Fseq versus MACS2 for ATAC-seq

Question

[a.rex]

I wish to compare two biological conditions with respect to their open chromatin. I use ATAC-seq as my assay. My strategy has thus far been to:

1. call peaks on replicates of each condition
2. find common peaks between replicates
3. merge peaks between conditions
4. count reads in peak intervals for each replicate in each condition using featuresCount
5. Perform DESeq2 differential analysis between the conditions

However, when I use MACS2 I get a lot fewer peaks (-200,000) than if I use FSeq (-20,000,000). I have heard that FSeq is better for considering cutting sites as opposed to the pile-up of overlapping reads strategy that MACS uses. I was wondering, what do people generally favour? What are the benefits of using one over the other?

[ATpoint]

Can you give some details on the species? 200k is a lot. In human, you typically get something in the range of several tens-of-thousands, depending on the q-value cutoff and stringency in downstream filtering. I cannot comment on F-Seq as I never used it, but if peak calling gives you a hard time, maybe consider the R/BioC package CSAW, which uses a window-based approach over the entire genome to call differential regions, avoiding the need to call peaks. Also, what command did you use for ATAC-seq with macs?