Lexical mapping analysis and feedback/review

[hrshdhgd]

• [ ] Create a table
  • [x] All matches: For this one, we want general feedback, false assumptions and suggestions (do not spend more than 30 mins)
  • [ ] Table with just ORDO matches that are not already in MONDO. Goal is to remove everything in tmp/mondo.sssom.tsv.
  • [x] Compare tmp/mondo.sssom.tsv and src/mappings/mondo-sources-all-lexical.sssom.tsv
    • [x] Implement in sssom-py a method on MSDF msdf1.remove_mappings(msdf2) -> msdf output: all rows for KEY_FEATURES that is [msdf1.df - msdf2.df] Implement CLI as well (ideally)
  • [ ] Review mappings and get feedback for bad mappings

[hrshdhgd]

Do we still need just ORDO mappings?

[matentzn]

No all. I think we need to count how many mappings by prefix their are to make sure that all of the ontologies are taking into account by the mapping process

[matentzn]

https://github.com/monarch-initiative/mondo/issues/5232

Our first OAK run revealed mostly wrong matches, need to make sure this does not happen again

[matentzn]

Issues with the current matching to deal with

• [ ] missed synonyms (bag of words matching could be better than relying on synonyms?): neuroleptic malignant syndrome -> Malignant neuroleptic syndrome not caught
• [ ] irrelevant categorical information or synonyms in brackets: Leprosy [Hansen’s disease], Pregnancy with abortive outcome (O00-O08)
• [ ] Broken encodings Waldenström macroglobulinemia - any smart idea how to handle with now access to the source? Here we need a smart way. What comes to mind is replacing broken chars with regex wildcards like Waldenstr.m macroglobulinemia.

[hrshdhgd]

https://github.com/monarch-initiative/mondo-ingest/issues/78#issuecomment-1946865977