I had put our markduplicate bam files for input in edgeR and it generated 28,000 gene ids .But when I gave an input of MACS2 narrowpeaks for the same dataset in DiffBind it only gave us 42 chromosome start end regions.For both I used built in annotation file hg38.Why is there a difference as they both are used in differential analysis of ATAC seq.
We did our analysis using edgeR and we got many gene but when we used Diffbind we got less no of chromosome regions ?
I think that it can solve your question:
Based on our benchmarking results using both simulated data and real data, we recommend using edgeR for the differential analysis of chromatin accessibility when high sensitivity is required or in the condition of limited sample size. Meanwhile, we recommend using DESeq2 for the DAR analysis when the best specificity is required in the condition of large sample size.
Since the DiffBind Galaxy wrapper is built on DESeq2, it can be extrapolated to your results. According with it, edgeR provides a higher sensitivity, which could explain such differences. This is the original paper.