Heads up:
We’re moving the GATK website, docs and forum to a new platform. Read the full story and breakdown of key changes on this blog.
If you happen to see a question you know the answer to, please do chime in and help your fellow community members. We encourage our fourm members to be more involved, jump in and help out your fellow researchers with their questions. GATK forum is a community forum and helping each other with using GATK tools and research is the cornerstone of our success as a genomics research community.We appreciate your help!

Test-drive the GATK tools and Best Practices pipelines on Terra

Check out this blog post to learn how you can get started with GATK and try out the pipelines in preconfigured workspaces (with a user-friendly interface!) without having to install anything.
We will be out of the office for a Broad Institute event from Dec 10th to Dec 11th 2019. We will be back to monitor the GATK forum on Dec 12th 2019. In the meantime we encourage you to help out other community members with their queries.
Thank you for your patience!

VariantRecalibrator - "N" reference allele only in .recal files

astrandastrand New YorkMember


I ran VariantRecalibrator and ApplyRecalibration, and everything seems to have worked fine. I just have one question: if there are no reference alleles besides "N" in my recalibrate_SNP.recal and recalibrate_INDEL.recal files, and in the "alt" field simply displays , does that mean that none of my variants were recalibrated? Just wanted to be completely sure. My original file (after running GenotypeGVCFs) has the same number of variants as the recalibrated vcf's.


Best Answer


  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    Hi Alva,

    I'm not sure I understand your question. The best thing would be to post a few records so we can see exactly what you mean.

  • astrandastrand New YorkMember

    OK, so for example in the recalibrate_SNP.recal file, I only see this type of information:

    22 16050822 . N <VQSR> . . END=16050822;VQSLOD=-4.1403;culprit=MQ 22 16050933 . N <VQSR> . . END=16050933;NEGATIVE_TRAIN_SITE;VQSLOD=-3.0946;culprit=MQ 22 16051347 . N <VQSR> . . END=16051347;POSITIVE_TRAIN_SITE;VQSLOD=-1.1268;culprit=MQ 22 16051497 . N <VQSR> . . END=16051497;VQSLOD=-2.9562;culprit=MQ 22 16051556 . N <VQSR> . . END=16051556;VQSLOD=-2.4374;culprit=MQ

    In the recalibrate_INDEL.recal file, I see this:

    22 16052167 . N <VQSR> . . END=16052167;VQSLOD=-0.5069;culprit=MQRankSum 22 16052169 . N <VQSR> . . END=16052169;NEGATIVE_TRAIN_SITE;VQSLOD=-0.7924;culprit=SOR 22 16056854 . N <VQSR> . . END=16056854;NEGATIVE_TRAIN_SITE;VQSLOD=-2.0643;culprit=MQRankSum 22 16060236 . N <VQSR> . . END=16060248;POSITIVE_TRAIN_SITE;VQSLOD=4.3946;culprit=FS 22 16060479 . N <VQSR> . . END=16060491;VQSLOD=0.6134;culprit=DP

    I just want to understand the information contained in these recal files.


  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    Are you looking at the recalibration tables? Those just include information about sites that are used by the model. They are not the recalibrated files. The important (and informative) thing to look at is the recalibrated VCFs produced by ApplyRecalibration.

  • astrandastrand New YorkMember

    Yes, I did look at the recalibrated VCFs, like I said in my original post, and those are identical to the original VCF produced using GenotypeGVCFs. They have the same number of sites (I produced a vcf file called recalibrated_snps_raw_indels.vcf and then finally recalibrated_variants.vcf, and they are both identical to the original vcf file).

  • astrandastrand New YorkMember

    OK, thanks, that makes sense. So can future tools/programs recognize the fact that some sites are designated as "PASS" or non-PASS (with the tranche tag)? Or do I have to add the --excludeFiltered flag to make sure that the non-PASS sites are filtered out no matter what?


  • SheilaSheila Broad InstituteMember, Broadie admin

    Hi Alva,

    I think you will have to use the --excludeFiltered flag in future versions as well.


  • astrandastrand New YorkMember

    I see. I guess I'm wondering what the use of doing the recalibration is if we don't end up removing the non-PASS sites in the end (that is, if we don't use the --excludeFiltered flag). If we don't use the flag, then we just have information about whether or not the sites passed, that's it. Why is --excludeFiltered just an option, why aren't the non-PASS sites just removed (assuming that we agree with VQSR's recalibration method)? If we didn't want them to be removed in the end, why go through VQSR at all?

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    Because you might want to do some QC analysis to compare the ones that passed and the ones that didn't, or apply additional filters that may in some cases rescue things filtered out by VQSR, depending on what you know about your specific dataset. The tool is geared to provide maximal flexibility to experienced analysts.

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    I should add -- some of the GATK tools that operate on VCFs will automatically ignore filtered sites; this should be indicated in each tool's Tool Documentation page. Also, you have the option of generating a PASS-only VCF at any time using SelectVariants once you are happy with your filtering. That is the expected workflow.

Sign In or Register to comment.