Hi,

I would like to use csaw to calculate effective library sizes for my ATAC-seq data so that I can then use these to generate normalized BigWig files. Would you recommend normalizing for composition biases or efficiency biases? I am struggling to decide which is most relevant for ATAC-seq.

Many thanks,

Lucy

Seeing as how no one else has answered, I'd suggest applying the same criterion used for deciding how to normalize ChIP-seq data. That is, do you expect global changes in accessibility between your conditions? If not, then use the efficiency approach; otherwise normalize for composition biases.