The current GATK version is 3.6-0
Examples: Monday, today, last week, Mar 26, 3/26/04

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Powered by Vanilla. Made with Bootstrap.
Last chance to register for the GATK workshop next week in Basel, Switzerland! http://www.sib.swiss/training/upcoming-training-events/training/gatk-workshop-lecture

Multi-allelic sites being dropped in PhaseByTransmission?

mlindermmlinderm Posts: 29Member

In switching to the 2.x series of GATK, I noticed that PBT now drops multi-allelic sites entirely from the output. Shouldn't the correct behavior be to write them out unmodified? Or is there a specific reason multi-allelic sites are not being written out?

Specifically, here is the current code

if (vc == null || !vc.isBiallelic())
    return metricsCounters;

But I think it should be something like this...

if (vc == null)
    return metricsCounters;
if (!vc.isBiallelic()) {
    vcfWriter.add(vc);
    return metricsCounters;
}

Best Answer

Answers

  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 10,468Administrator, Dev admin

    Sorry to get to your question so late, it got dropped during a shift change.

    I agree that it would make sense to write out multi-allelic sites as unmodified rather than drop them... I'll ask if the author of PBT, @Laurent, can shed some light on this.

    Geraldine Van der Auwera, PhD

  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 10,468Administrator, Dev admin

    Hi Laurent, thanks for answering! Yes, we think it would be preferable to have the multi-allelic sites included in the output by default. Although it might make sense to give the option to omit them from the output using a flag in the command... but that's up to you.

    Good to hear you're working on supporting multi-allelic sites. Good luck!

    Geraldine Van der Auwera, PhD

  • mlindermmlinderm Posts: 29Member

    Thanks for the responses. The change is so small, it is probably not worth submitting a patch...

Sign In or Register to comment.