I am working with some atac files from cellranger-arc v.2.0. I was wondering whether the atac_possorted_bam.bam produced as the output was deduplicated? I believe the fragment files that are generated detect duplicate reads (as represented by reads with the fifth column >= 2), but I was not sure whether the pipeline deduplicates reads within the output bamfile similar to how samtools dedup would.
If these bam files have not been deduplicated, would it be prudent to do so if I wanted to generate sample/count matrix based on these bam files using something like featureCounts?