Closed ningyifan closed 8 years ago
TODO: identify similar QA issues
I have updated the CSV file to include the information that was previously missing.
I attempted to do this manually, but it was taking too long, so I switched to a different approach and edited the query script that Yifan had made.
The issue was that sometimes, not all of the information was included, ie number of participants, which excluded the information from the CSV file, which is not how that should be. I edited the file so this is not the case.
However, this does not conclude the QA issue here. By using ChEBI identifiers, which we are deriving from DrugBank identifiers, we are missing a number of drugs, and, subsequently, assertions from the old DIKB. I have attempted to fix this problem by manually adding DB to ChEBI mappings for drugs that were common enough for me to recognize, but this is by no means complete.
We will have to identify a better mapping system or accept a loss of data in the transition from old to new DIKB.
ChEBI mappings updated. 8 metabolites that do not have ChEBI identifiers are still pending -- see JoCHEM http://biosemantics.org/index.php/resources/jochem
Here are the umapped metabolites....
object: demethylcitalopram object: R-demethylcitalopram object: S-demethylcitalopram object: clozapine-N-oxide object: demethylcitalopram object: desacetyldiltiazem object: N-demethyldesacetyl-diltiazem object: R-demethylcitalopram object: S-demethylcitalopram object: beta-hydroxy-lovastatin object: beta-hydroxy-simvastatin object: dehydro-aripiprazole object: demethylcitalopram object: N-desalkylquetiapine object: R-demethylcitalopram object: reduced-haloperidol object: S-demethylcitalopram object: hydroxybupropion
All drugs from old-DIKB are included. All metabolites are skipped.
Issues:
Claim-41:
Claim-105: