5111-929c-78fa82a74636",
+    "73f80ca8-2f2c-5ff4-9b65-2eeae1fd0b02",
+    "de94e095-34e7-537c-8c85-531bb17f4735",
+    "ffe01714-be5b-5aaa-889b-b83e97fc022c",
+    "35967ed4-335d-5b3a-b66f-97f3073a292d",
+    "8cc88dd8-4985-57f5-93db-4bbf171f938b",
+    "022e1268-76b1-590b-b73e-a096d4719c72"
+  ],
+  "document_id": [
+    "3b346320-36f0-593c-bb36-c40cc6e23715",
+    "0567de5c-e886-5660-82de-8b80d2b580a9",
+    "f50c4d62-acab-5024-8ec7-526fffbfbf25",
+    "6a81e435-bd17-558d-850a-44ee3dbab5bd",
+    "3b346320-36f0-593c-bb36-c40cc6e23715",
+    "05e764f5-4ae8-51b7-89f0-987c79f6ed8f",
+    "889af7dc-d665-59a8-8b32-d3a65a831c70",
+    "68e362a5-e580-5a4d-8d41-6a138c873ede",
+    "ef974b09-4ea2-5382-85e5-c2169f440fda",
+    "cf1fdd6b-e926-5e84-a6b1-a5e92abbd2f3"
+  ],
+  "id": [
+    "chatcmpl-ADZC0hJis0QrHtORi8K0UBB4TqKH0",
+    "66e86865-9c57-5ee7-883c-7bd1044fa708",
+    "83a31bf6-bd31-5a7b-ad2b-0f4223aa085a",
+    "21c0b3f1-a901-5a49-88ff-38963651d6cd",
+    "c43cf59c-5359-50cb-b9ee-73e74e3e1bd7",
+    "13a284d7-ff1c-5933-bce0-a69bbcee02cc",
+    "872237a6-b34e-57b4-bc4f-9967f8908796",
+    "940a31fb-adfd-558c-9c9d-39cb8d1ecee6",
+    "edcd5595-3b69-5ebe-b24f-a0c611f79606",
+    "16f7648c-92d7-5128-ae30-2a19ec89e04c",
+    "14cd9387-ac3c-52f9-81c3-c535925aeea8"
+  ],
+  "contexts": [
+    "When reliable prior knowledge exists about the variant composition in a pan-genome (typi- cally obtained via read-to-reference mapping), there are computational tools that can transform a linear reference sequence and a set of variant calls into graphs (18).This approach bypasses the computationallyexpensiveall-versus-allalignmentstepalongwiththeuncertaintiesofsubsequent graph construction, but the trade-off is increased reference bias and a potentially incomplete",
+    "(Karolchik et al. 2014 )] and Ensembl ( Flicek et al. 2013 ). Use of a single haploid reference  sequence as an anchor for all studies of genetic variation in mouse offers many practical  advantages. But the dependency on a reference genome requires several assumptions about  the nature of genetic variation which may be violated in practicethe strongest of which is  that of genomic collinearity (i.e., conserved marker order) between strains. We consider the",
+    "for at least 500 ancestrally diverse humans. This resource willalso provide a set of highly accurate genomes that can be used as a benchmarking dataset to improve short-read analysis tools. Even more importantly, these genomes allow completelynew designs for more effective short-read analysis strategiesthat overcome many of the limitations described above. Transitioning to a pan-genome reference will require develop-",
+    "2018;562(7726):203-209. http://doi.org/10.1038/s41586-018-0579-z 110. Li R, Li Y, Zheng H, et al. Building the sequence map of the human pan-genome. Nat Biotechnol . 2010;28(1):57-63. http://doi.org/10. 1038/nbt.1596 111. Vernikos G, Medini D, Riley DR, Tettelin H. Ten years of pan- genome analyses. Curr Opin Microbiol . 2015;23:148-154. http:// doi.org/10.1016/j.mib.2014.11.016 112. Miga KH, Wang T. The need for a human pangenome reference sequence. Annu Rev Genomics Hum Genet . 2021;22:81-102. http://",
+    "Whilemostpan-genomesconstructedtodateareprimarilygene-basedbecauseoftherelative easeofcomparingandcategorizingdiscreteunitsdefinedbytranscriptionandtranslation,theim- portanceofnoncodingandrepetitivesequencesisunquestionable.Itwouldthereforebeextremely powerfultodefineacomprehensivesequence-basedpan-genomethatincludesinformationabout therelativepositionofallsequences.Unfortunately,interpretingnoncodingsequencevariationischallenging.Indeed,evenforclassesofnoncodingsequencesofknownimportance,e.g.,promot-",
+    "assessment  will improve our understanding of the reference to  better assemble and interpret future  genome sequences.     We have previously developed a method to assess the risk of a patient for 55 diseases using a  quantitative human disease -SNP association database, and showed that we could suggest useful  and clinical relevant information using his personal genome sequence  (16). Here, we queried  the  reference genome sequence against our databa se and  identified  3,556 disease -susceptib ility",
+    "The shortcomings of a single, linear reference genome per species are well appreciated, and richer reference datastructures are an active area of research (Church et al. 2015 ). An alternative is de novo assembly of the genomes of commonly used strains. The Sanger Mouse GenomesProject is using a combination of long-insert jumping libraries and optical mapping to build de novo assemblies",
+    "undertake comprehensive and  powerful explorations rather than  being confined to testing hypoth - eses focused on candidate path - ways. With the completion of the  first reference sequence of the  human genome,3 attention shifted  from searching for genes to dis - covering their functions. System - atic genetic mapping in families  and populations helped scientists  pinpoint the genetic variants that  contribute to human disease.",
+    "points, nding statistical associations, modeling and run ning predic- tors, or constructing and pruning networks of detected rela tions. In the following paragraphs I will explore these opportunities in detail. 1.4.1 Population reference genomes Genomes are relatively similar between individuals, there fore, instead of assembling the complete sequence for each person, we only de termine points of DNA variation compared to a reference genome. Subs equently,",
+    "having a reference genome for a related specie s certainly makes the process easier. The  availability of long-read sequences vastly improves our ability to assemble new genomes, and new  technologies, such as PacBio  and Nanopore, are now able to give reads between 100-1000  kilobases, an order of m agnitude longer than current Illumina sequencing (Shendure et al.  2017).  Combining these new technologies with traditional short read NGS will greatly improve our ability"
+  ],
+  "answer": "Using a pangenome-based reference can be more useful than using a single linear reference for several reasons:  1. **Capturing Genetic Diversity**: A single linear reference genome represents just one version of a species' DNA, which might not include all the genetic variations found in different individuals. A pangenome, on the other hand, includes multiple versions of genes and sequences from various individuals, capturing a broader spectrum of genetic diversity [3].  2. **Reducing Reference Bias**: When using a single linear reference, there's a risk of reference bias, which means that genetic variations not present in the reference genome might be missed or misinterpreted. A pangenome reduces this bias by including multiple genetic variants, making it more likely to accurately represent the genetic makeup of different individuals [1].  3. **Improving Disease Research**: By including a wider range of genetic variations, a pangenome can help researchers better understand the genetic basis of diseases. This comprehensive approach allows for more accurate identification of disease-associated genetic variants, which can lead to better diagnostics and treatments [6].  4. **Enhanced Analysis Tools**: The use of a pangenome can improve the performance of genetic analysis tools. For example, it can provide a more accurate benchmark for evaluating new sequencing technologies and analysis methods, leading to more reliable results [3].  5. **Overcoming Limitations of Single References**: A single reference genome assumes that the order and structure of genes are the same across all individuals, which is not always true. This assumption can lead to errors in genetic studies. A pangenome, by including multiple sequences, can account for these structural variations and provide a more accurate representation of the genome [2].  In summary, a pangenome-based reference is more useful because it captures the genetic diversity of a species, reduces reference bias, improves disease research, enhances analysis tools, and overcomes the limitations of using a single linear reference. This makes it a more comprehensive and accurate tool for genetic studies.",
+  "question": "Can you explain why using a pangenome-based reference might be more   useful than simply using a single linear reference? Please make you   answer accessible to a non-expert."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_31 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_31
new file mode 100644
index 0000000..82a1daf
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_31
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2016 - The Genomics of Type 1 Diabetes.pdf",
+    "2015 - Quantitative and logic modelling of molecular and gene networks.pdf",
+    "2016 - The Genomics of Type 1 Diabetes.pdf",
+    "2013 - Genetic and Genomic Approaches to Understanding Macrophage Identity and Function.pdf",
+    "2016 - The Genomics of Type 1 Diabetes.pdf",
+    "2010 - The Role of Epigenetics in the Pathology of Diabetic Complications.pdf",
+    "2018 - Molecular Brain Adaptations to Ethanol_ Role of Glycogen Synthase (2).pdf",
+    "2011 - The age of the \u201come\u201d Genome, transcriptome and proteome data set collection and analysis.pdf",
+    "2009 - Detection and interpretation of expression quantitative trait loci (eQTL).pdf",
+    "2005 - Part I Previous Research Track Record.pdf"
+  ],
+  "extraction_id": [
+    "24a08eeb-d72d-5ff6-97e3-d5f07795db7a",
+    "2ba86c45-9754-5300-8052-8b9c2765ecbc",
+    "cc42c6bf-d890-5a83-9598-b1a518f097b1",
+    "4c2afa3b-cf31-58ba-8ae8-2bf609f25dbc",
+    "24a08eeb-d72d-5ff6-97e3-d5f07795db7a",
+    "5c84ddde-f0cc-51fc-927c-0d5a11de0c07",
+    "ce57e949-0eeb-575a-9d67-5e21d08a47a2",
+    "772a0ef3-9be9-5b05-ad99-01c5c06347a0",
+    "a744cab4-ff8a-59e2-a789-ef7911c0a9ea",
+    "0853c5ab-3d98-565c-ba1f-50e5bd91d14c"
+  ],
+  "document_id": [
+    "4933cdc2-7d36-5181-87c9-63b58498839f",
+    "8db6a373-be03-5653-beaf-1b2ae1d98c31",
+    "4933cdc2-7d36-5181-87c9-63b58498839f",
+    "1526d201-2f4e-5e6c-b2c8-8c825e741401",
+    "4933cdc2-7d36-5181-87c9-63b58498839f",
+    "766edfd5-4756-51bf-b636-c94b041d030c",
+    "cc2690a9-5a87-5f09-87d5-115a6a6b8349",
+    "ca99ed69-ee09-5717-95ed-c26eefb5e42d",
+    "ef974b09-4ea2-5382-85e5-c2169f440fda",
+    "1875d68b-adeb-5f91-8a67-91d881906238"
+  ],
+  "id": [
+    "chatcmpl-ADZCAQL2r34gNZf3z3cSmtRoVHASI",
+    "1b7834f6-6ffa-5ebe-8b53-e590f43a13e0",
