Filtering overlapping calls in Pindel

Question

[volvox]

I'm running Pindel on some tumor/normal pairs. I noticed that the output contains some overlapping calls as shown below. Start positions of the deletions are the same, but the ends shift by 1bp.

217    D    4    NT    0        ChrID    chr1    BP    7828173    7828178    BP_range    7828173    7828192
218    D    5    NT    0        ChrID    chr1    BP    7828173    7828179    BP_range    7828173    7828192
219    D    6    NT    0        ChrID    chr1    BP    7828173    7828180    BP_range    7828173    7828192
220    D    7    NT    0        ChrID    chr1    BP    7828173    7828181    BP_range    7828173    7828192

How would you pick one? Is it reasonable to use the SUM_MS per supported read?

Thanks.

Answer 1

[liangkaiye]

this is a homopolymer region so that we would probably see all possible alleles if you sequence ultra deep. If this sample is MSI, have you tried our recent MSIsensor? This tool approaches the variant detection differently.