We've moved!
This site is now read-only. You can find our new documentation site and support forum for posting questions here.
Be sure to read our welcome blog!

VQSR for indels

pedropedro Member
edited January 2013 in Ask the GATK team


I am running Variant Quality Score Recalibration on indels with the following command.

java -Xmx8g -jar /raid/software/src/GenomeAnalysisTK-1.6-9-g47df7bb/GenomeAnalysisTK.jar \
    -T VariantRecalibrator \
    -R /raid/references-and-indexes/hg19/bwa/hg19_lite.fa \
    -input indel_output_all_chroms_combined.vcf \
    --maxGaussians 4 -std 10.0 -percentBad 0.12 \
    -resource:mills,known=true,training=true,truth=true,prior=12.0  /raid/Merlot/exome_pipeline_v1/ref/Mills_and_1000G_gold_standard.indels.hg19.sites.vcf \
    -an QD -an FS -an HaplotypeScore -an ReadPosRankSum  \
    --ts_filter_level 95.0 \
     -mode INDEL \
    -recalFile /raid2/projects/STFD/indel_output_7.recal \
    -tranchesFile /raid2/projects/STFD/indel_output_7.tranches \
    -rscriptFile /raid2/projects/STFD/indel_output_7.plots.R

My tranches file reports only false positives for all tranches. When I run VQSR on SNPS, the tranches have many true positives and look similar to other tranch files reported on this site. I am wondering if anyone has similar experiences or suggestions?


Post edited by Geraldine_VdAuwera on


  • darbrobdarbrob Member

    Hey - for what its worth - I've been using those settings for indel VQSR and I also observe that the plots don't look as good as the ones for SNVs. Sadly, I don't know what settings do work better. Maybe you could try calling indels with multiple callers, taking the intersection, and then creating a "per sample" gold standard set of indels to use in place of the mills ones (but maybe using the mills indels as an additional training set).

  • Same issue here using UG and GATKv 2.2 (10 whole genomes, Illumina platform)

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    Hi folks, if you mean your tranches plots look bad when you run on indels, don't worry -- it's completely expected, as tranches plots are only meant to be used for SNPs. They are meaningless for indels; you shouldn't be looking at them at all. We're going to clarify this in the docs, and in the next version (2.4) the tranches plots won't be generated at all anymore. Sorry about the confusion, and not picking this thread up earlier. It slipped through the net somehow.

  • blueskypyblueskypy Member ✭✭

    so -tranchesFile option should not be used with indel in VariantRecalibrator and ApplyRecalibration?

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    @blueskypy, the tranches file is necessary for the recalibration. It's just that the plots only apply to SNPs, but tranches as such are used for both types of variants.

Sign In or Register to comment.