Polymerase slippage filtering

Question

[marina-orlova]

Hi! Is there any tools to filter out false positive variants due to polymerase slippage? (AAAAAAinsA for example) Or any callers which perform such filtering. I have total-rna, and abnormally high percent of mutations (about of 30%) are indels in homopolymers. Maybe a bad polymerase..I don't know.

[Kevin Blighe]

Which sequencing platform have you used?

[marina-orlova]

the platform was illumina

[Kevin Blighe]

The mapping quality (MAPQ) of the reads containing these reads should be low, perhaps less than Phred-scaled 30. You could, therefore, indirectly filter them out by filtering out reads below a certain MAPQ. From my experience, filtering based on MAPQ 50 or 60 eliminates such variants.

Even Illumina's sequencing chemistry has problems with the accurate sequencing of homopolymer regions.

[finswimmer]

While filtering by this high MAPQ would indeed filter out this type of false-positive indels, depending on your target region, you will also filter out a lot of true-positives. The usual problem: When you increase the specificity you will decrease the sensitivity.

At least bcftools tries to find out, whether there is a homopolymer error. From the manual for bcftools mpileup

-h, --tandem-qual INT
Coefficient for modeling homopolymer errors. Given an l-long homopolymer run, the sequencing error of an indel of size s is modeled as INT*s/l [100]

What variant caller have you used, marina-orlova ?

fin swimmer