Salmon and DNA contaminated RNAseq

[Giovanni Bacci]

Hi all,

I was planning to use Salmon for analyzing a set of RNAseq samples coming from four different bacterial strains of the same species. Unfortunately, samples seem to be contaminated with DNA, indeed, if I map the reads on the whole genome (and plasmids) I end up with an almost uniform coverage both inside CDS and outside. Attached to the post you can find a plot showing the coverage of CDS (in blue) and of regions outside the CDS (in other words the complement of the CDS, negative values in red). Each row of the plot is a different strains whether each column is a different chromosome. I was thinking about subtracting the median of the coverage outside CDS from the coverage inside CDS but I don't know if this approach can be feasible with Salmon. Does Salmon take into account possible DNA contamination in its pipeline? I would like to use it for my experiment but I don't know if it can be used in this kind of situation.

Thanks in advance,

Giovanni