Normalization for bulk ATAC-seq

[ma-ta-ta]

I have bulk ATAC-seq from two cell populations (two replicates each). Papers suggest for bulk counts to apply quantile normalization followed by GC bias normalization using CQN.

Does chromVAR account for varying read depth between samples or should counts be normalized before passing them to chromVAR? If a normalization step before running chromVAR is appropriate, what kind of normalization would you suggest? Should both, quantile normalization and GC bias normalization, be applied or only quantile normalization or would it make sense to downsample all BAM files to have equal read depth?

Thanks a lot for your help!

[AliciaSchep]

chromVAR does not require read depth or GC bias normalization as it will do both of those