Missing 'common knowledge' Drug-Disease Connections

[thiemeaunc]

The top 10 most commonly prescribed drugs in the US (https://www.drugreport.com/50-commonly-prescribed-drugs-in-america/) ['lisinopril','atorvastatin','L-thyroxine zwitterion','metformin','amlodipine','metoprolol','omeprazole','simvastatin','losartan','albuterol'] were identified on DrugBank (https://go.drugbank.com/), and all 'Associated Condition' terms were pulled to generate a 2 * 108 matrix where the first column contains the names of the 10 drugs listed above, and the second column contains the names of all disease/phenotype terms associated with each drug on DrugBank.

Between the 10 drugs, there are a total of 3,746 total pathways returned for the query: MATCH (cs:chemical_substance)-[r0]-(x1) WHERE cs.name IN ['lisinopril','atorvastatin','L-thyroxine zwitterion','metformin','amlodipine','metoprolol','omeprazole','simvastatin','losartan','albuterol'] RETURN DISTINCT(cs.name),r0.relationship_label,LABELS(x1),(x1.name). Out of this set, 2,683 pathways were within the (chemical_substance)--(disease_or_phenotypic_feature) subset. Only 7 out of 2,683 of these connections were identified as one of the 108 potential 'validation connections'. 'atorvastatin', 'L-thyroxine zwitterion', 'metformin', 'simvastatin', 'losartan' were included within these 7, but the remaining 5 drugs are not. Due to the wide usage of these drugs, and the relatively high amount of information pertaining to their pharmacology, the misalignment between these Top 10 drugs and the indications for which they are prescribed within the database is indicative of a larger problem regarding chemical--disease/phenotype relationship information processing.

ROBOKOP_csb_dpf_connection_validation_query_results.xlsx