How to select or subset process outputs in Nextflow DSL2?

Question

[Eliveri]

I have a DSL2 Nextflow workflow. I would like to use just the outputs named "paired.fq.gz" ( index 0 and 2 in the tuple) in downstream processes.

Is there a way to filter or select a subset of the process outputs?

    process trimreads {

        input:
        tuple val(pair_id), path(reads)

        output:
        set pair_id, path("trimmed_${pair_id}_R{1,2}_paired.fq.gz"),
        path("trimmed_${pair_id}_R{1,2}_unpaired.fq.gz")

        script:
        """
        #!/usr/bin/env bash
        trimmomatic ...
        """
    }

workflow {
   ...
    trimmed_reads_ch = trimreads(read_pairs_ch)

    // selected_trimmed_reads_ch = filter trimmed_reads_ch?

    fastqc_ch = fastqc(...)

[Matthias Zepper]

How about declaring multiple output channels? You can even name them to easily reuse them later:

output:
tuple val(pair_id), path("trimmed_${pair_id}_R{1,2}_paired.fq.gz"), emit: paired
tuple val(pair_id), path("trimmed_${pair_id}_R{1,2}_unpaired.fq.gz"), emit: unpaired, optional: true

Answer 1

[Pierre Lindenbaum]

use the map operator https://www.nextflow.io/docs/latest/operator.html#map

trimmed_reads_ch = trimreads(read_pairs_ch)

pair_id_unmapped_channel = trimmed_reads_ch.map{T->[T[0],T[2]]}

[Eliveri]

Thank you. This worked great!

Answer 2

[Marcel Ribeiro-Dantas]

Check the code below:

process FOO {
  input:
    val id
  output:
    tuple val(id), path("${id}_paired.fq.gz"), emit: paired
    tuple val(id), path("${id}_unpaired.fq.gz"), emit: unpaired, optional: true
  script:
  """
    echo ${id} > ${id}_paired.fq.gz
    echo ${id} > ${id}_unpaired.fq.gz
  """
}

workflow {
  Channel
    .of(1..5)
    .set { ch }
  FOO(ch).paired.view()
}

It outputs:

N E X T F L O W  ~  version 22.10.3
Launching `ex4.nf` [astonishing_einstein] DSL2 - revision: 3f012d70fa
executor >  local (5)
[2f/e395a9] process > FOO (3) [100%] 5 of 5 ✔
[1, /Users/mribeirodantas/work/74/497cebcf0cbe44316c52d877302762/1_paired.fq.gz]
[2, /Users/mribeirodantas/work/2f/127b25707472a7780416bea05b042c/2_paired.fq.gz]
[4, /Users/mribeirodantas/work/94/5106f1547fd087979003be47787c3c/4_paired.fq.gz]
[5, /Users/mribeirodantas/work/aa/886d44ebdd497b6ff43ae5c56fc2a2/5_paired.fq.gz]
[3, /Users/mribeirodantas/work/2f/e395a9283e532c1ef9fc7bac30242a/3_paired.fq.gz]

This is the easiest way to do it. Depending on your case, there are channel operators that can help you better such as filter, branch, take, buffer and so on.