Picard Base Distribution by Cycle and adapter contamination

Question

[ElCascador]

Hi,

I have trouble interpreting the CollectMultipleMetrics.base_distribution_by_cycle plot from picard for atac-seq data

In my example, there's weird patterns at what looks like the begining of each paired end sequence. Is it a direct reflection of tthe fastqc sequence content across all bases? I am worried about adapter contamination but the picard plot is the same with out without adapter removal.

The picard plot:

The fastqc plot:

Can I just take to heart this post about the fastqc metric and call it a day ?

[GenoMax]

If these are libraries made by nextera (transposon) then they show a similar pattern as the random primed ones (in blog post you linked). You can move forward with the rest of analysis.

Answer 1

[h.mon]

As genomax pointed out, most likely you can move forward with your analysis without further concern. However, if you really want to check, you can use bbmap.sh (from the BBTools/BBMap suite) with mhist=mhist.txt, the mhist.txt file will contain an histogram of matches / mismatches by position of the reads in relation to the reference genome. From the bbmap.sh help:

 mhist=<file>        Histogram of match, sub, del, and ins rates by read location.