+    "93732536-96d1-5b06-a73b-06387cdfc139",
+    "4c7d9d92-5685-5560-8ec3-5c0df73a7a4f",
+    "6baacb55-1480-592e-9b6b-87c3e3980e19",
+    "8bb2457a-69e7-5e4a-9cc3-f51b8db47a89",
+    "8ddad342-ceca-56a1-afd6-6a3218436c1c",
+    "49fe2fc3-a95f-5de1-a261-bf596e48e6ff",
+    "0423eecb-319c-598b-a1d6-2ca3ecee84ce",
+    "b46d80d5-c968-5541-b658-d9a2dcd86e37",
+    "b9320635-76da-5a0c-a1fc-ea3a11cc7068"
+  ],
+  "contexts": [
+    "al., 2012 ; Hindhorff, 2009; Barrett et al ., 2007 ). Recent efforts by the Encyclopedia of DNA  elements (ENCODE) consortium, to characterise the human genome, have revealed that most  of the non -coding part of the genome is not inactive but is associated with  different forms of  regulatory activity (ENCODE, 2012 ; Thurman, 2012 ). One important regulatory process that  takes place within the genome is the (in-) activation  of gene expression through the interaction",
+    "network of transcriptional regulators. Nature 403,  335338 (2000). 18. Gardner,T ., Cantor,C. & Collins,J. Construction of a  genetic toggle switch in Escherichia coli. Nature 403,  339342 (2000). 19. Kauffman,S.A. Metabolic stability and epigenesis in  randomly constructed genetic nets. J.Theor. Biol. 22,  437467 (1969). 20. Thomas,R. Boolean formalization of genetic control  circuits. J.Theor. Biol. 42, 563585 (1973). REVIEWS NATURE REVIEWS | GENETICS   ADV ANCE ONLINE PUBLICATION | 11",
+    "25   2.8 REGULATION OF GENE EXPRESSION     Apart  from the protein coding sequences, there are other biologically relevant nucleic acid  sequences that play other important roles in the genome such as regulation of gene expression  and maintenance of the chromatin structure (Pique -Regis  et al., 2011). Regu lation of gene  expression involves  a process that leads to increase or decrease in the production of specific",
+    "expression is regulated at many levels, but gene transcription  represents an essential and, in many cases, dominant point of control. Protein-coding genes are transcribed from promoters,  which represent genomic regions that recruit basal transcrip- tion factors and RNA polymerase II. Physiological levels of gene expression and responses to internal and external signals require the actions of additional sequence-specific transcrip- tion factors that recruit nucleosome-remodeling complexes,",
+    "regulatory elements  and variants thereof that may affect gene expression particularly through  the binding of transcription factors (TFs) to DNA.   The suggestion that the  genetic determinants of complex diseases are perh aps better sought in  problems associated with gene regulation is due to findings that many of the disease associated  variants occur in non -coding DNA sequences within the genome  (ENCODE, 2012; Schuab et",
+    "through multiple cell divisions at the transcriptio nal and epigenetic level need to be more 204  carefully examined and have evolved as an exciting area of research. 205   206  Epigenetics and transcriptional regulation  207  Regulation of gene expression relies on the ac cessibility of DNA to various transcription 208  factors, co-activators/co-repressors, and the transcriptional machinery. DNA is first wrapped 209",
+    "post-translationally, translationally, transcriptionally, or epigenetically  (Lempradl   et al, 2015; Zong  et al, 2017) . It seems likely that these different layers of  regulation can operate cooperatively on different time- scales . More permanent   adaptations might be expected following persistent regulation on a more transient  levelfor example,  lowered transcriptional activity of a gene might follow  a  period of low functional  activity of its protein. Elucidating the means of such",
+    "important  component in the regulation of gene expression with between  10  and  20%  of  the  transcriptome  being  regulated  by  DNA variation. 2. Technologies The  study  of  DNA  and  its  downstream  effects  is  very  much a  technology  driven  process.  Most  of  the  rst  screens  looking  at DNA  changes in disease involved looking at segregation in fam- ilies  because  there  were  no  reasonable  technologies  at  the  time",
+    "the cytosine and adenine nucleotides[31]. In addition, the c hromosomal structure of DNA can be decondensated by histone acetylatio n (trans- fer of acetyl groups to DNA organizational elements), makin g it more accessible for transcription[87]. The transcriptional ex pression of genes is further regulated by genetic variants themselves[7]. Fi nally, proteins form a complex network of interactions[265] that, in turn, a lso regulate gene expression[331].",
+    "eterogeneity and common, small effect genetic variants will be assessed.  h  D (c) Regulatory Signals:  Co-regulation of genes via shared transcriptional networks provides the basis for context-dependent gene  expression, an understanding of which is vital to the understanding of disease etiology and disease progression. In  particular, transcription factors (TF) and their transcription factor binding sites (TFBS) provide a key component in the understanding of how co-regulation is achieved."
+  ],
+  "answer": "Not all genetic regulation is done solely through DNA elements like promoters, repressors, and activators. There are other forms of genetic regulation that play crucial roles in controlling gene expression. Here are some key points to consider:  1. **Non-Coding DNA**: Recent research has shown that much of the non-coding part of the genome, which does not code for proteins, is still active in regulating gene expression. This includes various regulatory activities that are not directly related to the traditional DNA elements like promoters and repressors [1].  2. **Chromatin Structure**: The structure of chromatin, which is the complex of DNA and proteins in the cell nucleus, plays a significant role in gene regulation. For example, histone acetylation, which involves adding acetyl groups to histone proteins, can decondense chromosomal structure and make DNA more accessible for transcription [9].  3. **Epigenetic Regulation**: Epigenetics involves changes in gene expression that do not alter the DNA sequence itself. This can include modifications like DNA methylation and histone modification, which affect how tightly DNA is wound around histones and thus its accessibility for transcription [6], [7].  4. **Post-Transcriptional Regulation**: After DNA is transcribed into RNA, there are additional layers of regulation. This includes processes like RNA splicing, editing, and degradation, which can influence how much of the RNA is available to be translated into protein [7].  5. **Translational and Post-Translational Regulation**: Even after RNA is translated into protein, there are mechanisms that regulate the activity, stability, and localization of proteins. These include modifications like phosphorylation and ubiquitination, which can alter protein function and lifespan [7].  In summary, while DNA elements like promoters, repressors, and activators are important for genetic regulation, there are multiple other layers of regulation involving chromatin structure, epigenetic modifications, and post-transcriptional and post-translational processes that also play critical roles in controlling gene expression.",
+  "question": "Is all genetic regulation done through DNA (e.g., prompters,   repressors, activators) or are there other forms of genetic regulation?   Please make you answer accessible to a non-expert."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_32 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_32
new file mode 100644
index 0000000..ee973b3
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_32
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2015 - Linking traits based on their shared molecular mechanisms.pdf",
+    "2015_GN_Diabets_notheses.pdf",
+    "2015 -Oren- Linking traits molecular mechanisms.pdf",
+    "2008 - Rutter_s child and adolescent psychiatry-Blackwell Pub (2008).pdf",
+    "2004 - Combining QTL and Microarray Data.pdf",
+    "2015 - Linking traits based on their shared molecular mechanisms.pdf",
+    "2015_GN_Diabets_notheses.pdf",
+    "2015 -Oren- Linking traits molecular mechanisms.pdf",
+    "2012 - Needs Analysis of Genetics and Genomics in Communication Sciences and Disorders.pdf",
+    "2007 - An Informatics Approach to Systems Neurogenetics.pdf"
+  ],
+  "extraction_id": [
+    "d54f1d09-1606-5cc3-92b7-e68d46a9f705",
+    "089366c5-7d36-5621-b463-6d1ad16d98cd",
+    "d1b6294a-3096-5385-8ae1-cf229e122f83",
+    "248fc08f-6f5e-5367-97dd-bdab0ca49699",
+    "d0be4c6e-9821-53e5-a770-3a68a06d4d84",
+    "6060c5cc-b34f-57ee-89e3-266cc660aa7e",
+    "0f901755-7d06-5722-8d57-49b7da4bc35f",
+    "65feb87b-411c-5835-9bb4-d6b14115f86d",
+    "2a443a7a-f71c-5a52-adda-085fa9199b40",
+    "6e361ba5-ecb0-51e6-b178-6f244743383f"
+  ],
+  "document_id": [
+    "162160c3-1395-58da-98bd-b8450ec923d6",
+    "1744d9bf-29f9-52e3-a7c9-62a916999cda",
+    "ebf5e07f-4b24-5955-a2a6-fc8b9d5b5904",
+    "59daba11-206e-5bbc-8833-9d1b661532b5",
+    "5aeba67e-2338-5add-b8ee-ea304020834b",
+    "162160c3-1395-58da-98bd-b8450ec923d6",
+    "1744d9bf-29f9-52e3-a7c9-62a916999cda",
+    "ebf5e07f-4b24-5955-a2a6-fc8b9d5b5904",
+    "c8a76cb1-506d-57e4-a18e-548e777898e2",
+    "4d0692d8-74ef-58a3-82ec-d198b1d1c758"
+  ],
+  "id": [
+    "chatcmpl-ADZCLib2EWT6WxKcrpOOYTqqeufBy",
+    "0c2422d5-9509-58b6-9ae2-615d4ee0ddee",
+    "359c496d-0870-5f54-82bd-3ce31e12863f",
+    "8e339342-5a42-5332-8c7d-3cea19e5f11b",
+    "bfe0accf-3aa8-5d95-97ce-6761e88c526a",
+    "edf0249b-5a8b-5050-b8fb-56a8304cbb23",
+    "651a0790-ec4f-5615-8aaa-90293e45ae42",
+    "b457363b-69ea-5b9e-9a48-06ae89034def",
+    "59e00799-df5e-52c0-882d-5c1eefc74e8b",
+    "1a01fece-3b4b-5b36-b994-e0fe945bdbf9",
+    "aa0a5df1-8084-579d-9d31-40d3bc9bee4d"
+  ],
+  "contexts": [
+    "3, 4 and 5 suggest previously unknown connections between traits. We next characterized pairs of traits within each group of traits (trait pairs) to show that the quality of these pairs is not lower than in existing methods. We focused on three main properties of trait pairs: the correlation among traits in a pair; the correlation between a trait pair and the",
+    "3, 4 and 5 suggest previously unknown connections between traits. We next characterized pairs of traits within each group of traits (trait pairs) to show that the quality of these pairs is not lower than in existing methods. We focused on three main properties of trait pairs: the correlation among traits in a pair; the correlation between a trait pair and the",
