I would like to clarify something about the

--postbinning_input

. I get the sense that in most cases, if we trust the bin refinement, we should pass the refined bins or both for downstream analysis. Am I correct in that assumption? If so, shouldn’t the default option be either of those instead of

raw_bins_only

You're correct, but I think the emphasise is on trusting the bin refinement, which I think is still a big area that's not clarified what are best practises, and thus it's an opt-in step - so we can't set psotbinning input to that or

both

We don't do

both

by default either because that can be HUGE computationally, and we don't want users to accidently waste 1000s of currency on wasted computation time

That makes sense, you are leaving the decision to the user. Fair enough. PS. I realized we cannot do a stub-run for the pipeline as some processes (the bowtie2 ones AFAIK) do not follow updated nf-core standards and do not have a stub script

Yes, we are slowly trying to strip out all the old local modules and update the ones we can't remove, but it's taking time to do so while keeping the rest of the pipeline up to date at the same time (we are all volunteers, it's no one's main job). We will get there eventually! Mag is just old and complicated