After democratically reviewing and deciding on the first point, here is the second one 😛 https://github.com/nf-core/modules/pull/9665#discussion_r2708432728 Should I keep and update one variable (

ch_reads

) throughout the execution, or create and update a new variable, for every downstream input? Pros and cons on the link (don't want to sway opinions 😛 )

I personally would create new variables for each stage... it makes debugging what is in each channel at each time point a bit clearer. I tend to favour an approach like the following (variable naming could be improved):

ch_reads_for_umitools = ch_reads.branch { tuple ->
    umitools: !skip_umitools
    skip: true
}
UMITOOLS_EXTRACT( ch_reads_for_umitools.umitools )

ch_reads_after_umitools = UMITOOLS.out.reads.mix(ch_reads_for_umitools.skip)

I don't like how that bloats the code, and I am not sure how and if garbage collector properly cleans the variables that won't be used any more in Groovy 😞

That's a fair point

Could I have a couple more opinions on this one, and hopefully merge this PR soon? 😅