diff options
| -rwxr-xr-x | cnt_gene_addiction_abstracts.py | 33 | ||||
| -rwxr-xr-x | ratspub.py | 2 |
2 files changed, 34 insertions, 1 deletions
diff --git a/cnt_gene_addiction_abstracts.py b/cnt_gene_addiction_abstracts.py new file mode 100755 index 0000000..c216862 --- /dev/null +++ b/cnt_gene_addiction_abstracts.py @@ -0,0 +1,33 @@ +#!/bin/env python3 +import os +import re +import time +from ratspub import * +## turn dictionary (synonyms) to regular expression + +def gene_addiction_cnt(gene): + q="\"(" + addiction.replace("|", "[tiab] OR ") + ") AND (" + drug.replace("|", "[tiab] OR ", ) + ") AND (" + gene + ")\"" + count=os.popen('esearch -db pubmed -query ' + q + ' | xtract -pattern ENTREZ_DIRECT -element Count ').read() + return (count) + +out=open("gene_addiction_cnt_result_part1.tab", "w+") + +cnt=0 +with open ("./ncbi_gene_symb_syno_txid9606_part1.txt", "r") as f: + for line in f: + line=re.sub(r"\)|\(|\[|\]|\*|\'","",line) + (gene, synoms)=line.strip().split("\t") + if (gene[0:3] != "LOC"): + cnt+=1 + for syno in synoms.split("|"): + if len(syno)>3: + gene+="|"+syno + gene_q=gene.replace("|", " [tiab] OR ") + gene_q+="[tiab]" + count=gene_addiction_cnt(gene_q) + if (len(count)==0): + print("pause") + time.sleep(10) + print(gene+"\t"+count) + out.write(gene+"\t"+count) + @@ -105,7 +105,7 @@ addiction=undic(addiction_d) drug_d = {"alcohol":"alcohol|alcoholism|alcoholic", "nicotine":"smoking|nicotine|tobacco", "cocaine":"cocaine", - "opioid":"opioid|opioids|fentanyl|oxycodone|oxycontin|heroin|morphine|methadone|buprenorphine", + "opioid":"opioid|opioids|fentanyl|oxycodone|oxycontin|heroin|morphine|methadone|buprenorphine|vicodin|hydrocodone|hycodan|kadian|percoset|hydromorphone|naloxone|codeine|suboxone|tramadol|kratom", "amphetamine":"methamphetamine|amphetamine|METH", "cannabinoid":"endocannabinoid|cannabinoids|endocannabinoids|marijuana|cannabidiol|cannabinoid|tetrahydrocannabinol|thc|thc 9|Oleoylethanolamide|palmitoylethanolamide|acylethanolamides" } |