To celebrate the release of GATK 4.0, we are giving away free credits for running the GATK4 Best Practices pipelines in FireCloud, our secure online analysis portal. It’s first come first serve, so sign up now to claim your free credits worth $250. Sponsored by Google Cloud. Learn more at

VQSR with missing annotation fields

I am calling variants (non-model organism) following the best practice workflow. After haplotypecaller (with GVCF) and GenotypeGVCFs, I want to perform VQSR (separately for SNPs and INDELs) to the raw vcf file (10 samples) as an alternative for filtering manually. As a resource I use a subset of very high quality variants (obtained with hard filtering, coming from different samples). However, I have noticed that not all annotation fields are present in the raw vcf file. For example, not all sites have a MQRankSum field, ReadPosRankSum field or MQ field. As I want to use these annotations for recalibration, I was wondering how the model handles such sites (I have been using GATK3.5 so far) and how this would affect the filter field.
I guess running variantAnnotator on the raw vcf file will not add the missing fields (for example ReadPosRankSum cannot be calculated when no reads are found with the reference allele).
Thanks a lot for your help!

Issue · Github
by Sheila

Issue Number
Last Updated


  • SheilaSheila Broad InstituteMember, Broadie, Moderator


    Sorry for the late response.

    With only 10 samples, it will be very hard to make a solid known sites resource file. Please validate your results thoroughly, and definitely compare your results to the results you get from hard filtering.

    If you do not have some annotations at some sites, those annotations will be ignored at the site. This should not affect the filtering results too much, as long as you have some other annotations that are used at the site.

    You are correct VariantAnnotator cannot annotate the Rank Sum tests when there is no evidence for both ref and alt alleles.


Sign In or Register to comment.