Specifying MaveDB Data with VA-Spec

[bencap]

Submitter Name

Ben Capodanno

Submitter Affiliation

University of Washington

Project Details

With new functionality to map new and existing variant objects to human genome, the generation of VRS objects to represent these pre- and post-mapped variants, and new structured score set data to handle information about score range and variant classification, MaveDB would benefit from integrations with VA-Spec that would provide a way to export this information in a consistent, structured format.

For VA-Spec, this would be a good opportunity to develop or tweak VA profiles relevant to MAVE data and for MaveDB to adopt a reference implementation of VA-Spec information models that could provide an example for other adopters looking to implement VA-Spec on their projects.

Required Skills

• Python
• Pydantic
• VA-Spec

[bencap]

Work during the hackathon focused on bringing forward the MaveDB implementation of VA-Spec based on an earlier version to the latest ballot release. Attached below is a draft VA object for a variant in MaveDB, which is heavily inspired from the ACMG Variant Pathogenicity Statement. MaveDB can export evidence lines for data related to pathogenicity, which is supported by a statement on variant function, an evidence line for variant function, and a study result for a MaveDB variant.

Some implementation details and work remain outstanding, but the last few days have moved MaveDB x VA-Spec integration forward significantly and we're confident we'll be able to deliver MaveDB data in the spec to consumers in the near future.

va-draft-by-variant.json