+    "3, 4 and 5 suggest previously unknown connections between traits. We next characterized pairs of traits within each group of traits (trait pairs) to show that the quality of these pairs is not lower than in existing methods. We focused on three main properties of trait pairs: the correlation among traits in a pair; the correlation between a trait pair and the",
+    "taxonomy of traits is that it allows researchers to turn theirattention to the ways temperament and personality traitsexpress themselves in daily life and to the fundamental pro-cesses underlying variations in these traits. In this section, we rst describe the traits and then review some of the mostinteresting current work on the psychological and evolutionaryunderpinnings of each trait. A more detailed description of thecomponents of these traits is found in Caspi and Shiner (2006).Because relatively less",
+    "ditions and related totraits ofinter est,often bycomparing two groups differing forthetrait. Darvasi (2003) states that thereisanundeclar eddispute among resear chers who study complex traits :::Onone side areclassical geneticists :::ontheother areproponents ofgene expr ession analysis :::.Darvasi goes ontooutline thepossible advantages ofcombining these techniques over and above either technique alone. Inaddition tobetter correlating ge-",
+    "three types of high-order organization of traits. (i) Groups of tightly related traits that share thesame transcripts mechanisms (modules 1, 2, 6, 7, 8, e.g., Figure 3 ). (ii) Groups of distinct traits that share the same transcripts mechanism, but not necessarily high correlations among them (modules 3, 4, 5, e.g., Figure 4 ). (iii) Different groups commonly have overlapping traits, but typically differ in their underlying mechanisms ( Figure 2B ).",
+    "three types of high-order organization of traits. (i) Groups of tightly related traits that share thesame transcripts mechanisms (modules 1, 2, 6, 7, 8, e.g., Figure 3 ). (ii) Groups of distinct traits that share the same transcripts mechanism, but not necessarily high correlations among them (modules 3, 4, 5, e.g., Figure 4 ). (iii) Different groups commonly have overlapping traits, but typically differ in their underlying mechanisms ( Figure 2B ).",
+    "three types of high-order organization of traits. (i) Groups of tightly related traits that share thesame transcripts mechanisms (modules 1, 2, 6, 7, 8, e.g., Figure 3 ). (ii) Groups of distinct traits that share the same transcripts mechanism, but not necessarily high correlations among them (modules 3, 4, 5, e.g., Figure 4 ). (iii) Different groups commonly have overlapping traits, but typically differ in their underlying mechanisms ( Figure 2B ).",
+    "of varying effect sizes (small to moderate), interact with each other across time to manifest as individual genotypic and phenotypic traits. These traits contribute to normal variation in human behavior. Yet, these trait variants also increase the susceptibility of a disorder or a condition for many others.",
+    "action will open a Correlation Plot page in which you can examine the relationship between the two traits. Look for linearity and outliers. 3.3.1. Selection and Saving Multiple Traits The list of traits on the Correlation Results page represents traits that may be related in some way. You may want to select a group of them for further analysis. For example, use the checkboxes to the left of each entry to check entries 1, 9, 10, 14, 16, 18, traits related to brain size. Click the Add to collection"
+  ],
+  "answer": "The different relationships between traits can be categorized into several types based on the provided context:  1. **Correlation Among Traits in a Pair**: This refers to how traits within a pair are related to each other in terms of their correlation [1], [2], [3].  2. **Correlation Between a Trait Pair and Other Factors**: This involves examining how a pair of traits correlates with other external factors or conditions [1], [2], [3].  3. **High-Order Organization of Traits**:    - **Groups of Tightly Related Traits**: These are traits that share the same transcript mechanisms and are highly correlated with each other (modules 1, 2, 6, 7, 8) [6], [7], [8].    - **Groups of Distinct Traits with Shared Mechanisms**: These traits share the same transcript mechanisms but do not necessarily have high correlations among themselves (modules 3, 4, 5) [6], [7], [8].    - **Overlapping Traits in Different Groups**: Different groups of traits may have overlapping traits but typically differ in their underlying mechanisms [6], [7], [8].  These relationships highlight the complexity and interconnectedness of traits, showing that they can be related through direct correlations, shared mechanisms, or overlapping characteristics.",
+  "question": "What are the different relationship between traits?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_33 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_33
new file mode 100644
index 0000000..f781f08
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_33
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2014 - A Population Genetic Signal of Polygenic Adaptation.pdf",
+    "2021 - Correlational selection in the age of genomics.pdf",
+    "2012 - Functional genomics research in aquaculture principles and general approaches.pdf",
+    "2011 - Genetical genomics approaches for systems genetics.pdf",
+    "2020 - A Multi-Omics Perspective of Quantitative Trait Loci in Precision Medicine.pdf",
+    "2010 - Systems genetics, bioinformatics and eQTL mapping.pdf",
+    "2022 -Chunduri- Drugs Animal Models.pdf",
+    "2022 - New Insights on Gene by Environmental Effects of Drugs of Abuse in Animal Models Using GN.pdf",
+    "2022 - New Insights on Gene by Environmental Effects of Drugs of Abuse in Animal Models Using GeneNetwork.pdf",
+    "2016 - Mouse genome-wide association and systems genetics identifies Lhfp as a regulator of bone mass.pdf"
+  ],
+  "extraction_id": [
+    "c28f56f2-4e3d-5c8c-afcc-c6ac1dc43074",
+    "2aea6aad-eaf7-5e30-b505-4c08b47a8e98",
+    "0c3d0cb3-d4b0-5655-8b04-285a87710636",
+    "ea23303c-d909-5bda-9a48-8c78fb60cf8c",
+    "2d44caa2-d625-5252-87a1-a9691af99e36",
+    "298ee1f5-58a9-567c-86ba-8ac5967e1718",
+    "3df0d755-b4aa-5635-a223-3bc6d454a196",
+    "d01794ca-a660-5319-af06-8f0b9ee8e060",
+    "407f64ca-3b4b-57b8-954c-b5a58132d458",
+    "2cecc2f8-8211-544f-88e7-23e270d34f63"
+  ],
+  "document_id": [
+    "5760b25c-236b-527d-98d6-563a85888727",
+    "5449975c-261a-5e45-a979-04fad61cefd8",
+    "a39b4cc1-8661-578b-a61b-b9962e45fc33",
+    "de78a01d-8d03-5afb-af5b-ce2ed2167766",
+    "8503b166-b917-5efb-a356-5ba371504cc1",
+    "27c922c6-e449-5f83-868a-3ad7284facc8",
+    "9cfa4f4c-37ce-5c0f-9da6-3bbb075fdc45",
+    "6f5d0c5b-0bbb-5eca-9e3e-73c3b0675472",
+    "d71efa0d-5de8-549c-964d-489ef6b73a1f",
+    "a554412b-b074-5bcd-9617-06ea69647b8a"
+  ],
+  "id": [
+    "chatcmpl-ADZCTjdARUSr934Zl60dbSl3iWvA2",
+    "75b9b0fa-38e8-5674-8000-ae14f26a1780",
+    "1641cea6-8773-516e-b08e-fad820ebfdb9",
+    "fa1981fe-6730-59a1-b331-c6c7250b0f2c",
+    "72b37b21-1d41-55bd-b835-f0bd267a3970",
+    "108483cf-404b-5a9c-bf1f-be58ebf6d16d",
+    "68a13597-c223-54d9-9664-604d69b97c50",
+    "ed4ddc1b-45f9-5d9c-8969-e881d96edc4e",
+    "73b8e482-b204-5da6-b92d-f090efb622f1",
+    "60a84952-41ed-57ee-b689-6da313793843",
+    "616a41e7-df46-54d5-979e-1654973aa642"
+  ],
+  "contexts": [
+    "ST, see [40,120122]). Such tools may also offer a way of incorporating GxE interactions, as multiple GWAS for the same trait in different environments can be treated as correlatedtraits [123]. As association data for a greater variety of populations, species, and traits becomes available, we view the methods described outhere as a productive way forward in developing a quantitativeframework to explore the genetic and phenotypic basis of local adaptation. Materials and Methods",
+    "has been achieved by quantitative trait loci mapping, admixture  mapping and GW AS131, which have limited power to detect  small-effect-size genes. Newer approaches map pleiotropy by simultaneously associating genomic loci with multiple traits 54  and can also detect epistatic interactions using machine learning algorithms 132.Detecting the genomic signatures of correlational selectionCorrelational selection could potentially be inferred from  signatures of selective sweeps at loci under strong selection",
+    "pairs that include many genes within the seg- ment. On the other hand, GWAS may point to several or even many genomic locations for the trait of interest, complicating further functional analysis. Analysis of Quantitative Trait Loci (QTL) QTL analysis reveals statistically signicant linkage between phenotypes and genotypes, thereby providing explanation for the genetic basis of variation in complex traits (Falconer and Mackay, 1996; Lynch and Walsh, 1998). In a sense, QTL analysis can be viewed as incom-",
+    "studies.    There are  many possible causal networks even in a simple syst em consisting of  a genomic locus (QTL) and two traits, T1 and T2 ( Figure 1 ). Causal inference in  GWLS and GWAS involves, in its simplest form, the i dentification of pairs of traits  with a common QTL (QTL-trait-trait triads) and dete rmining whether the QTL  directly affects each of two traits (independent), or if the QTL affects only one trait",
+    "tions by matching patterns of expression QTL and GWAS. Am. J. Hum. Genet. 92, 92 160. Giambartolomei, C. et al. (2014) Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS Genet. 10, e1004383 161. Porcu, E. et al. (2019) Mendelian randomization integrating GWAS and eQTL data reveals genetic determinants of com-plex and clinical traits. Nat. Commun. 10, 3300 162. Zhu, Z. et al. (2016) Integration of summary data from GWAS",
+    "knowledge of the true QTL location (Doss et al. 2005 ), which can be used to empirically estimate the power of aGWAS performed at a similar scale (Hao et al. 2008 ; Schadt et al. 2008 ). A GWAS on its own does little more than establish correlations between changes in DNA at agiven locus and changes in a disease trait of interest, with respect to populations of interest. Further, these studies on",
