search

monarch-initiative / mondo-ingest

Coordinating the mondo-ingest with external sources
https://monarch-initiative.github.io/mondo-ingest/
6 stars 3 forks source link

Update ingest (and maybe other things) related to DO switching to using "MIM" prefixes #441

Open sabrinatoro opened 4 months ago

sabrinatoro commented 4 months ago

DO switched to using "MIM" prefixes instead of "OMIM" prefixes. We need to review how this change will affect Mondo ingest (and Mondo in general).

One aspect that will be affected is the exact lex match on the OMIM cross-ref for DO alignment: For the alignment with DO, we allow exact matching between OMIM cross-ref, as reported in the exact lex match file: src/ontology/lexmatch/unmapped_doid_lex_exact.tsv This exact matching will not work since the prefixes are different.

Note: I don't know how this change in prefix will affect the rest of Mondo.

Related

joeflack4 commented 4 months ago

I saw that as well (slack). My initial thought is that we should update our EPM (extended prefix map) and add MIM as a synonym on OMIM. But I don't know if that alone is sufficient for everything to continue to work.

twhetzel commented 4 months ago

Bioregistry does contain mim as an alternative for OMIM: https://bioregistry.io/registry/omim and this is in the EPM.

@hrshdhgd mentioned that the semsql db would need to be updated with the prefixes.csv.

Sabrina mentions need to see if things break or if some alignments for exact matching are missing due to the change to MIM.

matentzn commented 4 months ago

I would suggest to put this on ice until someone presents a case for this at a tech call.. There are a number of complicated implications, including the fact that this is merging the OMIM and OMIMPS namespaces which will have major consequences on our user base.

joeflack4 commented 4 months ago

Mmm, I didn't consider that. The set of their MIM is our OMIM + OMIMPS. I suppose the solution we may need, when we implement it, is simply adding triples xref or exactMatch triples, e.g. MIM:<NUM> skos:exactMatch [OMIM | OMIMPS]:<NUM>

matentzn commented 4 months ago

is simply adding triples xref or exactMatch triples, e.g. MIM: skos:exactMatch [OMIM | OMIMPS]:

I would really not be a great fan of abusing SSSOM like that (using it to essentially map between the same identifiers and different prefix synonyms). But yeah, it may be an option.. In any case, I dont really see this as a high priority until some user pushes us to make the move.

joeflack4 commented 4 months ago

I wasn't even thinking of it in the SSSOM context, and they're close to synonyms, but not technically synonyms since the set of terms in the namespaces is not the same. Anyway, will defer to the future!