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.

Setting a specific seed when using GATK

NawarDalilaNawarDalila Member
edited August 12 in Ask the GATK team

Hi,
I just heard that one should set a "fixed seed" when analyzing data using GATK (HapltypeCaller) in order to get always the same results (reproducibility issue). First, is that true? Second, If yes, how can one do that please?
I am using GATK 4.1.0.0 and the commands are executed using WDL.
Best regards
Nawar

Best Answer

  • bhanuGandhambhanuGandham Cambridge MA admin
    Accepted Answer

    HI @NawarDalila

    HaplotypeCaller should automatically set fixed seed. There is a bug in reproducibility in some cases when using very large number of bams, and this is being worked on by the dev team. I am not sure about the timeline on when this would be worked on, but for all intents and purposes HaplotypeCaller should automatically set fixed seed.

Answers

  • bhanuGandhambhanuGandham Cambridge MAMember, Administrator, Broadie, Moderator admin
    edited August 19

    Hi @NawarDalila

    Could you please post a link to the fixed seed suggestion.

    Post edited by bhanuGandham on
  • NawarDalilaNawarDalila Member

    Hi @bhanuGandham
    I don't have a link but I have a similar example from a statistical program we use called STATA were we write before any iteration/imputation this line of code:
    set seed 26000
    The 26000 is just a random number that we use all the time to get reproducible results.
    I hope that example cleared my previous question.
    Best/Nawar

  • bhanuGandhambhanuGandham Cambridge MAMember, Administrator, Broadie, Moderator admin
    Accepted Answer

    HI @NawarDalila

    HaplotypeCaller should automatically set fixed seed. There is a bug in reproducibility in some cases when using very large number of bams, and this is being worked on by the dev team. I am not sure about the timeline on when this would be worked on, but for all intents and purposes HaplotypeCaller should automatically set fixed seed.

Sign In or Register to comment.