Process xrefs correctly in OBO files, e.g. alt_ids

[julie-sullivan]

I don't see that the OboConverter handles subset definitions in obo files. Did that slip >my eye or is this on purpose? I'd like to include subset information for each term that is loaded. Christian

[zfinintermine]

It would be nice to see the "alt_ids" parsed and stored as well. This would allow us to link MEDIC terms with OMIM diseases terms. Sierra

[Term] id: MESH:C531617 name: Amyotrophic lateral sclerosis 1 alt_id: OMIM:105400 is_a: MESH:D000690 ! Amyotrophic Lateral Sclerosis synonym: "ALS1" EXACT [] synonym: "AMYOTROPHIC LATERAL SCLEROSIS 1, AUTOSOMAL DOMINANT AMYOTROPHIC LATERAL SCLEROSIS 1, AUTOSOMAL RECESSIVE, INCLUDED" EXACT [] synonym: "AMYOTROPHIC LATERAL SCLEROSIS 1, FAMILIAL" EXACT [] synonym: "Amyotrophic Lateral Sclerosis, Autosomal Dominant" EXACT [] synonym: "Amyotrophic Lateral Sclerosis, Familial" EXACT [] synonym: "AMYOTROPHIC LATERAL SCLEROSIS, SPORADIC, INCLUDED" EXACT [] synonym: "FALS" EXACT []

[julie-sullivan]

@zfinintermine Sierra, I looked at the code and it looks like alt_ids are being parsed. See here:

https://github.com/intermine/intermine/blob/dev/bio/core/main/src/org/intermine/bio/ontology/OboParser.java#L306

They are saved as synonyms. Can you verify that this is not happening? Or did you want the model to be different?

[julie-sullivan]

@cmpich @zfinintermine How did you want to load the GO slim data? How do you want to model it?

[julie-sullivan]

Duplicate of #178

[zfinintermine]

I want to be able to use the alt_id as a key. Can I do that with a synonym collection? Also, zebrafishmine has a possible model for OntologySubset that is working for us...

[cmpich]

Each term has a collection of subsets which we at ZFIN model through a many-to-many relationship between term and subset. Each ontology has a one-to-many relationship to subset, i.e. an ontology can have one or more subsets.

[julie-sullivan]

Sorry, I meant in terms of intermine. which collection would they go in, what would that be called, etc.

Each term has a collection of subsets which we at ZFIN model through a many-to-many relationship between term and subset. Each ontology has a one-to-many relationship to subset, i.e. an ontology can have one or more subsets.

---

Reply to this email directly or view it on GitHub: https://github.com/intermine/intermine/issues/412#issuecomment-22333878

[cmpich]

I think you need a new entity, called OntologySubset. ZFATerm e,g, then would have a collection of Ontologsubsets as well as OntologySubset would have a collection of ZFATerm entities. Is that what you mean?

[julie-sullivan]

This is a good idea because a lot of the ontologies are referencing each other.

[julie-sullivan]

Another use case: #1421

Where BDGP expression terms are using Fly anatomy terms, but not exclusively. Want to be able to reference one term from another.

Tell @rachellyne

[julie-sullivan]

@rachellyne Requires a model change so putting in queue for 2.0. Also requires some discussion.

[julie-sullivan]

@rachellyne We already have a OntologyTerm.crossReferences collection, which is currently only populated by Uberon terms. Just add these there?