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!

the threshold of variant classification in MuTect

jingmengjingmeng AustraliaMember

When MuTect performs variant classification in normal data, it distinguished two cases. I understand how to get the prior probability of a non-dbSNP site being a gremline (0.00005) and the prior probability of a dbsnp site being a gremline (0.095). But I do not know how to get the threshold of 2.2 and 5.5 respectively. Can you please show me what is the probability of a site not being a germline variant?

Another question is similar. When performing variant detection in tumor data, what is the probability of a site not being a variant? Thanks for your help!

Issue · Github
by Sheila

Issue Number
Last Updated
Closed By


  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    @jingmeng The LOD thresholds that MuTect uses have been determined empirically as being values that work well.

    We currently don't provide more detailed support than that, but you can try reading the MuTect paper (Cibulskis et al) for more details on the underlying model.

Sign In or Register to comment.