Override options don't always take

[s-andrews]

We got caught out with a pipeline recently, for example running the nf_bisulphite_scBSseq pipeline adding options for trim_galore (increasing the trim_R1 length) didn't have any effect.

Looking at the pipeline it seems that because some of the internal pipeline options also set trim_galore options (in this case the --singlecell flag), then I think it then duplicates the options to trim_galore, so you get something like:

trim_galore --trim_R1 12 --trim_R1 6

..and the second option take precedence, even though the first was the one which was manually set. The effect this has is a combination of the way the options are added in nextflow, and how the end program interprets duplicated options.

A fix for this could be to check if an option is set already before changing it, or maybe changing the order in which options are added? In any case this is an issue because the failure here is completely silent, the pipeline takes the arguments and then just does the wrong thing.