+    "Another method to identify candidate genes is to leverage data generated in another population or species. Phenome-wide association studies (PheWAS) take a gene or variant of interest and nd all reported associations in GWAS datasets. A number of these GWAS tools exist, using either different methods, or different human cohorts (https://atlas.ctglab. nl/PheWAS, http://pheweb.sph.umich.edu/, accessed on 2 February 2022). Mouse QTL mapping has high power but low precision (i.e., we can detect a QTL, but",
+    "Another method to identify candidate genes is to leverage data generated in another population or species. Phenome-wide association studies (PheWAS) take a gene or variant of interest and nd all reported associations in GWAS datasets. A number of these GWAS tools exist, using either different methods, or different human cohorts (https://atlas.ctglab. nl/PheWAS, http://pheweb.sph.umich.edu/, accessed on 2 February 2022). Mouse QTL mapping has high power but low precision (i.e., we can detect a QTL, but",
+    "Another method to identify candidate genes is to leverage data generated in another population or species. Phenome-wide association studies (PheWAS) take a gene or variant of interest and nd all reported associations in GWAS datasets. A number of these GWAS tools exist, using either different methods, or different human cohorts (https://atlas.ctglab. nl/PheWAS, http://pheweb.sph.umich.edu/, accessed on 2 February 2022). Mouse QTL mapping has high power but low precision (i.e., we can detect a QTL, but",
+    "narrow regions ofthegenome harboring trait associated genetic variants. Itisstill, however, a challenge toidentify causal genes and several approaches have been developed that canassist inbridging thisgap. Specifically, systems genetics approaches involving theintegration of other types of-omics data have proven useful [25]. Two systems genetics approaches for informing GWAS areexpression quantitative trait loci(eQTL) discovery and co-expression"
+  ],
+  "answer": "Yes, the landscape of QTL (Quantitative Trait Loci) and GWAS (Genome-Wide Association Studies) hits can be used to find relationships between traits. This can be achieved through several methods:  1. **Correlated Traits in Different Environments**: Multiple GWAS for the same trait in different environments can be treated as correlated traits, which helps in exploring the genetic and phenotypic basis of local adaptation [1].  2. **Mapping Pleiotropy**: Newer approaches map pleiotropy by simultaneously associating genomic loci with multiple traits, which can reveal relationships between traits [2].  3. **QTL-Trait-Trait Triads**: Causal inference in GWAS and QTL studies involves identifying pairs of traits with a common QTL and determining whether the QTL directly affects each of the two traits independently or if it affects only one trait, which then influences the other [4].  4. **Colocalization and Integration of Data**: Methods such as Bayesian tests for colocalization between pairs of genetic association studies using summary statistics, and Mendelian randomization integrating GWAS and eQTL data, can reveal genetic determinants of complex and clinical traits, thereby identifying relationships between traits [5].  These methods collectively demonstrate that the landscape of QTL and GWAS hits can indeed be used to find relationships between traits.",
+  "question": "Can landscape of QTL and GWAS hits be used to find relationships between traits?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_4 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_4
new file mode 100644
index 0000000..3fa0eae
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_4
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - Opening Up the Conversation on Genetics.pdf",
+    "2009 - Opening Up the Conversation on Genetics.pdf",
+    "2013 - ACMG recommendations for reporting of incidental findings.pdf",
+    "2008 - Genetic and Genomic Healthcare Ethical Issues of Importance to Nurses.pdf",
+    "2009 - From Genetics to Genomics Ethics, Policy, and Parental.pdf",
+    "2009 - Opening Up the Conversation on Genetics.pdf",
+    "2020 - Informed Consent for Genetic and Genomics Research.pdf",
+    "2008 - Canada Public Health Genomics.pdf",
+    "2018 - Ethical_Social_and_Legal_Consequences.pdf",
+    "2009 - From Genetics to Genomics Ethics, Policy, and Parental.pdf"
+  ],
+  "extraction_id": [
+    "92f0e5a8-a086-5236-962d-9b11584c65f4",
+    "92f0e5a8-a086-5236-962d-9b11584c65f4",
+    "1c2f1490-e98b-5ff8-8af0-edda9d5f3993",
+    "b7cdc9f8-d2b3-5ba3-a15f-6ef8d0c4f398",
+    "fd6ec3a0-cd63-5a6c-915b-7560fee0206e",
+    "93dd42f5-abb6-5a66-975f-1c1056c57173",
+    "c6ea7f08-9005-547c-b7a2-8011d7b67221",
+    "94557cbe-8255-53c3-8f56-5ea21b1075c6",
+    "8c1a83eb-78c2-56ba-b572-53753ea5bf2d",
+    "a7b4b523-5ec0-5212-80d7-2796d2799e44"
+  ],
+  "document_id": [
+    "b62a8f54-c2f5-5bbb-9324-af80f7537167",
+    "b62a8f54-c2f5-5bbb-9324-af80f7537167",
+    "ef8364d6-d413-5150-8ad5-034a62bf787e",
+    "5e5322c1-f77d-5f71-8b1d-1a6eb0b9744a",
+    "4e65d14b-5c3f-547f-909d-d07064e59a47",
+    "b62a8f54-c2f5-5bbb-9324-af80f7537167",
+    "05a8ff6b-0de7-53af-a403-79eb2a9e9b33",
+    "ed94a341-3cf7-529a-827b-3ef681ce72c4",
+    "3fb80410-0b56-5c01-b3d6-9388b6029a77",
+    "4e65d14b-5c3f-547f-909d-d07064e59a47"
+  ],
+  "id": [
+    "chatcmpl-ADZ9d7iHwaIjxfj8twd8wFGyZBbWb",
+    "45c7b32c-b032-5031-9e74-9a50fb63543f",
+    "da82c453-a630-5708-8b8a-93c6cd5e4ce4",
+    "9ebb7f76-4039-541f-b5d6-0bf16b3b1f35",
+    "9700655c-04a3-58ed-af01-79629c936f2d",
+    "96371003-643f-5d58-ba56-73dadd8fe812",
+    "f536f809-6b9a-5602-8915-5a47210a361b",
+    "838f60f9-1253-58a5-9cf6-36568f0cf07f",
+    "fcd27256-ac2e-5fbb-bf71-5357836e5e02",
+    "1f807579-9f8c-53e6-a35d-8d426024b71a",
+    "73dcb25d-3f7a-50ec-a0a4-b27669015092"
+  ],
+  "contexts": [
+    "1999) raises practical and ethical issues of access to resulting opportunities and creates family communication challenges. Currently, prenatal testing for chromosomal diseases has become increasingly common (Moyer et al., 1999). Options such as pre-implantation genetic diagnosis (PGD) can identify over 1,250 disease-related mutations creating an opportunity for parents to select unaffected embryos for implantation in the womb (R. M. Green, 2008). Test results provide potential parents with information",
+    "undergo prenatal testing have determined that partners base their decision upon several factors, including, but not limited to: parental beliefs about abor-tion, attitudes regarding disability and their perceptions of the usefulness of having the information revealed by genetic tests (Moyer et al., 1999, p. 522). Abortion beliefs constitute a key issue in the decision-making process. Even though a majority of parents receiving abnormal prenatal test results terminate their pregnancies (Redlinger-Grosse,",
+    "Hum Genet 1995;57:12331241. 24. Committee on Bioethics. Ethical and policy issues in genetic testing and  screening of children. Pediatrics 2013;131:620622. 25. Ross LF, Saal HM, David KL, Anderson RR. Technical report: ethical and policy issues in genetic testing and screening of children. Genet Med 2013;15:   234245. 26. Wilfond B, Ross LF. From genetics to genomics: ethics, policy, and parental decision-making. J Pediatr Psychol 2009;34:639647.",
+    "Informed Consent and Genetic Testing    Genetic testing is increasingly used across the life continuum  for screening, diagnosis, and de termining the best treatment  of diseases. Obstetric and pediat ric nurses have traditionally  been involved in the genetic testing process with prenatal  screening for genetic conditions such as spina bifida and Down  syndrome, and newborn screening for genetic conditions such",
+    "Objective Ethical evaluation of genetic testing in children is traditionally based on balancing clinical benefits and risks. However, this focus can be inconsistent with the general practice of respecting parentaldecision-making about their childrens health care. We argue that respect for parental decision-making should play a larger role in shaping pediatric genetic testing practices, and play a similar role regarding decisions",
+    "prenatal decisions. Further research needs to investigate how different families engage in such discussions and decision-making pro-cesses, especially as prenatal testing becomes more common and better able to predict or prevent a wider range of genetic conditions.",
+    "all of the complex ethical and legal issues rel- evant to genetic testing would disappear if there were effective preventions or treatments available for genetic conditions. The ability to predict future disease in conjunction with a limited ability to do much about it has im- portant social and psychological implications that must be addressed in conducting genetic research. One final factor worth consideration in un- derstandingthesensitivitytogeneticmedicine",
+    "Newborn screening by tandem mass spec-trometry: ethical and social issues. Can J Public Health 2007;    98:   284286.   65 Belle-Isle L: Genetic testing for late onset dis- eases: a population and public health per-spective. Health Policy Res Bull 2001;    1:   11 12.   66 Williams-Jones B: Private genetic testing in  Canada: a summary. Health Law Rev 2001;   9:    1013.   67 Begleiter ML: Training for genetic counsel- lors. Nat Rev Genet 2002;   3:   557561.   68 Carroll JC, Reid AJ, Woodward CA, Per-",
+    "Although risk-based genetic testing for common diseases raise similar ethical issues to more  traditional genetic testing for rare diseases, new challenges are raised due to the type of information  revealed and access to tests. With thoughtful deliberation with health professionals, patients and  families, test developers and laboratories, insurers and other stakeholders, these issues can be  addressed to ensure the safe and appropriate use of these promising new clinical applications.     REFERENCES",
+    "against testing, parents should generally be given final decision-making authority. Ethical Considerations in Developing Policy for Comprehensive Genomic Testing In the near future, genomic testing is likely to become more accessible and will provide both information aboutthe risks of common conditions such as heart disease, diabetes, and hypertension as well as predictions aboutindividual responses to specific pharmaceuticals and other medical therapies (Aspinall & Hamermesh, 2007)."
