ga4gh / va-spec

An information model for representing variant annotations.
Apache License 2.0
17 stars 4 forks source link

Variant pathogenicity statements should have a classification #208

Closed Mrinal-Thomas-Epic closed 1 month ago

Mrinal-Thomas-Epic commented 1 month ago

Variant pathogenicity is reported using classification terms like "Pathogenic", "Benign", etc. There used to be a property for this, but I can no longer find it.

ahwagner commented 1 month ago

Do you mean as an enum? I see the classification field here, for example. Though this has been turned into a Coding and some examples where such a code is provided would be helpful. And if a coding doesn't exist, it might be worth recreating the Mappable Entity class that allowed for non-coded entities so they may be representable as strings.

Mrinal-Thomas-Epic commented 1 month ago

The property you linked to seems like a good approach to me. I couldn't find the property in the 1.0.0-connect.2024-09 schema for va-spec (or gk-pilot main), so I was trying to use this issue to ask, where did it go in the pre-release?

Edit: I see it in the JSON schema in the 1.0.0-connect.2024-09, just don't see where its coming from in the source yaml

Mrinal-Thomas-Epic commented 1 month ago

Nevermind, I see it now in core-im-source.yaml. Although I am curious why its on all statements, rather than a special type of ClassificationStatement like before

ahwagner commented 1 month ago

I think that @mbrush and @larrybabb flattened a lot of the VA Spec abstract classes in their rework leading up to Connect.

I agree that the "Statement" class doesn't seem like a great fit for the classification attribute, and this is a decision I would also like some additional clarity / refinement on.