"FilteredInAll" VariantEval category

eflynn90eflynn90 Washington DCMember

I am running CombineVariants followed by VariantEval on two VCF files (GATK version 3.1-1). I noticed that none of my "set=FilteredInAll" variants are included in the output of VariantEval. I suspect this is because the FILTER column is not "PASS" because the variants were not "PASS" in either of the input VCF files.

Is there an argument to pass into VariantEval so that it does not filter variants out based on the FILTER column? I was using the arguments from a guide article (http://gatkforums.broadinstitute.org/discussion/48/using-varianteval). I saw that in the comments you mentioned an --ignoreFilters argument, but when I tried to use it I got the GATK error: Argument with name 'ignoreFilters' isn't defined.

I can just change the FILTER column entries in the combined VCF file, but it seems like a problem that others might have as well.

Thank you!

Answers

  • eflynn90eflynn90 Washington DCMember

    Just saw the "Bug Bulletin"; I'll retry with 3.2 and post an update.

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    Hi @eflynn90,

    Good on you for updating to 3.2 but as it happens, in this case I doubt you'll see a difference (there were no issues in VariantEval 3.1 as far as we can tell).

    A quick clarification about how it works: having the FILTER field be something other than PASS is precisely what signals to VariantEval that a variant is filtered.

    Can you please post the commands you ran to make your evaluation? Starting from the CombineVariants step if you were following the article exactly, and including a few representative lines from each VCF as well as the relevant output table. Hopefully with that I can help you figure out what's wrong.

  • eflynn90eflynn90 Washington DCMember
    edited July 2014

    3.2-2 results were the same

    (oops didn't see your comment until after I posted!)

    Post edited by eflynn90 on
  • eflynn90eflynn90 Washington DCMember
    edited July 2014

    #!/bin/bash

    gatk.sh 20 -T CombineVariants
    -R /data/Udpbinfo/WES/DiploidAlign/GATK-Resources/hs37d5.fa
    -V:IT0 IT0/final.VScompat.tranchemod.vcf
    -V:IT1 IT1/final.VScompat.tranchemod.vcf
    --genotypemergeoption UNIQUIFY
    -o IT0IT1.combined.vcf

    gatk.sh 20 -T VariantEval
    -R hs37d5.fa
    -D dbsnp_137.b37.vcf
    -select 'set=="Intersection"' -selectName Intersection
    -select 'set=="IT0"' -selectName IT0
    -select 'set=="IT1"' -selectName IT1
    -select 'set=="filterInIT0-IT1"' -selectName IT1-filteredinIT0
    -select 'set=="IT0-filterInIT1"' -selectName IT0-filteredinIT1
    -select 'set=="FilteredInAll"' -selectName FilteredInAll
    -o IT0IT1.combined.eval.txt
    -eval IT0IT1.combined.vcf

  • eflynn90eflynn90 Washington DCMember
    edited July 2014

    grepped for some "FilteredInAll" variants in the combined VCF file:

    1 14464 . A T 188125.27 VQSRTrancheSNP99.90to100.00 AC=86;AF=0.307;AN=280;DP=31289;InbreedingCoeff=-0.4433;MLEAC=43;MLEAF=0.307;MQ0=0;NEGATIVE_TRAIN_SITE;culprit=MQRankSum;set=FilteredInAll GT:AD:DP:GQ:PL ./.:.:0 ./.:.:0 ./.:.:0 ./.:.:0 ./.:.:0 ./.:.:0
    1 14470 . G A 105.62 VQSRTrancheSNP99.90to100.00 AC=4;AF=0.014;AN=284;DP=23938;FS=1.965;MLEAC=2;MLEAF=0.014;MQ0=0;NEGATIVE_TRAIN_SITE;culprit=QD;set=FilteredInAll GT:AD:DP:GQ:PL ./.:.:0 ./.:.:0 ./.:.:0 ./.:.:0 0/0:.:1:3:0,3,37 0/0:.:1:3:0,3,36
    1 14513 . G A 108.40 VQSRTrancheSNP99.90to100.00 AC=6;AF=0.021;AN=292;DP=24217;FS=20.301;MLEAC=1;MLEAF=6.849e-03;MQ0=0;NEGATIVE_TRAIN_SITE;QD=0.56;culprit=QD;set=FilteredInAll GT:AD:DP:GQ:PL 0/0:.:2:6:0,6,62 0/0:.:2:6:0,6,61 0/0:.:2:6:0,6,63 0/0:.:2:6:0,6,60 0/0:.:4:0:0,0,34 0/0:.:4:0:0,0,30
    1 14522 . G A 36190.29 VQSRTrancheSNP99.90to100.00 AC=124;AF=0.425;AN=292;DP=25831;MLEAC=63;MLEAF=0.432;MQ0=0;NEGATIVE_TRAIN_SITE;culprit=FS;set=FilteredInAll GT:AD:DP:GQ:PL 0/0:.:2:6:0,6,62 0/0:.:2:6:0,6,61 0/1:2,1,0:3:20:20,0,163 0/1:2,1,0:3:20:20,0,161 0/0:.:5:0:0,0,93 0/0:.:5:0:0,0,96

