Allow lists of reads in file?

[mhoban]

Would it be useful to be able to specify your reads in text files? This would be in addition to the current thing where you can do a directory or a glob.

This could be useful if you have a whole ton of reads but only some of them are relevant for your pipeline run, like if you have a folder with

marker1_r1.fastq
marker1_r2.fastq
marker2_r1.fastq
marker2_r2.fastq
marker3_r1.fastq
marker3_r2.fastq

And you just want to do a run with one of those markers, you could do something like:

$ rainbow_bridge.nf --reads-file reads.txt

Or:

$ rainbow_bridge.nf --fwd-file reads_fwd.txt --rev-file reads_rev.txt

Where reads.txt look like:

marker2_r1.fastq
marker2_r2.fastq

And reads_fwd.txt looks like (do s/r1/r2/ for reads_rev.txt):

marker1_r1.fastq
marker3_r1.fastq

This is obviously a simple case where you could just use a glob, but there are others where this could be more useful. Particularly if someone has used the 'combined' multiplexing method (indices AND barcodes)

[mhoban]

Maybe reads should actually be a csv:

sample	read1
sample1	marker2_sample1.fastq
sample2	marker3_sample2.fastq
...	...

or

sample	read1	read2
sample1	marker2_sample1_r1.fastq	marker2_sample1_r2.fastq
sample2	marker3_sample2_r1.fastq	marker3_sample2_r2.fastq
...	...	...

This is how nf-core/radseq does it. Another advantage of this is that you can use arbitrary file names/extensions