Jules Jacobsen 2:34 PM
Alex - what's the reason for having a MoleculeContext?
This has to be one of genomic, transcript or protein, but you can express all of these things in the expressions field for a single VariantDescriptor. (edited)
Alex 3:00 PM
briefly (next meeting in 2 minutes) it is for when you are describing a VRS variation without HGVS. I haven’t had much chance to extend the documentation to demonstrate VRS alleles, but they do not have a p/c/r/g HGVS-style indicator built-in. People that want to pass variants in VRS with such a context can use MoleculeContext.
Also, in some cases (e.g. Allele Registry, CIViC, ClinVar variants) a given variant identifier might have multiple HGVS expressions associated, and you want to indicate the context you think best applies–also useful for that.
Larry Babb 3:01 PM
the moleculeContext is an attribute on the “primary” vrs value object that the descriptor wraps. Whereas the p/c/g/r/n/m on the expressions are expression specific.
Jules Jacobsen 3:21 PM
OK, then it would be useful to have an unknown/unspecified value here too as being an enum it can not be null and therefore needs a default value. It is possible (like in most of our examples) to not have a vrs variation, so you need an unknown or equivalent for these cases. (edited)
Being protobuf, it has a rather unique global scope in enums so the value needs to be unknown_molecule_context as MoleculeContext.unknown will clash with AnythingElse.unknown
This is a very irritating feature. (edited)
I've updated the comment in the schema to // The molecular context of the vrs variation. Must be one of "genomic", "transcript", or "protein".
Alex 4:14 PM
That’s fine. Can we go with unspecified_molecule_context?
from the GA4GH slack thread that originated this change...
Jules Jacobsen 2:34 PM Alex - what's the reason for having a MoleculeContext? This has to be one of genomic, transcript or protein, but you can express all of these things in the expressions field for a single VariantDescriptor. (edited)
Alex 3:00 PM briefly (next meeting in 2 minutes) it is for when you are describing a VRS variation without HGVS. I haven’t had much chance to extend the documentation to demonstrate VRS alleles, but they do not have a p/c/r/g HGVS-style indicator built-in. People that want to pass variants in VRS with such a context can use MoleculeContext. Also, in some cases (e.g. Allele Registry, CIViC, ClinVar variants) a given variant identifier might have multiple HGVS expressions associated, and you want to indicate the context you think best applies–also useful for that.
Larry Babb 3:01 PM the moleculeContext is an attribute on the “primary” vrs value object that the descriptor wraps. Whereas the p/c/g/r/n/m on the expressions are expression specific.
Jules Jacobsen 3:21 PM OK, then it would be useful to have an unknown/unspecified value here too as being an enum it can not be null and therefore needs a default value. It is possible (like in most of our examples) to not have a vrs variation, so you need an unknown or equivalent for these cases. (edited) Being protobuf, it has a rather unique global scope in enums so the value needs to be unknown_molecule_context as MoleculeContext.unknown will clash with AnythingElse.unknown This is a very irritating feature. (edited) I've updated the comment in the schema to // The molecular context of the vrs variation. Must be one of "genomic", "transcript", or "protein".
Alex 4:14 PM That’s fine. Can we go with unspecified_molecule_context?
Jules Jacobsen 5:36 AM Works for me.