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!

Which variant caller should we use when ploidy is unknown?

snagsnag BaltimoreMember

Dear GATK team,
We want to know your recommendation as to what variant caller to use if the number of individuals (i.e. ploidy) is unknown, it can be one or many. Our bam files have human amplicons.
Any help would be appreciated. Thanks.


Best Answer


  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    You can use HaplotypeCaller and set the ploidy to whatever number you think would be appropriate. We can't give you any recommendation as to what that value should be since this is an experimental design that is completely unknown to us. Whether you set it high or low, the analysis will be complicated. What are you trying to determine with this experiment?

  • snagsnag BaltimoreMember

    Thanks for your response. We do realize our experiment differs from the traditional uses for GATK, however our goal at this stage is to obtain all variants with respect to the reference. Would setting the ploidy high increase sensitivity? Any other settings/steps that would be worth looking into? We were wondering about the possibility of using Mutect as module that makes no assumptions with respect to ploidy?
    Unfortunately the background information about the experiment that we can post on the public forum at this time is limited. Thanks again for all your help

  • snagsnag BaltimoreMember
  • isantanaisantana UMBMember

    Dear Geraldine, we have information that the amplicon is being used on environmental samples where the goal is to identify how many alleles (and thereby potential humans) have contributed to the sample.
    Any considerations on your part with that additional information?

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    @isantana Unfortunately we don't have any recommendations for this case figure; it's not a use case that we test for. So I really can't give you any guidance on this, sorry. I would recommend trying to find out what others have done in similar cases in the literature. Good luck!

  • isantanaisantana UMBMember

    Thank you for your reply!

Sign In or Register to comment.