  • eflynn90eflynn90 Washington DCMember
    edited July 2014

    After combining the files, the FILTER column in all the "FilteredInAll" variants is "VQSRTranche....", while the FILTER column in all the non-"FilteredInAll" variants is "PASS".

    The number of JexlExpression=none and Novelty=all variants in the VariantEval table is only equal to the number of FILTER="PASS" variants in combined VCF file.

  • eflynn90eflynn90 Washington DCMember
    edited July 2014

    #:GATKReport.v1.1:8
    #:GATKTable:11:21:%s:%s:%s:%s:%s:%d:%d:%d:%.2f:%d:%.2f:;
    #:GATKTable:CompOverlap:The overlap between eval and comp sites
    CompOverlap CompRod EvalRod JexlExpression Novelty nEvalVariants novelSites nVariantsAtComp compRate nConcordant concordantRate
    CompOverlap dbsnp eval FilteredInAll all 0 0 0 0 0 0
    CompOverlap dbsnp eval FilteredInAll known 0 0 0 0 0 0
    CompOverlap dbsnp eval FilteredInAll novel 0 0 0 0 0 0

    CompOverlap dbsnp eval IT0-filteredinIT1 all 1961 1036 925 47.17 916 99.03
    CompOverlap dbsnp eval IT0-filteredinIT1 known 925 0 925 100 916 99.03
    CompOverlap dbsnp eval IT0-filteredinIT1 novel 1036 1036 0 0 0 0
    CompOverlap dbsnp eval IT0 all 1795 1325 470 26.18 442 94.04
    CompOverlap dbsnp eval IT0 known 470 0 470 100 442 94.04
    CompOverlap dbsnp eval IT0 novel 1325 1325 0 0 0 0
    CompOverlap dbsnp eval IT1-filteredinIT0 all 2740 1468 1272 46.42 1244 97.8
    CompOverlap dbsnp eval IT1-filteredinIT0 known 1272 0 1272 100 1244 97.8
    CompOverlap dbsnp eval IT1-filteredinIT0 novel 1468 1468 0 0 0 0
    CompOverlap dbsnp eval IT1 all 1477 1068 409 27.69 373 91.2
    CompOverlap dbsnp eval IT1 known 409 0 409 100 373 91.2
    CompOverlap dbsnp eval IT1 novel 1068 1068 0 0 0 0
    CompOverlap dbsnp eval Intersection all 307966 69934 238032 77.29 230365 96.78
    CompOverlap dbsnp eval Intersection known 238032 0 238032 100 230365 96.78
    CompOverlap dbsnp eval Intersection novel 69934 69934 0 0 0 0
    CompOverlap dbsnp eval none all 315939 74831 241108 76.31 233340 96.78
    CompOverlap dbsnp eval none known 241108 0 241108 100 233340 96.78
    CompOverlap dbsnp eval none novel 74831 74831 0 0 0 0

  • eflynn90eflynn90 Washington DCMember

    Sorry, I'm having trouble getting the spacing to work correctly.

Sign In or Register to comment.