In my VCF, what specifically does the filter Indel_FS mean?

dauber

I am filtering looking for rare variants and found some frameshift variants in an interesting gene. Some of them are noted as PASS in the QC column of the VCF and some are noted as Indel_FS . What exactly does that second notation mean? I am almost positive that these will validate given how they segregate in my subjects.

pdexheimer

There are no standard filters in VCF, they're described in the meta-information lines of each file (see the spec: http://www.1000genomes.org/wiki/Analysis/Variant Call Format/vcf-variant-call-format-version-41).

If I had named that filter, it would be for Indel calls that fall outside some threshold on the FS metric - exactly what the threshold is should be defined in the meta-info.

dauber

Thanks for the heads up and quick reply. The metadata says Indel_FS,Description="FS>200.0" . Where can I find more info about the FS metric? Thanks. Andrew

pdexheimer

It's also defined in the meta-information, it's a measure of strand bias. There's documentation on this site (Guide/Technical Documentation/VariantAnnotator Annotations/FisherStrand), but it's pretty brief. If you really want to understand it, check out the source.

Geraldine_VdAuwera

Hi Andrew,

Here's the direct link to the doc in question:

http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_annotator_FisherStrand.html

Let me know if you know more detailed information, since the annotator docs are indeed rather thin.