search

ga4gh-beacon / beacon-v2-Models

Models that leverage the Beacon Framework v2
Apache License 2.0
4 stars 7 forks source link

Ensure consistent use of `geneSymbol` or similar alternative #28

Closed mbaudis closed 3 years ago

mbaudis commented 3 years ago

The models contain various parameters which in essence refer to a gene symbol/name, e.g.

... (other instances?).

I propose to use geneSymbol and geneSymbols throughout (or alternatively geneId | geneIds, if other identifiers would be permitted ... but IMO not). Solely gene is too fuzzy.

jrambla commented 3 years ago

Apart from the criteria of longer names in bodies and shorter in query string, not limiting to symbols was on purpose, given my ignorance on how much we could/should enforce using just the symbols. Could we? Having the opinion of the community would be interesting.

mbaudis commented 3 years ago

geneId(s) would be shorter & more permissive but may be problematic itself - I'd rather have something like this using ontologyTerm objects for scoping (again, not totally opposed).

However, the geneSymbol is in line with the current description for gene:

 "description": "Gene symbol following the HGNC (https://www.genenames.org) nomenclature.",

And I don't by into the length argument since

So 6 characters more isn't really an issue, given the can of worms opened by something like gene.

jrambla commented 3 years ago

@mbaudis @antbro manuel.rueda@crg.eu Skipping the GET length thingy... could we enforce using only gene symbols? or would this fireback with complaints from implementers?

jrambla commented 3 years ago

IMU, according to the GA4GH Beacon call this week, geneIdand geneIdsare preferable as we should not stick to HGVS gene symbols. I'll update the genomicVariation files to reflect that. The change will be reflected in an incoming PR.