Nextflow - Missing value declared as output parameter - output variable defined in script

Question

I've normal-tumor human FASTQ file(s). Currently, I work with only one sample (both normal and tumor fastq R1 and R2), only.
I have alignment gatk mark duplicates, setupnmdtags, applybqsr, recalibrate steps before calling mutect2.

My input files are:
tiny_tumor_R1.fastq.gz
tiny_tumor_R2.fastq.gz
tiny_normal_R1.fastq.gz
tiny_normal_R2.fastq.gz

The expected output from the mutect2 is tiny.mutect2out.vcf that is ${sample_name}".mutect2out.vcf"

I split tiny_normal using nextflow split and take first element of the split array, that is tiny.

However, I get error as:

Caused by: Missing value declared as output parameter: sample_name

I understand that the variable was defined in the script step. How do I retrieve at the output level?

Below is code of the mutect2_process.nf

process mutect2 {

        maxForks 3
        debug true

 memory '15 GB'

        input:
        tuple val(sample_id_tumor),path(bqsrbam_tumor)
        tuple val(sample_id_normal),path(bqsrbam_normal)
        
        output:
        tuple val(sample_name),path(${sample_name}".mutect2out.vcf")
        

        script:
        def avail_mem = task.memory ? task.memory.toGiga() : 0
                def java_options = [
        avail_mem ? "-Xmx${avail_mem}G" : "",
        "-XX:+UseSerialGC",
    ]
        def sample_name=sample_id_normal.split('_normal')[0]

        """
                
        ml gatk/4.1.3.0
gatk --java-options "${java_options.join(' ')}" Mutect2 -L $params.target -I ${bqsrbam_tumor} -I ${bqsrbam_normal} -normal ${sample_name} -tumor ${sample_name} --output ${sample_name}".mutect2out.vcf" -R $params.hg38genome

        """
        }

Answer 1

You would like to use the sample_name variable outside the script block, but the def keyword defines a variable locally. All you'd need to do is remove the def keyword to make it work, but usually what you'd want to do is to define your sample_name variable outside of your process block in your workflow block like in my previous answer. I would also call this variable something else like pair_id, donor_name, case or patient perhaps. I would say you actually have two sample names here: one tumor sample (i.e. tiny_tumor) and one normal sample (i.e. tiny_normal) from the same patient/donor/case.

You may also run into issues if both of your input channels are queue channels. You may not notice any issues if you're testing only a single pair of samples. Try just copying your tiny files a couple of times and giving your workflow a try, for example:

for i in tiny_{tumor,normal}_R{1,2}.fastq.gz ; do
    for j in foo bar baz; do
        cp "${i}" "${j}_${i#*_}"
    done
done

Most of the time what you'll want is one queue channel and one or more value channels when you require multiple input channels.