Help passing a tuple with paths as input for a process without splitting the files (Invalid method invocation 'call' with arguments: ...)

Ask for help
,
1

Hi! :upside_down_face:

I’m quite new in nextflow, and I already saw some other similar topics here, but none seem to work for me.
I am trying to run a process that takes the following tuple as input:

input:
tuple val(meta), path(samplesheet), path(run_dir)

For this, I defined two parameters, that I pass from the command line as:

-- samplecsv <path to csv file> --bcldata <path to BCL tar.gz or folder>

These parameters are passed from my main workflow to a subworkflow like this:

    bcl_dir = params.bcldata
    sample_sheet = params.samplecsv

workflow MYPIPELINE_MAIN {
    take:
    bcl_dir         // path: BCL input
    sample_sheet    // path: SampleSheet CSV

    main:
    MYPIPELINE_CORE(bcl_dir, sample_sheet)

    emit:
    demux_fastq = MYPIPELINE_CORE.out
}

Inside the subworkflow (MYPIPELINE_CORE ), I pass them into the process using a tuple:

workflow MYPIPELINE_CORE {

    take:
    bcl_dir       // path to BCL tar.gz
    sample_sheet  // path to samplesheet CSV

    main:
    Channel
        .of(tuple([id: 'run1'], file(sample_sheet), file(bcl_dir)))
        .set { ch_combinedTuple }

    ch_combinedTuple.view { "Combined tuple-channel content: $it" }

    BCL2FASTQ(ch_combinedTuple)
}

I verified via println statements that sample_sheet and bcl_dir are java.lang.String and correctly point to valid files.
However, when the process has to be run i come across this error:

Combined tuple-channel content: [[id:run1], /beegfs/home/lrenteria/my_pipeline/data/Samplesheet.csv, /beegfs/home/lrenteria/my_pipeline/data/BCL_data.tar.gz]
ERROR ~ Invalid method invocation `call` with arguments: [<data row from samplesheet>] (java.util.ArrayList) on _closure6 type

Interestingly, when I run a minimal test with only the process and the same tuple construction, everything works fine:

include { BCL2FASTQ } from './modules/nf-core/bcl2fastq'
Channel
    .of( tuple([id: 'run1'], file(params.samplecsv), file(params.bcldata)) )
    .set { ch_test }

workflow {
    ch_test.view { "Test channel content: $it" } 
    BCL2FASTQ(ch_test)
}

Any idea what might be going wrong in the subworkflow context?

Thanks a lot in advance! :grin:

2

First, I think you are using Nextflow DSL1 syntax, not DSL2.
Could you try it in DSL2? the error might be you are trying to use DSL2 syntax in combination of DSL1.

3

The issue isn’t using DSL1 syntax.

    bcl_dir = params.bcldata
    sample_sheet = params.samplecsv

This part isn’t recognised inside the workflow.

Ideally the call to MYPIPELINE_MAIN should be like:

workflow {
    MYPIPELINE_MAIN (
        params.bcldata,
        params.samplecsv
    )
}

It’s also worth noting that params.bcldata and params.samplecsv are String objects that contain a path. The file function you’re using inside the tuple is what coverts the String into a Path object for the process to stage the file correctly.

1 Like
4

@mahesh is right, I tested the following, it works. Thanks.
The MYPIPELINE_MAIN and the main workflow serve the very similar wrapper purposes, probably redundant.

workflow MYPIPELINE_CORE {

    take:
    bcl_dir
    sample_sheet

    main:
    Channel
        .of(tuple([id: 'run1'], file(sample_sheet), file(bcl_dir)))
        .set { ch_combinedTuple }

    ch_combinedTuple.view { "Combined tuple-channel content: $it" }
    
    emit:
    ch_combinedTuple
}

workflow MYPIPELINE_MAIN {
    take:
    bcl_dir
    sample_sheet

    main:
    MYPIPELINE_CORE(bcl_dir, sample_sheet)

    emit:
    demux_fastq = MYPIPELINE_CORE.out
}
workflow {
    bcl_dir = params.bcldata
    sample_sheet = params.samplecsv
    MYPIPELINE_MAIN(bcl_dir, sample_sheet)
}
1 Like