I'm working with data from a targeted sequencing panel (HaloPlex) to generate germ-line variant calls. The library for sample A17789 was in a batch that initially failed and tech support assisted our sequencing core in retrieving the samples via another round of PCR (not clear on exact details). I'm trying to make sure these samples are OK to use.There is not an abundance of variants called for this sample and other metrics such as coverage, average depth etc. look on par with the rest of the samples.
I've ran Pierre's BamCmpCoverage to get this. A17789 is in the middle, and if I understand correctly the graph is one samples depth at a particular position compared to another samples depth at the same position. I interpret the large spread for this sample to mean the depth at each position is different than most other samples; some regions having more depth, some less.
I've also used Stephen Turner's blog post about visualizing coverage to generate this. It's hard to tell with the colors, but A17789 is the line that starts on top and has a steeper curve.
So, what other types of QC would you do to make sure this sample is OK to use? Or do I simply make note of these differences and trust variant calls for this sample?
Thanks, I'll look more closely at this. From a quick look at the FASTQC results it looks like the duplication results are inline with the other samples, but I'll look closer.