Representing intronic variants in VRS

[Mrinal-Thomas-Epic]

When working on representing molecular consequences from VEP in VA-spec (PR 336), we hit some friction while representing intronic variant molecular consequences. We can represent the transcript (using SequenceReference) and molecular consequence terms, but aren't able to represent intronic variant (currently returned as an HGVS by VEP, e.g. ENST00000650946.1:n.439-1599G>C)

[ahwagner]

Today we discussed initial drafts for this as part of a broader model for representing variants on any set of aligned sequences. The conceptual model draft is publicly available here.

It was noted that we will need a mechanism for describing alignment orientation in alignmentBlocks, and that we should consider scenarios where you have 1-to-many sequence alignments as commonly seen in liftover operations.

[Mrinal-Thomas-Epic]

In the meeting today, we discussed data structures to describe aligned alleles, sequence locations, and sequence references. I was wondering if its needed to provide alignments across the whole sequence reference when describing a single aligned allele. For example,