Outage tasks for 2022-12-15

[kltm]

• Send reminder email to go-consortium mailing list
• Run noctua-models-migration pipeline for @vanaukenk (day before? morning of?) to get report
  http://skyhook.berkeleybop.org/noctua-models-migrations/reports/
• Refresh minerva code with latest from minerva master
• Run replaced_by term update on blazegraph (@vanaukenk)
• Cycle minerva to get latest ontology
• Update NEO
  ontology build outages
• Prep ticket for next outage

[kltm]

@vanaukenk The report is completed

[vanaukenk]

I've checked over the diff and most of the updates are okay and consistent with the work in these two ontology tickets:
geneontology/go-ontology#24152
geneontology/go-ontology#24486

However, we also have two automatic model updates in which a BP term is being replaced by an MF term, as a result of the work in this ticket:
geneontology/go-ontology#24546

This update will result in either logically incorrect models or models with possibly incorrect biological information.
The two models affected are:
http://noctua.geneontology.org/editor/graph/gomodel:R-HSA-5221030
http://noctua.geneontology.org/editor/graph/gomodel:63894f2500001622

For the Reactome model, presumably this will be corrected with the next Reactome import, but for the second, mouse model, I think an MGI curator will need to review the model after the update to fix it.

I'm not sure how we should handle these types of replacements, i.e. MF for BP, in the future as they will likely result in similarly incorrect models.

@kltm @ukemi @cmungall @pgaudet @deustp01 @balhoff

[kltm]

@vanaukenk I think you might be thinking more of the day-to-day but, not to sound like a broken record, this would be in the scope of geneontology/project-management#47. Theoretically, I assume that we'd go ahead with the breakages, but have fixing them be a high priority; larger changes would be shelved until there was enough resources to accomplish them. We would also be able to continuously run these checks against new versions of the ontology, so there would be more lead time and a way of testing changes.

[vanaukenk]

I think you might be thinking more of the day-to-day

@kltm I'm actually thinking more about the consequence of replacing a term from one ontology with another, i.e. BP with MF, and whether we should do that in the ontology or whether these types of changes should instead make use of the 'consider' tag.

[kltm]

@vanaukenk Is a main consequence under consideration that we'd be building up a body of "broken" (i.e. shex-illegal) models and what to do with them? Or is this purely a best practices for the ontology and how migrations operate? (If the latter, I'll hide this off-topic subthread.)

[vanaukenk]

@kltm and I talked about this.
We will go ahead with the model updates and alert the MGI curator to manually review the model post-update.
We'd also like to talk this finding over with the ontology editors to see what they think is the best thing to do going forward.

[pgaudet]

I shouldnt have replaced a BP by a MF, I'll change that

[vanaukenk]

Thanks @pgaudet

[ukemi]

I went in and removed the mouse annotations.

[vanaukenk]

See also: geneontology/go-ontology#24546 (comment)

[deustp01]

For Reactome see comment in #26546

[vanaukenk]

@kltm - I'm closing this outage ticket in favor of any new work going into the next outage ticket #811