diff options
| -rw-r--r-- | topics/database/mariadb-database-architecture.gmi | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/topics/database/mariadb-database-architecture.gmi b/topics/database/mariadb-database-architecture.gmi index 22427b6..c9d532b 100644 --- a/topics/database/mariadb-database-architecture.gmi +++ b/topics/database/mariadb-database-architecture.gmi @@ -15,8 +15,8 @@ These are the terms we use * Genotypes * Case or genometype: individual, strain, sample -* ProbeData: obsolete -* ProbeSetData: trait/case values +* ProbeData: Now almost obsolete. [Comment by RWW perhaps for a footnote: In GeneNetwork 1 we had built and maintained a table for individual "Probe level" data simply because the Affymetrix data sets were so large. For example, the BXD Family: "UMUTAffy Hippocampus Exon 9Feb09)RMA" array data consists of 1.236 million "probesets" each of which is a summary of many individual probe assays (ProbeData)—a total of 4.5 million probes (see https://www.thermofisher.com/order/catalog/product/900817). In GN1 we built a special interface to interrogate these 4.5 million indivdual probes--extremely useful to studing the fine-structure of mRNA expresswion. We thought it best to split these very large "pro-level" data sets from the much smaller and more widely use "ProbeSetData". The term "Probe" in this particular context (Affymetrix Exon arrays) refers to short nucleotide probes used by Affymetrix and other microarray vendors. Affymetrix "Exon"-type arrays consist of 25 nt hybridization probes that target relatively specific parts of RNAs--mainly exons but also many intronic sequences.] +* ProbeSetData: trait/sample values almost exclusively used for molecular data types (mRNA, protein, methylation assays, metabolomics, etc). [Comment by RWW perhaps for a footnote: The term "ProbeSetData" should ideally be changed to "High_Content_Data_Assays. In 2003 the only high content data assays we had were Affymetrix microarrays that measured mRNA level, and the vendor called their assays "ProbeSets". We used this now obsolete term. Most ProbeSetData in GN1 and GN2 as of 2024 are measurments of molecular traits that can be tagged to a single genome location—-the location of the gene from which the mRNA and its derivative protein are transcribed and translated, or in the case of epigenomic studies—the site at which the genome is methylated. When these three types of molecular traits are mapped, we typically add a mark all graphic output maps that highlight the location of the "parent" gene. For example, the sonic hedgehog gene in mice is located on chromosome 5 at about 28.457 Mb on the mm10 assembly (aka GRCm38). When we measure the expression of Shh mRNA, we place a purple triangle at the coordinate of the Shh gene. Two notes: 1. There are at least three ProbeSetData types do NOT have parent genes--metabolomic data, and metagenomic data, and new high-content brain connectome data. When we do NOT know the location of a parent gene, we should NOT place any mark along the X-axis. 2. Ideally GN databases would define the TYPE of high-content data, so that the code could fork to the correct GUI for that particular data type. Connectome data for the brain is an example of a data type that is very large (40,000 measurements per brain), that is truly high-content data, but that is NOT molecular. Time series data may also fall into this category.] * ProbeSetFreeze: points to datasets ## More on naming |