Indel realignment for a normal-tumor pair of BAM files

dgacquerdgacquer BelgiumMember

Dear GATK team,

I have noticed that sometimes, when I run IndelRealigner independently on a tumor BAM file and its matching normal BAM file, I get different haplotypes, leading to false somatic variant or indel calls. Is there a way to ensure that both BAM files are realigned the same way ? For instance, merging both BAM files together, running IndelRealigner and splitting resulting BAM file by read group ?

Best regards

David

Tagged:

Best Answer

Answers

  • dgacquerdgacquer BelgiumMember

    Thanks Geraldine,

    I did not known about the -nWayOut argument, this is exactly what I was looking for.

    Best regards

    David

  • artitandonartitandon Member

    Is this the standard way to run Indel realignment for the tumor/normal samples, i,e. run it for both the normal and tumor sample together, and then use the -nWayOut option?

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin
  • IrantzuIrantzu Member
    edited December 2016

    Hi there!
    Maybe this has been asked before but if I have paired normal/tumor datasets, should I run both RealignerTargetCreator and IndelRealigner commands using both bam (N/T) files at time (for the both steps)?
    Thank you in advance,

  • shleeshlee CambridgeMember, Administrator, Broadie, Moderator admin

    Hi @Irantzu,

    If you are using a mutation caller that reassembles haplotypes, e.g. MuTect2, then there is no longer a need to perform indel realignment. If you are using a pile-up caller, e.g. MuTect1, then yes, please run RealignerTargetCreator and IndelRealigner on both your matched T/N BAMs concurrently. You can pass in multiple alignment BAMs each with the -I parameter. Generate one output file for each input by specifying -nWayOut instead of -o for IndelRealigner. For detailed documentation, see Tutorial#7156.

Sign In or Register to comment.