EMIT_ALL_CONFIDENT_SITES for bacteria

ProtaeusProtaeus Posts: 17Member
edited October 2012 in Ask the GATK team

Hello,
I am running the variant caller to identify SNPs and Reference Calls in a bacterial genome, which means I am running with -ploidy 1, -glm POOLSNP and -prnm POOL as suggested in other regions of this forum. The tool does an excellent job when just looking for Variants, but when I attempt to EMIT_ALL_CONFIDENT_SITES, it just spits out the SNPs and not the reference calls. When I remove the arguments stating that it is ploidy of 1, it works fine but calls SNPs that shouldn't be there since it's assuming diploid. Thus, I would really like to be able to emit all sites in ploidy=1 mode. Any reason why this is not possible?
Thanks for you help!
John

Post edited by Geraldine_VdAuwera on

Best Answer

  • delangeldelangel Posts: 71GATK Developer mod
    Answer ✓

    Hi - yes, we also just noticed that emitting confident reference sites is broken when ploidy != 2. This should be fixed in the next couple of days, I'll let you know when this is ready.
    Note that when you update to the newer version (2.1.x) you'll have to change your command line back to -glm SNP, the POOL* arguments have been deprecated.

Answers

  • delangeldelangel Posts: 71GATK Developer mod
    Answer ✓

    Hi - yes, we also just noticed that emitting confident reference sites is broken when ploidy != 2. This should be fixed in the next couple of days, I'll let you know when this is ready.
    Note that when you update to the newer version (2.1.x) you'll have to change your command line back to -glm SNP, the POOL* arguments have been deprecated.

  • ProtaeusProtaeus Posts: 17Member

    Thanks for the info and working to address the issue.
    Cheers,
    John

  • dparkdpark Posts: 9Member

    I noticed the same thing earlier on v2.1-8. I always run the UnifiedGenotyper with the EMIT_ALL_SITES option. But I've found that if I specify ploidy=1, every position where the genotype is same-as-reference is marked with LowQual. In many cases, the QUAL is NaN, in other cases it's just an extremely low number (<10). But the QUALs for SNPs look normal. Interestingly, when I run the same data with ploidy=2, everything looks great (normal/high QUALs for both SNPs and same-as-ref positions).

    I'll keep an eye out for 2.1-9.

    Danny Park, PhD -- Broad Institute, IDI, Sabeti Lab

  • delangeldelangel Posts: 71GATK Developer mod

    Yes, we are aware of these issues when ploidy !=2. We are working on fixing the problem but may or may not be ready for 2.1-9, we will post here when the fix is out.

  • dparkdpark Posts: 9Member

    Has there been any progress on this front? The Sep 6th response on this thread sounded like it would be somewhat straightforward to fix the calculation of QUAL for same-as-ref sites for ploidy != 2.

    Danny Park, PhD -- Broad Institute, IDI, Sabeti Lab

  • delangeldelangel Posts: 71GATK Developer mod

    It turned out that fix wasn't straightforward - it's been fixed but it will make it when we release 2.2, no ETA yet

  • ProtaeusProtaeus Posts: 17Member

    I hate to dwell, but if I am reading the updates to2.2 correctly, this bug has not yet been addressed? Eta?

  • ebanksebanks Broad InstitutePosts: 684Member, Administrator, GATK Developer, Broadie, Moderator, DSDE Member, GP Member admin

    It is in 2.2 as Guillermo said.

    Eric Banks, PhD -- Senior Group Leader, MPG Analysis, Broad Institute of Harvard and MIT

Sign In or Register to comment.