The frontline support team will be slow on the forum because we are occupied with the GATK Workshop on March 21st and 22nd 2019. We will be back and more available to answer questions on the forum on March 25th 2019.

Infuence of Coverage Depth Difference in Tumor & Normal Sequencing Data in Somatic Mutation Calling

Hi, GATK team.

I'm planing to use Mutect & Mutect2 to call somatic mutations in Tumor and mathched Normal samples. I have a question about the coverage depth difference of Tumor and Normal samples. Since our team want to sequencing Tumor sample in 200X ~ 400X WES, when Normal sample in about 30X ~ 50X WGS. Tumor samples are about 5 times to 10 times deeper than Normal samples. Is that okay for the Statistical Model in Mutect/Mutect2?

I have tested 150X tumor with 120X normal, and the same sample 150X tumor with 26X normal, only ~75% sites in concordence. In addition, more low variant-allele-frequence sites are detected in the 150X tumor with 26X normal samples. This is out of our expectation (we thought low coverage depth of normal may miss part of the low variant-allele-frequence sites).

Hopefully, your team could explain about this (in the Statistical Modelling), and give us some advise in sequencing depth design.

Thank you. Looking forward to your reply.


  • AdelaideRAdelaideR Unconfirmed, Member, Broadie, Moderator admin

    Hi @jemimalwh, I believe the information on this page can help resolve some of these questions, if not all of them.

    Mutect2 differs from Mutect in that it uses a panel of normals (PoN) to filter out sites with multiple variant alleles and systematic artifacts of sequencing. A PoN is constructed with germline normal samples and variants are called with the same sensitivity used in somatic calling. The normal samples are then gathered into a cohort and sites which are called in two or more samples are retained. This also filters out common germline variant sites.

    Please feel free to post a follow up question after looking over the documentation.

Sign In or Register to comment.