search

HUPO-PSI / psi-mod-CV

Reactivated PSI-MOD ontology for molecular mass modifications
Creative Commons Attribution 4.0 International
8 stars 13 forks source link

UniProt PTM references are not written as CURIEs #68

Closed cthoyt closed 2 years ago

cthoyt commented 2 years ago

right now, UniProt post-translational modification references (https://bioregistry.io/registry/uniprot.ptm) are written as PTM-XXXX.

This isn't CURIE syntax and doesn't make clear what the resource is that these come from. It would be good to switch them all to CURIE syntax using an appropriate prefix. I will send a PR to demonstrate.

edeutsch commented 2 years ago

Is UniProt on board with this change? We are using PTM-XXXX because UniProt is using PTM-XXXX: https://www.uniprot.org/docs/ptmlist.txt

Will they be changing their ptmlist.txt? They may wish to keep the PTM- since this has been part of their identifiers for a long time?

I'm all for moving to proper CURIEs, but we should make sure they're on board, too.

cthoyt commented 2 years ago

I haven't seen a standardized approach to the way UniProt expects external resources to handle identifiers, and I've seen quite a few that omit these pseudo-prefixes. However, the Bioregistry also has a way of annotating this situation, so I'd also be happy just to slap a prefix in front of these as they were. I updated my PR in https://github.com/HUPO-PSI/psi-mod-CV/pull/69/commits/4213e9531534f25c1810e1daedb4b7705ace362f.

If the question was whether UniProt should sign off on using uniprot.ptm as the prefix itself, then I'd say that it's better to consult a registry since this is more of a descriptive thing than prescriptive thing (i.e., use whatever you want as long as it makes sense to others). However, UniProt hasn't engaged in minting prefixes for its various resources in registries, so I've added several of them myself to the Bioregistry