search

broadinstitute / seqr

web-based analysis tool for rare disease genomics
GNU Affero General Public License v3.0
176 stars 89 forks source link

Frameshift variant filtered out incorrectly #877

Closed gvannoy closed 5 years ago

gvannoy commented 5 years ago

In RGP_256 an initial search with parameters: [Inheritance - Recessive, Location - DEGS1, Call Quality - GQ: 20, AB: 25] yields 16 variants: 12 intronic, 3 3prime UTR, and 1 frameshift (search link: https://seqr.broadinstitute.org/variant_search/results/8cfd5bc58a0d726d4e423d414cd2f521?page=1&sort=pathogenicity)

A second search with the same parameters as above PLUS [Annotations: Frameshift] yields 0 variants. (search link: https://seqr.broadinstitute.org/variant_search/results/8ef2ac697460ccf471457910c905bc5f?page=1&sort=pathogenicity)

hanars commented 5 years ago

This is because in that first search the variant was returned as part of a compound het. When you apply the frameshift filter, it only looks for compound hets in variants with the filtered consequences, and since there is only one hit with that consequence that no longer counts as a compound het and so nothing is returned.