Update clingen prefixes

[cmungall]

Bioregistry has both

1. clingene (alleles)
2. clingen.affiliation (added in New prefix: clingen.affiliation #1314)

I think clingene is a mistake inherited from identifiers.org. I think clingen should be primary, clingene could be a synonym (even though I have never seen it written this way)

While we are breaking things, I would advocate for not assuming the top level refers to alleles and instead make it clingen.allele. Clingen have other kinds of identifiers, in particular curation IDs.

I'm also not sure the affiliation is the best term for the subnamespace. I know that is what is used in the URL but on the actual page itself we see that this is an ID for a CGEP (ClinGen expert panel). I will check with them.

My overall proposal would be:

1. clingen.allele
2. clingen.curation
3. clingen.cgep

[bpow]

I agree that clingene is an error, and should be clingen.

It also makes sense to me to avoid the top-level. clingen.allele and clingen.curation could make sense, but it may be worth having a discussion about what "curation" would be used to represent. This is one of those words that is a process but also the product of the process, which is the first thing that could lead to confusion. More to the point, I think of clingen as producing multiple types of data resources: curations that represent evidence aggregation, and assertions (or assessments?) that are the results of the expert panels weighing those evidence according to standardized plans/procedures. The plans/procedures are actually another resource produced by clingen (e.g., in the criteria specification (cspec) registry and in many SOPs for curation/assessment processes).

I don't know that bioregistry needs a namespace for the plan/procedure artifacts, but would be interested on thoughts on whether having a curation subnamespace separate from assertions might be warranted, or alternatives. Are the assertions what we are really most interested in conveying in this context (with curations being the evidentiary support)? Or is there a term we could use to combine both items more generically?

Regarding the affiliation subnamespace, some of this is because we have different types of groups that produce these resources in clingen. Essentially, these represent organizational agents from a provenance perspective. @cmungall proposed "cgep" (ClinGen Expert Panel?), which is curiously similar to "GCEP" (our Gene Curation Expert Panel), separate from VCEP (Variant Curation Expert Panel) or one of the other working groups that produce artifacts for clingen (Actionability Working Groups, Dosage Sensitivity Working Group). The "Expert Panel" / "Working Group" dichotomy in ClinGen is itself a bit historical/operational (the working groups are panels of experts, but Expert Panels in ClinGen have a certain application/approval process). I'd actually be OK calling all of them expert panels from an external standpoint, but that's me speaking as an individual without running it by others to think through implications.

So that's context for discussion, I'll discuss among others within ClinGen to get their thoughts.

[bgyori]

I agree with all the points except renaming clingen.affiliation to clingen.cgep: given @bpow's explanation, in my view, this could cause some confusion. I think it's best to stick to the terminology that is used by ClinGen in the URL pattern, which is affiliation, unless ClinGen changes the terminology on their end.

[courtneythaxton]

I would also favor clingen.affiliation as we have groups other than Gene Curation Expert Panels (GCEPs) that may work to help provide a ClinGen_label in Mondo.

[matentzn]

@courtneythaxton

that may work to help provide a ClinGen_label in Mondo.

That may not mean much to @bgyori and the bioregistry team :D (I mean, they wont know what you are talking about)

[bgyori]

Following up on this, @nagutm created #1412 to resolve the clingene->clingen.allele renaming. We also started looking into the proposal for clingen.curation and wanted to check in on the best way to capture that. Would it be useful to curate identifiers like 1806e24c-7e4f-4774-ac6d-77bacb921cdb which resolve on pages like https://erepo.clinicalgenome.org/evrepo/ui/classification/1806e24c-7e4f-4774-ac6d-77bacb921cdb as "curations"? Or did @cmungall have something else in mind?

[cmungall]

@bgyori that link is broken for me, here's an example of a curation: https://search.clinicalgenome.org/CCID:004126

[nagutm]

Thanks for the clarification @cmungall, you can see #1414 which curates clingen.curation