+  ],
+  "answer": "The ethical considerations surrounding prenatal genetic testing and the selective termination of pregnancies based on genetic factors are multifaceted and complex. Key considerations include:  1. **Parental Beliefs and Attitudes**: Decisions about prenatal testing and subsequent actions, such as termination, are heavily influenced by parental beliefs about abortion, attitudes regarding disability, and their perceptions of the usefulness of the information revealed by genetic tests [2]. These personal beliefs and attitudes play a crucial role in shaping the ethical landscape of prenatal genetic testing.  2. **Informed Consent**: Ensuring that parents are fully informed about the implications of genetic testing is critical. This includes understanding the potential outcomes, the accuracy of the tests, and the possible decisions they may face based on the results [4]. Informed consent is a cornerstone of ethical medical practice, particularly in the context of genetic testing.  3. **Respect for Parental Decision-Making**: Ethical considerations must balance clinical benefits and risks with respect for parental decision-making. Parents should generally have the final authority in making decisions about their children's health care, including whether to undergo genetic testing and how to respond to the results [5].  4. **Social and Psychological Implications**: The ability to predict future diseases through genetic testing, coupled with limited options for prevention or treatment, has significant social and psychological implications. These must be addressed to ensure that parents are not unduly burdened by the information provided by genetic tests [7].  5. **Access and Equity**: Ethical issues also arise from the accessibility of genetic testing and the opportunities it creates. There are concerns about equitable access to these technologies and the potential for disparities in who can benefit from them [1].  6. **Family Communication Challenges**: Genetic testing results can create communication challenges within families, as they navigate the complex information and make decisions that affect their future [1].  In summary, the ethical considerations surrounding prenatal genetic testing and selective termination involve respecting parental beliefs and decision-making, ensuring informed consent, addressing social and psychological impacts, and promoting equitable access to genetic testing technologies. These considerations must be carefully balanced to navigate the ethical complexities of prenatal genetic testing.",
+  "question": "What are the ethical considerations surrounding prenatal genetic testing and the selective termination of pregnancies based on genetic factors?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_5 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_5
new file mode 100644
index 0000000..3407a58
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_5
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2020 - Precision and Personalized Medicine How Genomic.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2015 -Pandey- Functional Analysis of Genomic Variation and Impact on Molecular.pdf",
+    "2015 - Functional Analysis of Genomic Variation and Impact on Molecular and Higher Order Phenotypes.pdf",
+    "2017 - Infection control in the new age of genomic epidemiology.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2016 - Coming of age ten years of next.pdf",
+    "2008 - Gene Expression Profiling.pdf"
+  ],
+  "extraction_id": [
+    "de09f30d-e9ba-5379-8c7a-85b2cd2ed6c8",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "f4e989e5-c3d0-5d5c-b8c3-95894a14630b",
+    "fa426831-7c04-56c1-a191-1ebbc35342ed",
+    "04f06fb6-b2ff-57d4-bac0-de5cf4782ff3",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "4cdf13c0-c505-5ff9-9a6e-b10e5d1c8819",
+    "5da5fc5d-1fe6-58f0-9141-72b9b2996fff",
+    "cad1dcca-621d-5003-ba3a-81950819bc52",
+    "c14d1c74-a14a-5037-8d3f-f32a60faa9a5"
+  ],
+  "document_id": [
+    "cd11028a-933b-52a0-9534-c173323056ef",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "93381e23-494b-5bc2-9d09-ef315506601f",
+    "263d327b-f5db-54e4-a215-b3f8a51cd7d6",
+    "8f028916-b990-5e95-b2a6-e69f451cc291",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "9dd6e4e9-d136-507b-b628-68c8e1461bd0",
+    "59f3b969-089b-5258-93ad-892dbc9ffa9c"
+  ],
+  "id": [
+    "chatcmpl-ADZ9nZqF5q344Dbv0zjyqijEIpDdi",
+    "8477a774-dddb-5541-b8d7-d51a7e56b0af",
+    "13a6b6f9-4a9a-5eb9-ac79-d986d9e613f0",
+    "d1158643-3625-5855-a03d-eec4ac96eb4d",
+    "cdf2b80f-1509-50a2-9cb2-a36dd6f3f2cc",
+    "f8ae01ae-cea8-5b8b-95c0-7147055de596",
+    "d2540614-9397-5e3e-8b5f-ad328ca973b2",
+    "199e1929-dc7c-58d4-8c8d-1c931e658e9c",
+    "1e324977-2ca5-5062-8a09-7659d516e899",
+    "98010acc-fd11-5d33-bced-626ef29f2896",
+    "3e782f01-a06e-51b6-ac8a-0e0a56939d08"
+  ],
+  "contexts": [
+    "36. Sequencing, H.G. Finishing the euchromatic sequence of the human genome. Nature 2004 ,431, 931945. 37. Heather, J.M.; Chain, B. The sequence of sequencers: The history of sequencing DNA. Genomics 2016 ,107, 18. [CrossRef] 38. Rothberg, J.M.; Leamon, J.H. The development and impact of 454 sequencing. Nat. Biotechnol. 2008 ,26, 11171124. [CrossRef] [PubMed] 39. Shendure, J.; Ji, H. Next-generation DNA sequencing. Nat. Biotechnol. 2008 ,26, 11351145. [CrossRef] [PubMed]",
+    "22. Karow, J. Qiagen launches GeneReader NGS System  atAMP; presents performance evaluation by broad.  GenomeWeb  [online], https:// www.genomeweb.com/ molecular-diagnostics/qiagen-launches-genereader- ngs-system-amp-presents-performance-evaluation   (4Nov 2015). 23. Smith,D.R. & McKernan,K. Methods of producing  and sequencing modified polynucleotides . US Patent  8058030 (2011). 24. Margulies,M. etal.  Genome sequencing in  microfabricated high-density picolitre reactors. Nature  437, 376380 (2005).",
+    "11 BIOINFORMATIC CHALLENGES FOR GENOMIC MEDICINE      Processing and managing of high-throughput sequence data     High throughput sequencing offers severa l advantages relative to array-based  genotyping or expression assays. First, unlike genotyping arrays, whole genome sequencing is not limited to interrogating onl y known sequence variants. Similarly, RNA- sequencing (RNA-seq) enables expression quanti fication of novel transcripts that are not",
+    "11 BIOINFORMATIC CHALLENGES FOR GENOMIC MEDICINE      Processing and managing of high-throughput sequence data     High throughput sequencing offers severa l advantages relative to array-based  genotyping or expression assays. First, unlike genotyping arrays, whole genome sequencing is not limited to interrogating onl y known sequence variants. Similarly, RNA- sequencing (RNA-seq) enables expression quanti fication of novel transcripts that are not",
+    "High-throughput bacterial genome sequencing: an embarrassment of choice, aworldof opportunity.NatRevMicrobiol2012;10:599-606. 11.CroucherNJ,DidelotX.Theapplicationof genomicstotracingbacterialpathogen transmission.CurrOpinMicrobiol2015;23:62-7. 12.ShendureJ,JiH.Next-generationDNAsequencing.NatBiotechnol2008;26:1135- 45. 13.MillerJR,KorenS,SuttonG.Assemblyalgorithmsfornext-generationsequencing data.Genomics2010;95:315-27. 14.OlsonND,LundSP,ColmanRE,FosterJT,SahlJW,SchuppJM,etal.Bestpractices",
+    "sequencing. Genome Res. 20, 11651173 (2010). 64. English,A.C. etal.  Assessing structural variation in a  personal genome-towards a human reference diploid  genome. BMC Genomics 16, 286 (2015). 65. Carneiro,M.O. etal.  Pacific Biosciences sequencing  technology for genotyping and variation discovery in  human data. BMC Genomics 13, 375 (2012). 66. Quail,M.A. etal.  A tale of three next generation  sequencing platforms: comparison of Ion T orrent,  Pacific Biosciences and Illumina MiSeq sequencers.",
+    "Nat. Biotechnol. 30, 10331036 (2012). 111. Chrystoja,C.C. & Diamandis,E.P . Whole genome  sequencing as a diagnostic test: challenges and  opportunities. Clin. Chem. 60, 724733 (2014). 112. McGuire,A.L. etal.  Point-counterpoint. Ethics and  genomic incidental findings. Science 340, 10471048  (2013). 113. Bowers,J. etal.  Virtual terminator nucleotides for  next-generation DNA sequencing. Nat. Methods 6,  593595 (2009). 114. Heger,M. Chinas Direct Genomics unveils new",
+    "sequencing. Bioinformatics 31, 20402042 (2015). 46. Qiagen.  Oncology insights enabled by knowledge base- guided panel design and the seamless workflow of the  GeneReader NGS system  Press Release. Qiagen   [online], http://www.genereaderngs.com/PROM-9192- 001_1100403_WP_GeneReader_NGS_0116_NA.pdf   (2016). 47. Forgetta,V. etal.  Sequencing of the Dutch elm disease  fungus genome using the Roche/454 GS-FLX Titanium  System in a comparison of multiple genomics core",
+    "FURTHER INFORMATION 10X Genomics: http://www.10xgenomics.com 454 Sequencing: http://www.454.com Advances in Genome Biology and Technology (AGBT):  http://www.agbt.org BGISEQ500: http://seq500.com/en/portal/Sequencer.shtml Illumina: http://www.illumina.com Ion Torrent: https://www.thermofisher.com/us/en/home/ brands/ion-torrent.html Oxford Nanopore Technologies: https://www.nanoporetech. com Pacific Biosciences: http://www.pacb.com Personal Genome Project: http://www.personalgenomes.org",
+    "DNA), and provide the means to link sequences containing applications. First,  base- callers like Phred (4,5) extract raw sequences from raw data. There are  also contig assemblers like Phrap (University of Washington, http://bozeman. mbt.washington.edu/phrap.docs/phrap.html ) or CAP3 (6) that assemble frag- ments to contigs and packages like consed (7) or GAP4 (8), which are used to  finish sequencing projects. These programs are not explained in detail here."
+  ],
+  "answer": "### How-to Guide for Genetic Sequencing  Genetic sequencing is a powerful tool used to determine the order of nucleotides in DNA. This guide will walk you through the basic steps involved in genetic sequencing, referencing key resources for further information.  #### Step 1: Sample Preparation 1. **Collect DNA Sample**: Obtain a sample from the organism or tissue of interest. This could be blood, saliva, or tissue biopsy. 2. **Extract DNA**: Use a DNA extraction kit to isolate DNA from the sample. This involves cell lysis, removal of proteins and other contaminants, and purification of DNA.  #### Step 2: Library Preparation 1. **Fragment DNA**: Break the DNA into smaller fragments. This can be done using mechanical shearing or enzymatic digestion. 2. **Add Adapters**: Ligate specific adapter sequences to the ends of the DNA fragments. These adapters are necessary for the fragments to bind to the sequencing platform.  #### Step 3: Amplification 1. **PCR Amplification**: Amplify the DNA fragments using Polymerase Chain Reaction (PCR). This step increases the quantity of DNA to ensure there is enough material for sequencing.  #### Step 4: Sequencing 1. **Choose a Sequencing Platform**: Select an appropriate sequencing technology based on your needs. Common platforms include Illumina, Ion Torrent, and Pacific Biosciences [6], [9]. 2. **Load DNA onto the Sequencer**: Follow the manufacturer's instructions to load your prepared DNA library onto the sequencing machine. 3. **Run the Sequencer**: Initiate the sequencing run. The machine will read the nucleotide sequences of the DNA fragments and generate raw sequence data.  #### Step 5: Data Analysis 1. **Base Calling**: Use software like Phred to convert raw data into nucleotide sequences [10]. 2. **Sequence Assembly**: Assemble the short DNA sequences into longer contiguous sequences (contigs) using tools like Phrap or CAP3 [10]. 3. **Variant Calling**: Identify genetic variants by comparing the assembled sequences to a reference genome. This can be done using various bioinformatics tools.  #### Step 6: Interpretation 1. **Annotation**: Annotate the sequences to identify genes, regulatory elements, and other functional regions. 2. **Data Management**: Store and manage the high-throughput sequence data efficiently. This is crucial for handling large datasets generated by next-generation sequencing technologies [3], [4].  #### Additional Resources - **Next-Generation Sequencing (NGS)**: For a detailed understanding of NGS technologies and their applications, refer to Shendure and Ji's review on next-generation DNA sequencing [1], [5]. - **Sequencing Platforms**: For comparisons of different sequencing platforms, see the work by Quail et al. [6]. - **Bioinformatics Tools**: Explore various bioinformatics tools and software for sequence analysis and data management [10].  By following these steps and utilizing the referenced resources, you can successfully perform genetic sequencing and analyze the resulting data.",
+  "question": "Create a how-to guide for genetic sequencing."
