Our documentation websites are currently offline due to a data center fire. We do not yet have an ETA for restoring service; we’ll update this message when we know more.

Run GenotypeGVCFs by populations separately or all populations together ?

Guillaume_DGuillaume_D Lausanne, SwitzerlandMember

Hi GATK team,

I have a question about joint genotyping which I think has not been addressed yet in this forum. I want to do population genomics on a non model species. Once I have my gVCFs files for each individual of my five populations, should I run GenotypeGVCFs in all the individuals of each population separately or in all the individuals of all populations together? And if the 2nd solution is the correct one, is there a way to specify the population so that GenotypeGVCFs takes it into account?

Thanks in advance
Guillaume

Best Answer

Answers

  • SheilaSheila Broad InstituteMember, Broadie, Moderator

    @Guillaume_D
    Hi Guillaume,

    It depends on what exactly your end goal is. Can you tell us more about what you are trying to do? Also, have a look at this article.

    -Sheila

  • Guillaume_DGuillaume_D Lausanne, SwitzerlandMember

    Hi Sheila,

    Thank you for your answer.
    The goal, at least in a first time, is to check populations structure and genomic differentiation between populations.

    Guillaume

  • Guillaume_DGuillaume_D Lausanne, SwitzerlandMember

    Hi Sheila,

    Ok cool, I just wanted to be sure, thanks a lot !

    Guillaume

  • liuqiliuqi 104 Building, 1# Courtyard, Beichen West Road, Chaoyang District, Beijing, ChinaMember

    @Sheila
    Hello Sheila,
    you say we should run GenotypeGVCFs on all of the populations together. after this step, we should filter variants by variant (Quality Score) recalibration. can we filter variants together? or split raw vcf into separate populations, and then update annotation and filter variants?

    I see an argument 'InbreedingCoeff' in tutorials which we should add. if we combine all of the populations together, and use VQSR, Do this argument maybe have influence?

    Thanks in advance

    Qi Liu

  • SheilaSheila Broad InstituteMember, Broadie, Moderator

    @liuqi
    Hi Qi,

    You can run VQSR on all variants/populations together. After filtering, you can separate the populations and compare them.

    For inbreeding coefficient, have a look at this thread.

    -Sheila

Sign In or Register to comment.