Asymmetric variant score quality distribution

[Pinko]

I've called WGS Drosophila variants using GATK UnifiedGenotyper.

My problem is that the distribution of the quality scores is very asymmetrical, effectively with one very high peak of comparatively low quality and one secondary peak of greater quality.

I don't strictly know if this a cause for concern but it really doesn't look right to me.

My read mapping pipeline includes removal of adaptors and low-quality bases, and local re-alignment. The variant call quality scores are from four individuals with an average of 30X coverage.

Any advice on whether I should be concerned and what I might be able to do to correct this would be much appreiciated.

[aeonsim]

I believe that's normal, it's similar to what we usually see when sequencing. I've never pull investigated the cause of it but I suspect it may be due to score differences between Heterozygous and Homozygous variants with possible some impact with read depth & structural duplications.

Remember QUAL score is simply an indicator of the probability that there is a Variant at that position (not if the variant is correct) and thus it is affected by Read Depth and a Homozygous variant will usually have 2x the number of Variant support reads at the same coverage as a Heterozygous Variant.