search

geneontology / neo

noctua entity ontology
9 stars 2 forks source link

Is NEO typing PRO identifiers correctly? #64

Closed vanaukenk closed 3 years ago

vanaukenk commented 3 years ago

@hdrabkin showed me a model this morning where the ShEx is not validating PRO identifiers as chemical entities, i.e. 'has input' PR:nnnnnnnnnnn, for a MF gives a ShEx validation.

How are PRO identifiers currently typed in neo?

balhoff commented 3 years ago

Can you paste some specific examples? I looked in NEO and all the PR terms I checked should fall under some CHEBI. Some are under protein while some are under amino acid chain.

deustp01 commented 3 years ago

some are under amino acid chain.

That looks like wrong chemistry. CHEBI:50047 "a compound formally derived from ammonia by replacing one, two or three hydrogen atoms by organyl groups" looks like a class of small molecules whose members include, for example, the various amino acid residues found in proteins.

vanaukenk commented 3 years ago

@hdrabkin - can you paste the PR ids for the specific examples you showed me? Thanks!

hdrabkin commented 3 years ago

The two in question were actually 'new' (they were the ones that were not in neo for several weeks and then the load was fixed:

[Term] id: PR:000050486 name: alanine--tRNA ligase, cytoplasmic isoform 1 methylated 1 (mouse)

PR:000050487 name: alanine--tRNA ligase, cytoplasmic isoform 1 unmethylated 1 (mouse)

balhoff commented 3 years ago

I think @hdrabkin's comment explains things. I'll close, but let me know if this continues to be a problem.