Thresholds for filtering of somatic variants from genomes

[kpjonsson]

There are hard treshholds in the scripts that filters false positive somatic SNVs/indels. These pertain to the read depth, allele counts and fraction as well as gnomAD population frequencies and PoN counts. What the appropriate values are for these filters needs to be investigated.

Also see: #401–should we construct a gnomAD VCF that combined exome/genome data?

[evanbiederstedt]

I feel this task is also complete, though the larger goal is ongoing

Sorry for kliver på @kpjonsson tår; we could re-open if necessary

[kpjonsson]

We haven't looked at output from genomes in depth, so I'm re-opening this.

I imagine we will have to parameterize some of the thresholds that pertain to read depth.

[evanbiederstedt]

Sorry for the tår stepping :/