I am not sure why HaplotypeCaller does not call my SNV mutation?

zhengzha2000zhengzha2000 Posts: 4Member

Hi Guys,
I am using HaplotypeCaller to call mutations for some of patient samples. I know that at MSH2 intron5 near splicing site there is a point mutation in one of the sample. however, this is also a region in Mills-1000g known indel file. Not matter how I try, Haplotypecaller can NOT call the SNV but always call the indel. the bam file is realigned and recalibrated using mills-1000g indel gold standard file.
the position is chr2 47641560
the snapshot is attached.
I am using 3.0 and gvcf mode on one single sample.
notice there are 158X coverage at this position, 72 show allele base T, 72 are deletion. 84 show reference base A.
mutect can easily pick it up though the reject reason is nearby-gap-event.
Any input is highly appreciated. Thanks in advance.

x.png
1106 x 838 - 81K
Tagged:

Answers

  • zhengzha2000zhengzha2000 Posts: 4Member

    btw, the lower panel in the igv snapshot shows the bam files by -outputbam. clearly haplotypecaller doesn't use the allele reads for local assembly.
    also the position is a well-known position for colon cancer risk predisposition.

  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 8,142Administrator, GATK Dev admin

    Hmm, what do the qualities look like? Can you post the gVCF record you get?

    Geraldine Van der Auwera, PhD

  • zhengzha2000zhengzha2000 Posts: 4Member

    Hi Gerald,
    thanks for your quick reply.
    I attached the gvcf output for your reference.

    txt
    txt
    a.txt
    11K
  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 8,142Administrator, GATK Dev admin

    Thanks. This looks pretty messy and I don't know what is the best way to deal with it, but I'll ask the team for some help. Might need to wait until after the weekend for an answer though.

    Geraldine Van der Auwera, PhD

  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 8,142Administrator, GATK Dev admin

    Hi there, we're not sure what's going on so we need a snippet of your bam file to debug this locally. Can you please upload a bug report to our FTP server? Instructions are here: http://www.broadinstitute.org/gatk/guide/article?id=1894

    Geraldine Van der Auwera, PhD

  • zhengzha2000zhengzha2000 Posts: 4Member

    Hello Geraldine:
    I uploaded a file names msh2-intron.tgz onto gsa ftp. see detail below.
    Thanks for your effort on this!
    Zheng

    LPTH0112:~ zhengzha$ ftp ftp.broadinstitute.org
    Connected to ftp.broadinstitute.org.
    220 ProFTPD 1.3.3g Server (Broad Institute of MIT and Harvard) [69.173.80.251]
    Name (ftp.broadinstitute.org:zhengzha): gsapubftp
    331 Password required for gsapubftp
    Password:
    230 Anonymous access granted, restrictions apply
    Remote system type is UNIX.
    Using binary mode to transfer files.
    ftp> put msh2-intron.tgz
    local: msh2-intron.tgz remote: msh2-intron.tgz
    227 Entering Passive Mode (140,163,99,30,68,216)
    150 File status okay; about to open data connection.
    100% |**************************************************************************************************************************************************************| 406 KiB 38.77 MiB/s 00:00 ETA
    226 Transfer complete, closing data connection.
    416015 bytes sent in 00:01 (335.73 KiB/s)
    ftp>

  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 8,142Administrator, GATK Dev admin

    Thanks for the test data. I've put this in the bug tracker; I can't guarantee we'll be able to look at this very soon as we are very busy right now, but I'll let you know in this thread when we know more.

    Geraldine Van der Auwera, PhD

  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 8,142Administrator, GATK Dev admin

    Hi @zhengzha2000,

    Sorry for the very late reply, your question dropped to the bottom of our priority queue by accident. I've had the devs look at your data; they don't think the SNP looks real, because there's a lot of PCR error in the data and none of the reads with the single mismatch actually reads through the repeat structure. To get GATK to call a SNP here, if it is actually real, you'd need to have cleaner data.

    Geraldine Van der Auwera, PhD

Sign In or Register to comment.