SEQTK_SAMPLE fails on `.fq.gz` inputs — output glob `*.fastq.gz` misses the preserved input extension

[adamjtaylor]

---

Description of the bug

When --sample_size is set to a non-zero value, SEQTK_SAMPLE fails for any input whose filename ends in .fq.gz (rather than .fastq.gz):

MissingFileException: Missing output file(s) `*.fastq.gz` expected by process
`NFCORE_SEQINSPECTOR:SEQINSPECTOR:SEQTK_SAMPLE (<sample>)`

The seqtk command itself succeeds (exit 0), but Nextflow cannot collect the output because the produced file ends in .fq.gz, while the process declares its output as *.fastq.gz.

This is an internal inconsistency: the samplesheet schema (assets/schema_input.json) explicitly accepts .fq.gz for fastq_1/fastq_2, but SEQTK_SAMPLE can only capture .fastq.gz. .fq/.fq.gz is a standard FASTQ extension and is common in public archives, so this breaks otherwise-valid inputs whenever subsampling is enabled.

Root cause — modules/nf-core/seqtk/sample/main.nf:

output:
tuple val(meta), path("*.fastq.gz"), emit: reads          // only *.fastq.gz

script:
"""
printf "%s\\n" $reads | while read f;
do
    seqtk sample $args \$f $sample_size \\
        | gzip --no-name > ${prefix}_\$(basename \$f)      // preserves input ext -> .fq.gz stays .fq.gz
done
"""

${prefix}_$(basename $f) keeps the input extension, so G073_1.fq.gz → <prefix>_G073_1.fq.gz, which *.fastq.gz never matches.

Suggested fix — broaden the output pattern:

tuple val(meta), path("*.{fastq,fq}.gz"), emit: reads

or normalize the emitted name to a canonical .fastq.gz in the script (also gives consistent downstream filenames). (Note: seqtk/sample is a shared nf-core/modules module, so the fix likely belongs there.)

Runs with the default --sample_size 0 succeed, since SEQTK_SAMPLE is not invoked.

---

Command used and terminal output

$ nextflow run nf-core/seqinspector -r 1.0.1 \
    -profile <...> \
    --input samplesheet.csv \        # fastq_1/fastq_2 point to *.fq.gz files
    --outdir results \
    --sample_size 2000000

ERROR ~ Error executing process > 'NFCORE_SEQINSPECTOR:SEQINSPECTOR:SEQTK_SAMPLE (<sample>)'
Caused by:
  Missing output file(s) `*.fastq.gz` expected by process
  `NFCORE_SEQINSPECTOR:SEQINSPECTOR:SEQTK_SAMPLE (<sample>)`

Command executed:
  printf "%s\n" G073_1.fq.gz G073_2.fq.gz | while read f;
  do
      seqtk sample -s100 $f 2000000 | gzip --no-name > <prefix>_$(basename $f)
  done

Command exit status: 0

---

Relevant files

.nextflow.log available on request. Minimal repro: any samplesheet whose fastq_1/fastq_2 end in .fq.gz, run with --sample_size > 0.

---

System information

• Nextflow version: 25.10.2
• Hardware: Cloud
• Executor: awsbatch
• Container engine: Docker
• OS: Amazon Linux 2 (AWS Batch)
• Version of nf-core/seqinspector: 1.0.1