Merging peak files for unionized set

Question

[aam458]

Hello,

I'm relatively new to atac and pretty much all of bioinformatics work but I've been picking up a lot of experience as I've gone and figured most things out on my own.

I'm beginning some work on some bed files I've (finally) attained for atac-seq analysis, I've used bedtools intersect to check how well the replicates overlap for downstream analysis (decently well at ~74%) and now I want to merge them (each replicate into a union set) so I can compare different timepoints.

Does anyone have any thoughts/experiences/procedures as to how to correctly merge two replicate peak bed files and maybe some hints as to common pitfalls to avoid? It occurs to me that it would be very hard to tell on first glance whether one has made a large mistake.

I was thinking about using bedtools merge function to do it without specifying any options (because none of them seem to apply, looking at the bedtools documentation).

Thanks!

Answer 1

[trausch]

I am not sure if that's necessarily the best solution but we indeed use bedtools merge to get a unified peak list, then count fragments for each sample at these peaks and then run differential peak calling.Here is the link to our ATAC-Seq workflow:

https://github.com/tobiasrausch/ATACseq