calling somatic mutations on ion proton data

Question

[CHANG]

I am trying to use Mutect to call somatic mutations on tumor/match normal pairs generated by ion proton ampliseq. However, none of the mutations pass mutect filter. I decided to take a lot at IGV to see to see if there's some quality issue and notice a lot of insertions in the read. Attached is a screen shot.

Att

I am guessing all these insertions is making it difficult to call mutations. So I'm thinking the following solutions

1. Re-map with another aligner (i.e. BWA) and re-align with GATK
2. Adjust TMAP settings. I'm currently using Version 4.0-6 with default settings
3. Come up with some filtering steps for processing Ion Var caller output.

Does anyone have suggestions for calling somatic mutations on proton data?

Thank you

Answer 1

[Bontus]

Dear Chang,

As far as I know, Torrent Suite produces two separate variant calls after TMAP, at least that is what I received from our sequencing provider and they have used only Proton software. Thus, you should check whether there is a separate TS option that returns somatic variants prior to variant calling. Otherwise, re-mapping all reads and then running a different tool is of course also an option, however, since you already noticed the homopolymer issue in IGV, this might not make a big difference (TMAP is performing on par if not better than other aligners for Proton), so adjusting Mutect settings may be the better choice here.

Best