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!

VariantAnnotator and Mutect2

BogdanBogdan Palo Alto, CAMember ✭✭

Dear Sheila, and Geraldine,

continuing our previous conversation about the fields in the VCF file produced by MUTECT2, please could you advise on why I see the following DIFFERENCES :

-- shall I run MUTECT2 with the options -A for FS, MQ, SOR, I am getting the following numerical values for a chr position :

FS=2.433; MQ=45.70; MQRankSum=-3.331 ; ReadPosRankSum=-0.654; SOR=1.489

-- shall I run MUTECT2 with NO -A options, but using VariantAnnotator on the VCF file in order to add FS, MQ, and SOR, the numerical values are :

FS=0.000; MQ=33.87; SOR=2.636;

the question would be of course, why these numerical values differ ? many thanks,

-- bogdan


  • BogdanBogdan Palo Alto, CAMember ✭✭

    Just to add one more little thing,

    shall I run MUTECT2 with -A options for FS, MQ, SOR, the FORMAT field in the VCF file is : GT:AD:AF:BCS

    while if I run MUTECT2 with NO -A options, the FORMAT field of the VCF file is

    how shall I think about these differences ? many thanks again for your kind help !

    Issue · Github
    by Sheila

    Issue Number
    Last Updated
  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    Regarding your first question, this is probably due to shifts in the read alignments that happen as a result of graph assembly as documented for HaplotypeCaller.

    Regarding the second, it's an unintended side effect of the way the annotations are wired up right now -- requesting some annotations disables some of the defaults. This will be addressed in the near future. In the meantime, the recommendation is to leave the annotations alone.

    What you're trying to do diverges significantly from our Best Practices; all the annotations that we want to use for evaluating variants are annotated by default, and any that are not may in some cases be inappropriate to use for variant evaluation. In particular, the filtering annotations recommended for germline analysis are generally not appropriate to use for somatic analysis, and it would be far safer not to try to use them in the same way.

  • BogdanBogdan Palo Alto, CAMember ✭✭

    Dear Geraldine, that is very helpful, many thanks ! Shall the GATK team plan to organize a workshop at Stanford at some moment, we would be happy to help with a part of the organization ;)!

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    You're welcome. We don't organize remote workshops ourselves, we just accept invitations from institutions that host and organize the workshops. We are now taking invitations for 2017, if you know anyone who would be interested in inviting us.

  • BogdanBogdan Palo Alto, CAMember ✭✭

    Thanks, would have to ask a few people, and we can talk about it later ;)

Sign In or Register to comment.