Mutect2 filter somatic mutations as multiallelic and clustered_events

Follow

Nesrine

• March 18, 2020 13:32

Hi,

I'm using mutect2 with a panel of normal (gatk/4.1.2.0) . However, I have an issue with filtermutectcalls step. It filters real somatic mutations like R132C for IDH1 and R172G for IDH2 as clustered_events and mutliallelic sites. I didn't change any parameters of mutect2 or filtermutectcalls. Also there is no variants of these two positions in the panel of normal that i have constructed. what can I do so that these two mutations and maybe, other real somatic mutations will not be filtered as clustered_events or multiallelic sites

Thank you

 

0

• 

  David Benjamin

  • March 18, 2020 18:17

  You can adjust the -max-events-in-region and -max-alt-allele-count arguments.  However, alleles filtered by these are overwhelmingly likely to be false positives.  I'm not disagreeing with you when it comes to known hotspots, but I am saying that relaxing these filters will greatly elevate the false positive rate across the genome.  

  0

  Comment actions Permalink
• 

  Nesrine

  • March 18, 2020 20:02
  • Edited

  -max-alt-allele-count is already fixed to 1. I changed -max-events-in-region which was fixed to 2  and unfortunately, I've had the same results...

  0

  Comment actions Permalink

Please sign in to leave a comment.