Hi, I noticed that the Qualimap reports calculate coverage including duplicate reads. Was this a deliberate choice, or is there a reason why Picard/MarkDuplicates-flagged reads aren’t excluded by default?

Off the top of my head, so take with a pinch of salt as it's been a while, I think that bismark deduplicates after the alignment step.

I forgot to mention that I was running taps and bwa-meth branches of the workflow. I guess, it shouldn’t affect bismark branch due to deduplication step but in other branches qualimap doesn’t exclude picard’s markduplicates. I checked yesterday by rerunning qualimap manually by adding

--skip-duplicated --skip-dup-mode 0

in addition to the default parameters and compared with default results. By default qualimap picks up duplicates but doesn’t remove them from the coverage calculations, while by adding options I mentioned above drops the coverage. I could create a PR and add these options :)

Qualimap is only a reporting / QC step, not preprocessing as I remember? In both branches I think.

Yes, qualimap doesn’t do any preprocessing, just qc reporting. So, I’m only concerned about qualimap’s reported coverage as it doesn’t include markduplicates, otherwise this doesn’t affect any downstream tasks.