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_6 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_6
new file mode 100644
index 0000000..ca72887
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_6
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2017 - Machine Learning and Data Mining Methods in Diabetes Research.pdf",
+    "2018 - Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls.pdf",
+    "2011 - Type 1 Diabetes Etiology, Immunology.pdf",
+    "2011 - Type 1 Diabetes Etiology, Immunology.pdf",
+    "2015 -precision-medicine-for-managing-diabetes.pdf",
+    "2016 - The Genomics of Type 1 Diabetes.pdf",
+    "2001 - The genetics of type 2 diabetes.pdf",
+    "2007 - Localization of type 1 diabetes susceptibility to the MHC Class 1 Genes.pdf",
+    "2019 - (Epi)genomic heterogeneity of pancreatic islet function and failure in type 2 diabetes.pdf",
+    "2010 - A recombination hotspot leads to sequence variability.pdf"
+  ],
+  "extraction_id": [
+    "46f1cae6-a01f-5445-b20f-0eadf892f8bf",
+    "43eecb5d-aca2-5c3e-9351-afbef000a795",
+    "682b7a19-c6f3-5773-8286-c027adef1fd3",
+    "69694cc4-e333-599c-9046-17a192ef3080",
+    "f53ccf4e-f47f-5b44-8b41-f7068efc8be3",
+    "e1274c5c-c854-52b0-83d9-72487111ba34",
+    "737e4fe2-91ba-50c5-8f64-1149944fb60c",
+    "92a54171-9f94-51ea-83cb-11698b1f0c21",
+    "7a2a9981-4096-5049-a717-3e69eb609777",
+    "7cf0ebfd-7231-540b-b44f-9c94316fdf80"
+  ],
+  "document_id": [
+    "e2dcbb80-5ad7-5441-b170-9b46607445b0",
+    "af63c74d-a204-5f9f-9a32-3451b112e5ba",
+    "3c9823cd-3615-53b6-96c8-b7d2123d3eb0",
+    "3c9823cd-3615-53b6-96c8-b7d2123d3eb0",
+    "80949bab-d085-5f61-b98a-4bee043bc4e2",
+    "4933cdc2-7d36-5181-87c9-63b58498839f",
+    "8ab06972-1c6c-5d68-a270-65fb0af0917b",
+    "3887995f-fa61-5472-b0a2-90b7b39592c2",
+    "b9bc63a5-e366-5685-bd7a-4732a8eeffb7",
+    "72115ac1-f66b-58c5-9a6f-2230ec7eacb4"
+  ],
+  "id": [
+    "chatcmpl-ADZ9vHPpqfR4t9mEA9x34UA73YZmX",
+    "abf69b53-da1e-5d4a-b957-e528cf986a22",
+    "2308bc87-b4e4-5e68-80f0-877bfd340377",
+    "bad4e085-d889-5a45-a5a4-f943a33bf72a",
+    "027471f3-0ccd-5b0d-b5d6-d8027ee07326",
+    "263dc0cb-dfa0-5ee2-b927-f9a196294d46",
+    "a76c839e-ec94-5fdb-b5b9-a3bd6eff1315",
+    "d8447ac5-d246-5cca-9336-693710b17f7a",
+    "4658d1c8-e096-54d3-8e93-4bf95a6ca114",
+    "84259ad2-080b-5f5c-82f3-0fe9a88500f4",
+    "ac8cfb6b-42cd-5c42-a4a3-b525790a22b1"
+  ],
+  "contexts": [
+    "are involved in the development of the disease [127 ].  There is evidence that more  than twenty regions of the genome are involved in t he genetic susceptibility to T1D.   The genes most strongly associated with T1D are loc ated in the HLA region of  chromosome 6 [128].  Similar to T1D, T2D has a stro ng genetic component.  To date,  more than 50 candidate genes for T2D have been inve stigated in various populations  worldwide.  Candidate genes are selected due to the ir interference with pancreatic",
+    "pre-existing statistical support for a role in T1D-susceptibility: these are the major histocompatibility complex (MHC), the genes encod- ing insulin, CTLA-4 (cytotoxic T-lymphocyte associated 4) and PTPN22 (protein tyrosine phosphatase, non-receptor type 22), and the regions around the interleukin 2 receptor alpha ( IL2RA/CD25 ) and interferon-induced helicase 1 genes ( IFIH1 /MDA5)94. However, these signals can explain only part of the familial aggregation of T1D.",
+    "C. The Insulin Gene A lesser genetic predisposition to T1D is conferred by the IDDM2 locus on chromosome 11 containing the insu-lin gene region. A polymorphic region located 5 =of the insulin gene was rst reported in 1984 to be associatedwith T1D in caucasoids (39). Now established as a pri- TYPE 1 DIABETES: FROM CAUSE TO CURE 81 Physiol Rev VOL 91 JANUARY 2011 www.prv.org Downloaded from journals.physiology.org/journal/physrev (041.090.188.152) on July 14, 2023.",
+    "ception of the insulin gene (434). The genetic susceptibil-ity component of T1D allows some targeting of primarypreventive care to family members of diagnosed T1Dpatients, but there is no complete inheritance of the dis-ease. Nevertheless, the risk for developing T1D comparedwith people with no family history is /H110111015 times greater. Although /H1101170% of individuals with T1D carry",
+    "Genes signifying increased risk for both type 1 and type 2 dia-betes have been identified. Genomewide association studies have identified over 50 loci associated with an increased genetic risk of type 1 diabetes. Several T1D candidate genes for increased risk of developing type 1 diabetes have been sug-gested or identified within these regions, but the molecular basis by which they contribute to islet cell inflammation and beta cell destruction is not fully understood. 12 Also, several",
+    "14   carried out on large cohorts including collections of families with  affected sibling pairs (Pociot   et al., 2010). These studies have provided evidence for over forty T1D susceptibility regions ,  but the  exact mechanisms by which the variation found in these regions  confer susceptibility to  T1D is still not clear (Noble and Erlich, 2012). The most important genes contributing to T1D  susceptibility are located in the MHC class II region , also  referred to as t he Human Leukocyte",
+    "The ultimate proof of an inherited contribution to disease pathogenesis comes from the identication of susceptibility genes. As described below, an increasing number of T2D susceptibility genes have been discovered in the past decade, especially,but not exclusively, in monogenic subtypes. Collectively, these probably account for294 A. L. Gloyn and M. I. McCarthy",
+    "loci contribute to Type 1 Diabetes (T1D) susceptibility and age at T1D onset. Hum. Immunol. 66,301313 (2005). 9. Aly, T. A. et al. Extreme genetic risk for type 1A diabetes. Proc. Natl Acad. Sci. USA 103, 14074 14079 (2006). 10. Noble, J. A. et al. The HLA class I A locus affects susceptibility to type 1 diabetes. Hum. Immunol. 63,657664 (2002). 11. Honeyman, M. C., Harrison, L. C., Drummond, B., Colman, P. G. & Tait, B. D. Analysis of families at risk for insulin-dependent diabetes mellitus reveals that",
+    "failure linked to T2D genetic risk and pathophysiology. Single celltranscriptome analysis of human islet cells indicate that multiplemonogenic diabetes genes are highly expressed in beta cells (e.g., PDX1, PAX4, INS, HNF1A, andGCK)[27]. However, other non-beta cell types express genes mutated in monogenic diabetes (such as PAX6 and RFX6 ), congenital hyperinsulinemia ( HADH, UCP2 ) and those implicated as T2D GWAS target/effector genes [28].",
+    "chain promoter (Serreze and Leiter 2001). This observation, alongwith human genetic studies, suggests that increased T1D risk in humans may also result from the combination of rare and common variants within the human population (Concannon et al. 2009b). Despite the identification of several Iddgenes to date, this limited collection does not fully explain T1D pathogenesis or the underlying genetic architecture for T1D risk. One of the many Idd"
+  ],
+  "answer": "The genes that give a predisposition to developing Type 1 Diabetes (T1D) include:  1. **Major Histocompatibility Complex (MHC)** genes located in the HLA region of chromosome 6 [1], [2], [6]. 2. **Insulin gene (INS)** located in the IDDM2 locus on chromosome 11 [3]. 3. **CTLA-4 (Cytotoxic T-Lymphocyte Associated 4)** [2]. 4. **PTPN22 (Protein Tyrosine Phosphatase, Non-Receptor Type 22)** [2]. 5. **IL2RA/CD25 (Interleukin 2 Receptor Alpha)** [2]. 6. **IFIH1/MDA5 (Interferon-Induced Helicase 1)** [2].  These genes are strongly associated with T1D and contribute to the genetic susceptibility to the disease.",
+  "question": "Which genes give a predisposition to developing T1D?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_7 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_7
new file mode 100644
index 0000000..17c628d
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_7
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf",
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf",
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "2021 - Old data and friends improve with age Advancements with the updated tools.pdf"
+  ],
+  "extraction_id": [
+    "4ef84d22-b428-5386-bbc0-39dbd364d3d7",
+    "7524bdfa-63f8-57c1-b5fe-1edcf11c275e",
+    "c8e9c4b7-19c6-5426-83a2-6f8628b68ceb",
+    "9c89683f-aca5-57f9-b28d-62e9eb64377b",
+    "a3ae6875-b0fc-5a4e-866f-4fee99c7d2a2",
+    "bb247bfe-333b-553a-94e6-2dc1b13b4723",
+    "bb247bfe-333b-553a-94e6-2dc1b13b4723",
+    "a3ae6875-b0fc-5a4e-866f-4fee99c7d2a2",
+    "9c89683f-aca5-57f9-b28d-62e9eb64377b",
+    "92339404-3864-5d8d-8731-8f8d0e9ac24c"
+  ],
+  "document_id": [
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0",
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0",
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "55cb2c81-b699-54df-96ab-2bf0b888031e"
+  ],
+  "id": [
+    "chatcmpl-ADZA23VOOb8blXNwMegY44QCzuw8S",
+    "c36215f6-2230-58ef-b3eb-44d1799ba5c2",
+    "89a578c7-5961-5b88-9a6d-f338216702c3",
+    "81e589eb-aa51-5f2a-966f-31928fb31943",
+    "20738786-99f6-573d-963e-377782eeb7a6",
+    "ae4d4109-66f7-59be-92f5-dc10c9dc2dd6",
+    "7e15e9b2-c731-5ab0-85c0-b6b432623220",
+    "f0c00edb-f07d-5975-a16b-16a072d0f2d4",
+    "1bf9bb72-ebaa-51d1-82ce-aae2f16dd92b",
+    "e2e526cb-0ac3-51ff-a1c5-43ff032b5558",
+    "28386b6b-e00f-5c0f-91d2-6fd031e9433a"
+  ],
+  "contexts": [
+    "supported by a signicant BLAST match to one or more expressed sequences or proteins. Ensembl also identies the positions of known human genes from public sequence database entries, usually using GENEWISE to predict their exon structures. The total set of Ensembl genes should therefore be a much more accurate reection of reality than ab initio predictions alone, but it is clear that some novel genes are missed (Hogenesch et al. , 2001). Of the many novel genes that are detected, some are",
+    "supported by a signicant BLAST match to one or more expressed sequences or proteins. Ensembl also identies the positions of known human genes from public sequence database entries, usually using GENEWISE to predict their exon structures. The total set of Ensembl genes should therefore be a much more accurate reection of reality than ab initio predictions alone, but it is clear that some novel genes are missed (Hogenesch et al. , 2001). Of the many novel genes that are detected, some are",
+    "supported by a signicant BLAST match to one or more expressed sequences or proteins. Ensembl also identies the positions of known human genes from public sequence database entries, usually using GENEWISE to predict their exon structures. The total set of Ensembl genes should therefore be a much more accurate reection of reality than ab initio predictions alone, but it is clear that some novel genes are missed (Hogenesch et al. , 2001). Of the many novel genes that are detected, some are",
+    "populations as Ensembl reects the progress of the International Haplotype Map Project (Thorisson et al. , 2005). More speculative data, such as GENSCAN-predicted exons that have not been incorporated into Ensembl-conrmed genes, may also be viewed. This means that the display can be used as a workbench for the user to develop personalized an- notation. For example, one may discover novel exons by nding GENSCAN exon predictions which coincide with good matches to a fragment of the draft mouse",
