cannot import existing annotations from GO database into a GO_CAM

[Pauldenny]

When trying to import existing annotations from GO database for P25051 e.g. these IEAs

noctua seems unable to find them.

[Pauldenny]

Kimberley @vanaukenk suggested this issue is a result of AMIGO2 including only EXP annotations for some species - would it be possible for these species to be added to the shortlist for which other non-EXP annotations could be included?
I am working on antimicrobial resistance, which can involve many non-reference / non-MOD species:

E. coli [taxID=562] - Not just the K12 MOD
Staphylococcus aureus [1280]
Klebsiella pneumoniae (High priority organism - WHO GLASS report 2022) [573]
Acinetobacter baumanii [470]
Pseudomonas aeruginosa - particularly high risk for multi-drug resistance [287]
Enterobacteriaceae [543]

NB - the reason for the high-level taxa requested is that many papers use bacterial species further down the taxonomic "tree" and so it's simpler to capture these as they will share a taxonomic parent, rather than try to second-guess an even larger number of more specific taxa?

[pgaudet]

Hi @Pauldenny
@vanaukenk correctly identified the issue you are observing - we only have EXP annotations for P25051

https://amigo.geneontology.org/amigo/gene_product/UniProtKB:P25051

I dont think we can load the additional data in the very short term, so you'll have to make that annotation again in Noctua.

WRT strains versus species, I am also in favor of GO central being strain-specific, and moving all 'strain' annotations up to the species. However that involves Uniprot chaning their reference proteomes, as far as I know.

Again, this is on the radar, but it may take a ltitle while to get done. However these use cases are extremely valuable in deciding which way to go.

Thanks for your patience - sorry we cannot help more in the immediate future.

@kltm anything you'd like to add? or correct if I stated anything incorrectly.

[Pauldenny]

One point I didn't explain clearly:
When I asked for all annotations to these taxa to become accessible in AMIGO2/Noctua, I meant that, ideally, I would want to be able to also access / annotate any "descendants" / children of these taxa e.g. taxID=562 has 2,926 direct descendants.
@kltm @pgaudet @vanaukenk

[pgaudet]

humm I'll check with @alexsign how this can be done

Do you need to annotate the strains? or are the species ok ? ie, could we map the annotations to the strains up to the species? (note that I dont even know if that's doable ;)

[Pauldenny]

@pgaudet yes - I'm not sure this could be done - UniProt is also about to change it's rules on importing new Proteome sets. However, reviewed entries already in Swiss_prot shouldn't be affected by the proteome changes.

Just hit a possibly related glitch:
I can see this in AMIGO https://amigo.geneontology.org/amigo/gene_product/UniProtKB:P52477
But the UniProt accession isn't recognised by Noctua.
Instead Noctua seems to be using a synonymous identifier derived from an Ordered locus name ID - PA0425

Confusing - and Noctua still doesn't find GO annotations displayed by AMIGO...

[kltm]

@Pauldenny The entities found in the autocomplete of Noctua and the annotations found in AmiGO have slightly different derivations, which is what I think is causing UniProtKB:P52477 to not show up in the autocomplete.
Grossly: these IBAs in AmiGO are coming in from PANTHER; the entities in the Noctua autocomplete are coming in from either self-created files by the resources for entities that they desire to annotate or (if that's not available) the upstream (pre-IBA) GAF from the resource.