Register GOCHE prefix, or change terms to GO?

[balhoff]

The GOCHE: terms in GO use an OBO namespace that has not been registered. I think this sets a bad example. I think we should try to register this namespace with OBO (describing it as a special circumstance originating in GO's long history before some of the OBO policies were settled). Or else change all these terms to GO terms (within some particular ID range, since they are automatically generated), or use some other (non-OBO) namespace entirely.

Same situation pertains to GOREL:, but we are actively working on obsoleting those (#19497).

@cmungall thoughts?

[pgaudet]

Discussion on the GO editors call is that we may not need GOCHE after the work with roles that @krchristie is doing, but @cmungall would like to punt on this one until we have finished the current work on roles.

[matentzn]

I would like to know - especially now that we change our pipelines to use GO plus instead of GO, the question is what I should do with GOREL and GOCHE IRIs - I could simply delete them in the preprocessing, but I want to make sure this is what is intended.

[balhoff]

Note - I need to add GOCHE terms to go-base.

[pgaudet]

It would be nice to document the content of the ontology here: http://geneontology.org/docs/download-ontology/

[cthoyt]

Currently looking at GOCHE in for the Bioregistry since it appears in OBA

update: see http://bioregistry.io/registry/goche

[matentzn]

Any updates on GOCHE? Still does not appear in base, which makes it hard to handle classes like GO:0042908 which do not have a subclass axiom, but only an equivalent class one including a GOCHE term.

[balhoff]

Thanks for pointing out that issue. I will try to make some changes here soon.