+    "populations as Ensembl reects the progress of the International Haplotype Map Project (Thorisson et al. , 2005). More speculative data, such as GENSCAN-predicted exons that have not been incorporated into Ensembl-conrmed genes, may also be viewed. This means that the display can be used as a workbench for the user to develop personalized an- notation. For example, one may discover novel exons by nding GENSCAN exon predictions which coincide with good matches to a fragment of the draft mouse",
+    "populations as Ensembl reects the progress of the International Haplotype Map Project (Thorisson et al. , 2005). More speculative data, such as GENSCAN-predicted exons that have not been incorporated into Ensembl-conrmed genes, may also be viewed. This means that the display can be used as a workbench for the user to develop personalized an- notation. For example, one may discover novel exons by nding GENSCAN exon predictions which coincide with good matches to a fragment of the draft mouse",
+    "Ostell/Spidey/ SSAHA at Sanger Institute http://www.sanger.ac.uk/Software/analysis/SSAHA/ human and mouse genomes, where there are large full-length cDNA collections to guide the hunt for genes, Ensembl should be very reliable. From the beginning, many genomic features other than predicted genes were included in Ensembl: different repeat classes, cytological bands, CpG island predic- tions, tRNA gene predictions, expressed sequence clusters from the UniGene database",
+    "Ostell/Spidey/ SSAHA at Sanger Institute http://www.sanger.ac.uk/Software/analysis/SSAHA/ human and mouse genomes, where there are large full-length cDNA collections to guide the hunt for genes, Ensembl should be very reliable. From the beginning, many genomic features other than predicted genes were included in Ensembl: different repeat classes, cytological bands, CpG island predic- tions, tRNA gene predictions, expressed sequence clusters from the UniGene database",
+    "Ostell/Spidey/ SSAHA at Sanger Institute http://www.sanger.ac.uk/Software/analysis/SSAHA/ human and mouse genomes, where there are large full-length cDNA collections to guide the hunt for genes, Ensembl should be very reliable. From the beginning, many genomic features other than predicted genes were included in Ensembl: different repeat classes, cytological bands, CpG island predic- tions, tRNA gene predictions, expressed sequence clusters from the UniGene database",
+    "comprehensive, powerful, flexible and interactive gene set enrichment analysis toolkit. Nucleic Acids Research ,45(W1), W130W137. [44] Zhang, B., Kirov, S., & Snoddy, J. (2005). WebGestalt: an integrated system for exploring gene sets in various biological contexts. Nucleic Acids Research ,33(Web Server issue), W741-8. [45] McLaren, W., Gil, L., Hunt, S. E., Riat, H. S., Ritchie, G. R. S., Thormann, A., Flicek, P ., et al. (2016). The ensembl variant effect predictor. Genome Biology ,17(1), 122."
+  ],
+  "answer": "Ensembl is a comprehensive genomic database that identifies the positions of known human genes from public sequence database entries, often using tools like GENEWISE to predict their exon structures. It aims to provide a more accurate reflection of reality compared to ab initio predictions alone [1]. Additionally, Ensembl includes various genomic features such as different repeat classes, cytological bands, CpG island predictions, tRNA gene predictions, and expressed sequence clusters from the UniGene database [7]. It also reflects the progress of projects like the International Haplotype Map Project and allows users to view speculative data and develop personalized annotations [4].",
+  "question": "What is ensembl"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_8 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_8
new file mode 100644
index 0000000..817666f
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_8
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2009 - Use of rat genomics for investigating the metabolic syndrome.pdf",
+    "2007 - The 20th International Mammalian Genome Conference Meeting Report.pdf",
+    "2009 - Use of rat genomics for investigating the metabolic syndrome.pdf",
+    "2009 - Use of rat genomics for investigating the metabolic syndrome.pdf",
+    "2007 - The 20th International Mammalian Genome Conference Meeting Report.pdf",
+    "2018 - Reproducibility and replicability of rodent phenotyping in preclinical studies.pdf",
+    "2014 - An evolutionarily conserved role for the aryl hydrocarbon receptor in the regulation of movement.pdf",
+    "2021 - Characterizing modifier genes of cardiac fibrosis phenotype in hypertrophic cardiomyopathy.pdf",
+    "2009 - Prioritizing genes for follow-up from genome wide association studies using information on gene expression in tissues relevant for type 2 diabetes mellitus.pdf",
+    "1999 - Functional Genomics and Rat Models.pdf"
+  ],
+  "extraction_id": [
+    "29832535-60a1-5d5f-9909-6b38160bb183",
+    "b846ba66-5f3b-5ff4-bf49-4324909d52c5",
+    "9de9e2d1-114a-5fa2-ae3f-e646f59ee116",
+    "6027b20e-d480-5485-874b-62cbe06c9c57",
+    "b846ba66-5f3b-5ff4-bf49-4324909d52c5",
+    "6af0332a-a004-5933-91e1-fb3fcd42fc2d",
+    "10934f40-1148-5e89-a06d-01909c6807e7",
+    "31573012-679a-513b-a878-882723f39855",
+    "9d081a37-83c4-52f5-9ed1-43a05a44a62c",
+    "2a252f5b-a6a1-54bd-bc0a-c25642002243"
+  ],
+  "document_id": [
+    "b06c0e90-1be1-5ba1-ad60-02b238070d07",
+    "d8b5b643-b7e7-5534-81fa-ee2e3679102d",
+    "b06c0e90-1be1-5ba1-ad60-02b238070d07",
+    "b06c0e90-1be1-5ba1-ad60-02b238070d07",
+    "d8b5b643-b7e7-5534-81fa-ee2e3679102d",
+    "2c03b37f-8c92-5fee-b19d-c582df5edb13",
+    "6a49b34d-b451-5b28-9e66-34c37b3ace6e",
+    "b29bc6c1-384d-5d91-bc0e-d6907116871c",
+    "4b1a56e7-6821-5504-b6da-27dcdf57c6a5",
+    "dd8b0499-f6d2-5202-8093-1a36d99796de"
+  ],
+  "id": [
+    "chatcmpl-ADZA6mykgNrrlE5Rh6Pwwt7u5tbjM",
+    "9a5513d0-5aeb-5c7e-9343-1794cee269d1",
+    "e47b58b3-214c-55a7-8a82-ea5d3b3e91db",
+    "ddc43bd2-6e83-5e79-9f3e-682a77398eeb",
+    "b35435ab-72c5-50c2-ab3d-df1f6c9fc445",
+    "74508b6c-cbb0-56ea-8acb-47a1c271e820",
+    "54b6e5a7-49e5-5e2a-9c35-86c10f671cd8",
+    "c9ca3828-4dcd-554c-97ef-5af644093f54",
+    "27781fa3-a3bd-5d17-9e77-b039ec04126b",
+    "5c92f513-fea8-51fa-8432-929553dc9e32",
+    "976a6422-6743-5d92-b368-3712cd13d3d2"
+  ],
+  "contexts": [
+    "417 Use of Rat Genomics for Investigating the Metabolic Syndrome and phenotypic traits are available to the scientific community  in databases, such as Ensembl ( http://www.ensembl.or g), the  Rat Genome Database ( http://www.rgd.mcw.ed u), eQTL  Explorer ( http://www. web.bioinformatics.ic.ac.uk/eqtlexplore r)  or GeneNetwork ( http://www.genenetwork.or g). Additional  online rat genetic resources have been recently reviewed by  Twigger et  al. (11).",
+    "Howard Jacob (Medical College of Wisconsin) discussed the Rat Genome Database disease portals, a platform for genetic and genomic research. Thereare 845 strains of rats, 573 of which are inbred,including substrains. Historically, biologists usingthe rat as a model have been disease focused,studying diseases, related phenotypes, pathways, and biological processes. The Rat Genome Database",
+    "10. Consortium STAR, Saar K, Beck A, Bihoreau  MT, Birney E, Brocklebank D, Chen Y et  al  (2008) SNP and haplotype mapping for  genetic analysis in the rat. Nat Genet  40:560566  11. Twigger SN, Pruitt KD, Fernndez-Surez  XM, Karolchik D, Worley KC, Maglott DR  et al (2008) What everybody should know about the rat genome and its online resources.  Nat Genet 40:523527  12. Butcher LM, Beck S (2008) Future impact of  integrated high-throughput methylome anal- yses on human health and disease. J Genet",
+    "for linkage analyses using new methods of efficient genotyping  based on genechip microarrays (10). In addition, over 800,000  ESTs and 5,000 annotated rat gene sequences are available for  functional analyses of candidate genes. Development of new  methodologies for high throughput phenotyping, such as expres- sion profiling, are becoming routinely used. Most of these genetic 2. Recent  Advances in Rat  Genetics and  Genomics",
+    "serves as a repository of all rat QTLs related to thedisease area as well as associated mouse and humanQTLs, strains used as disease models, phenotypedata, related references, expression data, genome-wide views of disease genes, and QLS via GViewer,comparative maps of disease-related regions, cus-tomization of data sets and download options, and analysis and visualization of function and cellular localization makeup of gene sets (http://www.rgd.mcw.edu/). ENU mutagenesis is now being done with rats.",
+    "3. Can data sharing in rodent phenotyping help with replicability? Laboratory mice and rats are the main mammalian models currently used for high-throughput genomic and behavior genetic research, and are employed primarily to explore and test gene function. This is con- sidered by some to be the great challenge facing biologists today  (Collins et al., 2007 ). Rodent models are used extensively as part of preclinical development and testing of treatments for disease in hu-",
+    "Bioinformatics and Statistical Analysis R was used for basic analysis of phenotypic data. GeneNetwork (www.genenetwork.org) was used for correlation and genetic analyses. The original phenotypes published in this paper and all microarray data generated in these cohorts are available for public analysis or download using the GeneNetwork database (Species: Mouse, Group: BXD, Type: Adipose mRNA, Liver mRNA, or Muscle mRNA, then select the EPFL datasets). The three",
+    "[23]. Shimoyama M, De Pons J, Hayman GT, Laulederkind SJ, Liu W, Nigam R, Petri V , Smith JR,  Tutaj M, Wang S-J, The Rat Genome Database 2015: genomic, phenotypic and environmental  variations and disease, Nucleic acids research 43(D1) (2014) D743D750. [PubMed: 25355511]  [24]. Dickinson ME, Flenniken AM, Ji X, Teboul L, Wong MD, White JK, Meehan TF, Weninger WJ,  Westerberg H, Adissu H, High-throughput discovery of novel developmental phenotypes, Nature  537(7621) (2016) 508. [PubMed: 27626380]",
+    "database (dbSNP) build 130 to identify genes located inthe vicinity of selected SNPs. Homologues of the genes formouse and rat were identified using the NCBI's Homolo-Gene release 64. We included only those genes that wereevolutionarily conserved in three different species namelyhuman, mouse and rat. Analysis of microarray data",
+    "(data not shown). Therefore, it seems logical to position the rat field so themechanistic, disease-based research canbe integrated into the awesome power ofthe human and mouse genome projects. Progress of the Rat Genome Project Recognizing the usefulness of the rat as amodel system, NIH, led by the NationalHeart, Lung, and Blood Institute(NHLBI), has funded the Rat GenomeProject (RGP), the Rat Expressed Se-quence Tag (RGP EST) Project, and the Rat"
+  ],
+  "answer": "You can use the Rat Genome Database (RGD) for genetic, genomics, phenotype, and disease-related data generated from rat research. The RGD is specifically mentioned as a platform for genetic and genomic research, including disease portals and phenotypic data [2], [5]. Additionally, it serves as a repository for various types of data such as QTLs, strains used as disease models, expression data, and more [5]. The database can be accessed at http://www.rgd.mcw.edu [1], [5].",
+  "question": "Which database can I use for genetic, genomics, phenotype, and disease-related data generated from rat research?"
