Mutect VCF output

sagrawalsagrawal new zealandMember
edited July 2014 in MuTect v1


Can someone please explain to me what 50 columns represents individually in variant output file of MuTect:

chr1 49530 TCGxTTG C T cancer nromal 0 NOVEL COVERED 0.997898 0.997898 1 1 1 145 1 5.400145 15.759951 4.864089 11.01621 0.095238 0.02 2.047348 73 57 6 2105 215 60 60 0 0 CC 16.992822 0.014085 71 70 1 2549 39 0.894338 0.871416 (28,29,2,4) 31 11.5 35.5 16.5 0 KEEP


Best Answer


  • It looks to me like "context" is also not explained on that page. Is the Context column (TCGxTTG) documented anywhere? I initially thought it was giving the bases before and after the SNV base but for many of my calls, that's not consistent with the ref_allele (or alt_allele) field.

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie

    Hi @Clare,

    The Context column should give you exactly that, the bases before and after your SNP (which is represented by x in the output). Have you checked the reference sequence around the position in question? The allele fields would not be informative to evaluate whether it is correct or not.

  • perryeperrye New Haven, CTMember

    Is it possible to generate an output that lists the reasons a particular mutation failed to pass the filters, so that we can exclude some filters in the analysis? For example, the strand bias filter is not appropriate for my dataset.

    Issue · Github
    by Geraldine_VdAuwera

    Issue Number
    Last Updated
    Closed By
  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie

    Hi @perrye,

    If you upgrade to MuTect 1.1.7 (available on the downloads page), you will get a "failure_reasons" column in the callstats output file.

    In the "failure_reasons" column (second to last before judgement) of the call_stats file, there's an identifier for the relevant filters. "strand_artifact" is the tag for the strand bias filter. As in the Nature Biotech paper: "Candidates are rejected when the strand-specific LOD [t_lod_fstar_forward or t_lod_fstar_reverse columns] is < 2.0 in directions where the sensitivity to have passed that threshold [power_to_detect_positive_strand_artifact and power_to_detect_negative_strand_artifact columns] is ≥ 90%.

    See the paper for more details on other filters or the existing forum synopsis:

Sign In or Register to comment.