GenotypeGVCFs using output from WholeGenomeGermlineSingleSample_v3.0.0

Post author
Don Conrad

I've created GVCFs using the WholeGenomeGermlineSingleSample_v3.0.0 workflow, and then processed these with CombineGVCFs and GenotypeGVCFs, all with default settings ("Functional Equivalence" version for the first step).

I’ve run into this behavior that very many of the sites with appreciable read depth (say > 20X) and 100% reference reads are being called as “./.” instead of “0/0”. I’m just running things with default settings, so I wouldn’t think these are being filtered in some way.

Here is an example of what I mean, based on a test with 4 samples

chr1    63643   .       A       G       90      .       AC=2;AF=0.500;AN=4;BaseQRankSum=1.63;DP=109;ExcessHet=4.7712;FS=2.766;MLEAC=2;MLEAF=0.500;MQ=22.46;MQRankSum=-6.030e-01;QD=1.80;ReadPosRankSum=1.51;SOR=1.329   GT:AD:DP:GQ:PL  0/1:10,5:15:29:57,0,29  0/1:26,9:35:29:45,0,29  ./.:29,0:29:40  ./.:30,0:30:40

the last two samples have read depth of 29 and 30, but are set to ./.

I’d prefer to have these called as “0/0” for a number of reasons.  Any ideas of how to change this behavior?

 

Comments

2 comments

  • Comment author
    Anika Das

    Hi Don, 

    Thank you for writing in! We will take a look at this and get back to you as soon as we can!

    Kind Regards, 

    Anika

    0
  • Comment author
    Samantha (she/her)

    Hi Don Conrad,

    This question is a bit out of scope for Terra support and better suited for the GATK forum.

    That said, I reached out to a colleague on the GATK support side and they informed me that "./." means no call, which is different from "0/0" (https://samtools.github.io/hts-specs/VCFv4.2.pdf).

    If you have additional questions, I would suggest posting to the GATK forum.

    Best,

    Samantha

    0

Please sign in to leave a comment.