Adding Clinvar to genome nexus

[jjgao]

17:g.41276045_41276046del is the most prevalent pathogenic germline mutation, but the links to ClinVar and dbSNP is missing. It is linked to gnomAD, which contains link to dbSNP and ClinVar.

Similarly: https://www.signaldb.org/variant/17:g.41209079_41209080insG

Solution: Adding clinvar to genome nexus.

• [x] Import clinvar VCF to genome nexus
• [x] Update genome nexus to query clinvar
• [ ] Find a way to automatically update clinvar data (github action or some other ways) (Create separate issue)

[leexgh]

@jjgao We get ClinVar from MyVariantInfo, but they don't have ClinVar for this particular variant. Actually I already asked them about the same variant before: https://github.com/biothings/myvariant.info/issues/106. Haven't heard back from them yet.

[inodb]

There might also be new updates to Ensembl's clinvar data so we might be able to use that over myvariant.info

[leexgh]

https://grch37.ensembl.org/info/genome/variation/species/sources_documentation.html GRCH37 ClinVar version: 12/2019

[leexgh]

1. XML:
   • Pros: Full data
   • Cons: Too big to parse(15GB); Genomic coordinate is not unique; Hard to find a tool to do streaming parsing.
2. VCF:
   • Pros: Easy to parse
   • Cons: Missing clinical significance data.
3. API
   • Pros: Full data
   • Cons: Have access control(3 queries per second); Still need a tool to parse XML.

[leexgh]

We decide to use VCF. We should be consistant with clinVar interpretation and have linkout to their website, therefor using VCF is better.