has these MuTect2 Pitfalls been solved?

Tagged:

Best Answer

  • davidbendavidben Boston ✭✭✭
    Accepted Answer

    @manba

    The allele frequency in the tumor sample is 58% (1528/2634). However, the variant is filtered out by MuTect2, because it fails both filters above. The allele frequency in the normal sample is 1.1% (21/1920).

    Fixed that.

    In some cases, we therefore end up calling variants with higher allele frequencies in the normal than in the tumor as somatic.

    Fixed that in GATK 4 beta.

    In fact, MuTect2 not only filters out V600E but also a variant in the neighboring codon. The explanation for both variants being filtered is indicated as clustered_events and multi_event_alt_allele_in_normal in the VCF file.

    Fixed that. GATK 4 Mutect2 has no problem with two nearby, phased SNVs.

    In general, anything you may read about Mutect2 before 2017 is obsolete. There's hardly any code remaining from that time. It's a completely different tool.

Answers

  • davidbendavidben BostonMember, Broadie, Dev ✭✭✭
    Accepted Answer

    @manba

    The allele frequency in the tumor sample is 58% (1528/2634). However, the variant is filtered out by MuTect2, because it fails both filters above. The allele frequency in the normal sample is 1.1% (21/1920).

    Fixed that.

    In some cases, we therefore end up calling variants with higher allele frequencies in the normal than in the tumor as somatic.

    Fixed that in GATK 4 beta.

    In fact, MuTect2 not only filters out V600E but also a variant in the neighboring codon. The explanation for both variants being filtered is indicated as clustered_events and multi_event_alt_allele_in_normal in the VCF file.

    Fixed that. GATK 4 Mutect2 has no problem with two nearby, phased SNVs.

    In general, anything you may read about Mutect2 before 2017 is obsolete. There's hardly any code remaining from that time. It's a completely different tool.

Sign In or Register to comment.