Decide on standard annotation curation policies for protein-containing complexes and make certain that standard annotation form is compliant

[vanaukenk]

We need to be clear on what annotations are allowed for each GO aspect for protein-containing complexes as subjects of standard annotation and make certain that the new standard annotation form is compliant.

Right now, curators can select a protein-containing complex from the subject autocomplete and add a GO term, ECO code, and reference, but no gp2term relations can be selected or are populated by default and the curator can still press the SAVE button although no annotation gets made as far as I can see.

@tmushayahama @thomaspd

[thomaspd]

@vanaukenk AFAIK it's only SGD who are annotating GO terms to complexes, and they are using SGD identifiers for those. Is that right? I see that GO complexes can currently be selected in the gene product field of the Standard Annotation form, but I don't think we would want curators to use those.

@tmushayahama just to clarify, this would be using an identifier for a complex in the "gene product" field of the form. An example is SGD:S000345854. When I select it in the gene product field, and then select a GO term, I also don't see a default relation appear. Is this because a complex identifier isn't expected to be selected in the gene product field?

[vanaukenk]

@thomaspd - yes, as far as I know, it's only SGD that currently has standard annotations to protein-containing complexes in Noctua, and the identifiers they use are of the format SGD:S000229701, etc.

I agree that we don't want GO protein-containing complexes as subjects of annotations in the standard annotation form. I suspect they are there because we are trying to restrict the autocomplete to allowed entity types for annotation and since both CPX entities and GO terms are typed as protein-containing complexes, they both appear in the autocomplete.

If we can somehow restrict IDs with prefix GO: from the annotation subject autocomplete, that should take care of the issue.

[vanaukenk]

From the 2024-05-03 workbenches call:

1) We now restrict the id space for subjects to EXCLUDE GO terms 2) The appropriate gp2term relations are available for protein-containing complexes as annotation subjects 3) Annotations are displaying properly on the form table

GO terms were removed with the following expression: fq: isa_closure:"CHEBI:33695" OR isa_closure:"GO:0032991" OR NOT idspace:"GO"

@kltm @balhoff - is this the correct way to filter out GO terms?

[kltm]

@vanaukenk As in something in a closure, or a specific term? The answer should be "yes" to both, but the latter may be more sensibly done at the UI end.

[vanaukenk]

This is now fixed. GO protein-containing complex terms are NOT available as subjects of standard annotation