Issue in PomBase GPI causes issues with NEO pipeline (duplicate entries)

[kltm]

In the file https://www.pombase.org/data/annotations/Gene_ontology/pombase.gpi.gz, lines like

PR:000060241 tht1 nuclear membrane protein involved in karyogamy Tht1 SO:0001217 NCBITaxon:4896 PomBase:SPAC13C5.03 PomBase:SPAC13C5.03.1.pep UniProtKB:Q09684 go-annotation-summary=nuclear membrane protein involved in karyogamy Tht1
PR:000060241 akr1 palmitoyltransferase Akr1 SO:0001217 NCBITaxon:4896 PomBase:SPAC2F7.10 PomBase:SPAC2F7.10.1.pep UniProtKB:Q09701 go-annotation-summary=palmitoyltransferase Akr1

generate obo like

[Term]
id: PR:000060241
name: tht1 Spom
synonym: "tht1" BROAD []
synonym: "000060241" RELATED []
synonym: "SPAC13C5.03.1.pep" RELATED []
xref: PomBase:SPAC13C5.03.1.pep
is_a: CHEBI:33695 ! information biomacromolecule
relationship: in_taxon NCBITaxon:4896
property_value: https://w3id.org/biolink/vocab/category https://w3id.org/biolink/vocab/MacromolecularMachine
property_value: https://w3id.org/biolink/vocab/category https://w3id.org/biolink/vocab/GeneProduct
relationship: has_gene_template PomBase:SPAC13C5.03

[...]

[Term]
id: PR:000060241
name: akr1 Spom
synonym: "akr1" BROAD []
synonym: "000060241" RELATED []
synonym: "SPAC2F7.10.1.pep" RELATED []
xref: PomBase:SPAC2F7.10.1.pep
is_a: CHEBI:33695 ! information biomacromolecule
relationship: in_taxon NCBITaxon:4896
property_value: https://w3id.org/biolink/vocab/category https://w3id.org/biolink/vocab/MacromolecularMachine
property_value: https://w3id.org/biolink/vocab/category https://w3id.org/biolink/vocab/GeneProduct
relationship: has_gene_template PomBase:SPAC2F7.10

Naturally, having duplicate entries causes parsing and resolution issues.

The following identifiers have multiple entries in the GPI:

      2 PR:000060241
      3 PR:000060258

Tagging @ValWood @kimrutherford @pgaudet @vanaukenk

[kltm]

Noting that this will block the NEO pipeline until we have a fixed upstream version or clean a version and point to that in the metadata instead.

[ValWood]

Oh shit, I have this in my inbox. I didn't think it was critical because I had assumed it was referring to different PRO iDs for the same gene, but this is basically an annotation error.

[ValWood]

I will fix it tomorrow, I have run out of energy today. But it is impressive that both pipelines pick it up ;)

[kltm]

No worries! I just wanted to make sure this didn't accidentally fly under the radar.

[ValWood]

Hi @nataled
I may need your help here.

https://proconsortium.org/app/entry/PR_000060241/
refers to
Palmitoylase=("ark1"; PR:000060242;
but this looks incorrect. Should it be
PR:000060240
PR:000060240

I'm not sure that this fixes the issue (because there might be some confusion between akr1 (palmitoyltransferase) and ark1 (protein kinase) too, either at my end, or your end, I will dig deeper)

v

[ValWood]

The annotation issue was that we had applied the modification to akr1

I deleted this. We only use the modified forms with the entities, not their substrates.

[nataled]

@ValWood is there anything still needed on my end, or does your action fix the issue?

[ValWood]

PRO is all correct except Palmitoylase=("ark1"; PR:000060242;
which I now realise refers to the palmitoylase modifying tht1 should be akr1 here in the comment:

Comment | Requested by=PomBase. Palmitoylase=("ark1"; PR:000060242; Cys-65/Cys-78). Evidence=(ECO:0001091, based on PMID:36650056).

let me know if you want me to transfer this to the original ticket.

[kimrutherford]

Hi All. The PomBase GPI file should be updated with Val's fix by 4am GMT tonight.

[nataled]

@ValWood already fixed in local copy of PRO, but for completeness I'll add a note to the original ticket submitted to PRO.

[kltm]

Hi All. The PomBase GPI file should be updated with Val's fix by 4am GMT tonight.

Cheers! This should get "automatically" tested this Wednesday.

[kltm]

The test cleared and we are building again. Thank you!