+}
+\ No newline at end of file
diff --git a/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_9 b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_9
new file mode 100644
index 0000000..cab698c
--- /dev/null
+++ b/gnqa/paper2_eval/data/dataset/human/intermediate_files/human_de_gn_9
@@ -0,0 +1,65 @@
+{
+  "titles": [
+    "2010 - Integrated genomic approaches to identification of candidate genes underlying metabolic and cardiovascular phenotypes in the spontaneously hypertensive rat.pdf",
+    "2015 - Multipronged approach to identify and validate a novel upstream regulator of Sncg.pdf",
+    "The FEBS Journal - 2015 - Chintalapudi - Multipronged approach to identify and validate a novel upstream regulator of Sncg.pdf",
+    "2007 - Bioinformatics_for_Geneticists.pdf",
+    "003 -Barnes- Bioinformatics_for_Geneticists.pdf",
+    "2007 - Bioinformatics_for_Genetices_MAZEN_SAEED.pdf",
+    "2018 - Genetic Networks Activated by Blast Injury to the Eye.pdf",
+    "2010 - Identification of a Chr 11 quantitative trait locus that modulates proliferation in the rostral migratory stream of the adult mouse brain_.pdf",
+    "2011 - Genetic Regulatory Network Analysis for Rpe65 in the Eye of BXD Mice.pdf",
+    "2013 - Effects of Glaucoma on Chrna6 Expression in the Retina.pdf"
+  ],
+  "extraction_id": [
+    "bb30622c-7f00-5ee4-928d-6f4f6f9f9e3d",
+    "ad4bf6de-f811-5ebc-82be-5fbd3aa1ba03",
+    "184d5422-8e35-57ca-b542-3bcfbd821b5a",
+    "af7722e9-a91e-533e-9403-e54ff59ffd73",
+    "9f3fd618-f56f-538a-b955-c7205a7c8107",
+    "d528a008-6931-562b-831c-f3c6dd925fac",
+    "c02d0625-3478-52d4-8339-78b2df351668",
+    "5eb43710-e1c0-5955-a34c-fb4b7204f421",
+    "86c6e14a-66bf-5a33-bcb6-750fbf259c87",
+    "1b9b34cc-e87b-53c2-aab5-2913d1e6fd25"
+  ],
+  "document_id": [
+    "ec54d632-be36-5d11-8437-2233e07049a0",
+    "803e3b96-d4ed-5f1f-b788-eb7564d4f6b4",
+    "734e6a57-5d63-5e10-b01d-1ccc04618c8a",
+    "4ea8e1a8-e113-5f02-ad78-880b9c51a101",
+    "045edae8-468b-5725-be06-8cb4b8f6a92b",
+    "139463d1-c63c-5c51-bf9c-9ccc356768e0",
+    "57e3820f-7a5d-51f1-a0c6-ecfbdf546005",
+    "0b7c325a-0be0-54a2-9c8f-d4607d0f7151",
+    "44e70f2d-3bda-563c-ae0b-83833b98529b",
+    "3d3e03db-a961-5668-bd69-44039142fb87"
+  ],
+  "id": [
+    "chatcmpl-ADZAAr5I5AVyGWpvGliJuDqL5V4HD",
+    "0b23b976-d97e-56ea-bb4d-6372a12cd48e",
+    "d71e9649-f56e-5376-9b97-79d450e932de",
+    "96700d1c-5c9b-545d-bec5-338a5aa8ea19",
+    "cbf58283-6ec3-5fc4-8a1e-73b1f0aa27f3",
+    "4a7fc44c-82d5-5808-a864-2dd4dd1ce33f",
+    "5008de52-b46c-5eb4-b033-66bdccda49a1",
+    "0db6fb13-b666-586a-bfe6-63b31e44ec5d",
+    "6d6bf436-2af4-5f8b-bf8b-81de331d2ad7",
+    "602fed11-6848-5916-89cd-67189890f37c",
+    "2c979a43-4536-5171-9f11-2c620a117551"
+  ],
+  "contexts": [
+    "were identied using the RGD (68). This resource provides infor-mation regarding the physiological trait studied, strain combina-tion used, associated linkage statistics, and the genomic coordi-nates of the pQTL region. For pQTL regions identied from RGD,the original data (Supplementary Table S3) were examined, and the99% condence interval [within the 2 logarithm of the odds (LOD)drop from the peak of linkage] was estimated. Cis-eQTLs were",
+    "RGCs. The discovery of this relationship may help inguiding studies that explore the disease mechanismsassociated with altered protein transport and foldingin RGCs. In glaucoma, the identication and conr-mation of these two proteins in RGC health and dis-ease holds great promise for the development ofmolecular targets to slow or reverse RGC damage, which, in turn, will preserve vision. Experimental procedures Human donor eyes Human donor eyes were collected in accordance with the",
+    "RGCs. The discovery of this relationship may help inguiding studies that explore the disease mechanismsassociated with altered protein transport and foldingin RGCs. In glaucoma, the identication and conr-mation of these two proteins in RGC health and dis-ease holds great promise for the development ofmolecular targets to slow or reverse RGC damage, which, in turn, will preserve vision. Experimental procedures Human donor eyes Human donor eyes were collected in accordance with the",
+    "(http://www.cbil.upenn.edu/PaGE/). All microarray platforms and image-analysis software are supported. In addition, RAD is being used for CGH, ChIP , and SAGE data. RAD can produce MAGE-ML les for export of data to other databases or software packages. RAD is part of a more general Genomics Unied Schema, which provides a platform to integrate gene and transcript data from a variety of organisms. Advantages RAD is a scalable, Web-accessible database that can accommodate data from sev-",
+    "(http://www.cbil.upenn.edu/PaGE/). All microarray platforms and image-analysis software are supported. In addition, RAD is being used for CGH, ChIP , and SAGE data. RAD can produce MAGE-ML les for export of data to other databases or software packages. RAD is part of a more general Genomics Unied Schema, which provides a platform to integrate gene and transcript data from a variety of organisms. Advantages RAD is a scalable, Web-accessible database that can accommodate data from sev-",
+    "(http://www.cbil.upenn.edu/PaGE/). All microarray platforms and image-analysis software are supported. In addition, RAD is being used for CGH, ChIP , and SAGE data. RAD can produce MAGE-ML les for export of data to other databases or software packages. RAD is part of a more general Genomics Unied Schema, which provides a platform to integrate gene and transcript data from a variety of organisms. Advantages RAD is a scalable, Web-accessible database that can accommodate data from sev-",
+    "differentiallysusceptibletodeath,withalpha-RGCsandintrinsicallyphotosensitiveRGCs (ipRGCs) being less sensitive to cell death than other RGC subtypes in a mouse model of glaucoma. Keywo rds: retinal ganglion cells, gene regulatory networks, transcription factors, recombinant inbred strain, subtypes INTRODUCTION Theretinalganglioncell(RGC)isthenaloutputneuronoftheretina,projectingthroughtheoptic nerve to the brain, where it targets a number of functionally distinct areas: for visual perception,",
+    "AG18245 (DG), NIAAA U01AA014425 (LL), and P20 DA021131 (RW). We thank Derek Rains, Gurjit Rai, Meifen Lu, Richard Cushing, Erich Brauer, and Alan Weatherford for their invaluable technical assistance. Abbreviations BrdU bromodeoxyuridine CV cresyl violet GF growth fraction LOD likelihood of the odds LRS likelihood ratio statistic NSCs neural stem cells OB olfactory bulb DG dentate gyrus QTL quantitative trait locus RI recombinant inbred RMS rostral migratory stream SGZ subgranular zone",
+    "Rdh10, Lrat,) whose biology functions are directly associated w ith the metabolism of  retinoid.  RGR (retinal G protein-coupled receptor, protein of  Rgr ) is a protein that structurally resembles  visual pigments and other G protein-coupled recepto rs. Light isomerizes 11- cis - into all-trans - retinal, triggering a conformational transition of the opsin molecule that initiates phototransduction .  After bleaching all- trans -retinal leaves the opsin, and light sensitivity mu st be restored by",
+    "GeneNetwork system, we were able to define robust expression covariance signatures for RGCs and confirmed membership of Chrna6 within the RGC cell type of the retina using new array data sets and RT-PCR tracking through a progressive RGC loss mouse line. Chrna6  can be added as reliable biomarker for RGCs and RGC loss secondary to glaucoma. It is important to note that in addition to providing evidence for Chrna6 expression as a"
+  ],
+  "answer": "RGD refers to a resource that provides information regarding physiological traits studied, strain combinations used, associated linkage statistics, and the genomic coordinates of the pQTL (protein Quantitative Trait Loci) region. For pQTL regions identified from RGD, the original data were examined, and the 99% confidence interval within the 2 logarithm of the odds (LOD) drop from the peak of linkage was estimated [1].",
+  "question": "What is RGD?"
+}
+\ No newline at end of file
author	ShelbySolomonDarnell	2024-10-17 12:24:26 +0300
committer	ShelbySolomonDarnell	2024-10-17 12:24:26 +0300
commit	00cba4b9a1e88891f1f96a1199320092c1962343 (patch)
tree	270fd06daa18b2fc5687ee72d912cad771354bb0 /gnqa/paper2_eval/data/dataset/human/intermediate_files
parent	e0b2b0e55049b89805f73f291df1e28fa05487fe (diff)
download	gn-ai-master.tar.gz