search

BigDataBiology / argNorm

ARG normalization by mapping to the ARO ontology.
MIT License
25 stars 4 forks source link

Drug Categorization Code Added #8

Closed Vedanth-Ramji closed 1 year ago

Vedanth-Ramji commented 1 year ago

The approach I used was to integrate the drug categorization functions in the normalizer itself. While adding the ARO columns, I also add 'OVERALL CATEGORY OF DRUG CLASS' and 'CONFERS RESISTANCE TO IMMEDIATE DRUG CLASS' columns and use my drug categorization functions to get the data to be displayed in those columns.

I have included the drug categorization in the 'run' function of the base normalizer class, as the drug categorization functions only need ARO numbers as parameters, hence after normalization, differences between database syntaxes wouldn't matter for categorization code.

luispedro commented 1 year ago

The tests are now passing, but this commit has a lot of unnecessary material (remember that whatever you add in a commit will be in git's history forever). All the extra files should be removed before we can merge it

Also, the comment about from_obo_library is very relevant in this context: the aro.obo file is not used at all. I'd prefer to change that code to actually load from the copy of aro.obo that argNorm provides (although you need to take care of installing it and finding it in the right path!) but including a copy of aro.obo is a worst-of-both-worlds solution.