We've moved!
This site is now read-only. You can find our new documentation site and support forum for posting questions here.
Be sure to read our welcome blog!

Any filter on the .call_stats.txt output?

helenehelene LAMember


I'm using MuTect 1.1.4 and it works well - except I have one question about the raw output. In my project, I don't always want to filter out the variants (shown in tumor, not in normal) frequently shown in dbSNP. (it gives high value of P(germ line) in the variant classification likelihood formula.)

Does that mean I would have to use the .call_stats.txt directly please? Is there any basic filter criteria applied to this raw output? e.g. Mapping quality, base quality, allele fraction, etc. Although I don't want the algorithm deals with dbSNP and COSMIC, but I do appreciate all the other part of the MuTect design, and would like to keep them. I tried to read the original paper, but couldn't get a complete answer for this question.

Thank you in advance for your help.



  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    MuTect does not blacklist based on dbsnp, it just raises the lod threshold used for calling somatic SNVs. You can adjust the parameter that controls by how much the threshold gets raised if you want.

    Also, check out the newest version of MuTect which calls indels. It is included in GATK 3.5.

  • helenehelene LAMember

    Hi Geraldine,

    Thank you for the reply. I guess my concern is more about the call_stats.txt output. Is this output a simple summary of all sites that differ between tumor and normal, without any LOD or filtration please?

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    It is a list of all sites that were seriously considered as possible somatic variants -- which is not exactly the same thing as a summary of differences between tumor and normal.

Sign In or Register to comment.