Maybe I'm not thinking about this correctly, but I don't understand the convention of breaking apart related items. Files and their indices are the most obvious example. I'm struggling to think of a situation where I'd want a channel of indices without their source. Every time we separate the index from the source, we're just making more ceremony for ourselves to have to join them back up again every time we need to use the index (because they're useless without the source). It also creates extra ceremony in the workflow outputs. We take a channel of fasta files (or more likely a single fasta file), send it through an index process, ditch the fasta file, keep the index, and then reunite the index to the fasta each time we use the index. For example, is there a reason we couldn't use

ch_indexed_fasta

rather than

ch_fasta

and

ch_fai

? e.g. replace

workflow BAM_MARKDUPLICATES_PICARD {
  take:
    ch_reads
    ch_fasta
    ch_fai
  main:
    PICARD_MARKDUPLICATES ( ch_reads, ch_fasta, ch_fai )
}

process PICARD_MARKDUPLICATES {
  input:
    tuple val(meta), path(reads)
    tuple val(meta2), path(fasta)
    tuple val(meta3), path(fai)
   ... 
}

with

workflow BAM_MARKDUPLICATES_PICARD {
  take:
    ch_reads
    ch_indexed_fasta
  main:
    PICARD_MARKDUPLICATES ( ch_reads, ch_indexed_fasta )
}

process PICARD_MARKDUPLICATES {
  input:
    tuple val(meta), path(reads)
    tuple val(meta2), path(fasta), path(fai)
    ... 
}

Same would go for things like • markdup_bam + markdup_bai • star_bam + star_bai • umi_dedup_bam + umi_dedup_bai • umi_dedup_flagstat + umi_dedup_idxstats + umi_dedup_stats • sorted_bam_flagstat + sorted_bam_idxstats + sorted_bam_stats • markdup_flagstat + markdup_idxstats + markdup_stats

In truth, it's probably a convention that's emerged rather than designed. Skinny tuples are easier to teach to newbies vs fat tuples are probably a better design pattern.

check the

modules.conf

,

workflows/rnavar.nf

and

<http://main.nf|main.nf>

the files captured in the snapshots just denotes path changes and rescuing a samtools stats file that somehow was not published before