fetchngs 1.12 produces empty runinfo.tsv because ENA no longer accepts parent_study

Description of the bug

fetchngs 1.8 produces empty runinfo.tsv because ENA no longer accepts parent_study

I encountered a failure in nf-core/fetchngs 1.12 while processing valid SRA run accessions such as:

SRR7299460
SRR7299461

The pipeline fails at:

NFCORE_FETCHNGS:SRA:SRA_RUNINFO_TO_FTP

with:

Missing output file(s) `*.tsv` expected by process
`NFCORE_FETCHNGS:SRA:SRA_RUNINFO_TO_FTP`

The preceding *.runinfo.tsv files are created, but contain only the header and no data rows, for example:

run_accession experiment_accession sample_accession ... fastq_ftp ...

The accessions themselves are valid. NCBI SRA returns complete metadata for them, and ENA also resolves both the run and experiment accessions correctly.

For example:

curl -L -G \
  'https://www.ebi.ac.uk/ena/portal/api/filereport' \
  --data-urlencode 'accession=SRX4201949' \
  --data-urlencode 'result=read_run' \
  --data-urlencode 'fields=run_accession,experiment_accession,fastq_ftp,fastq_md5,fastq_bytes'

returns the expected record for SRR7299460, including both paired FASTQ URLs.

However, reproducing the full ENA field request used by fetchngs 1.12 results in:

Invalid fieldName(s) supplied: parent_study

The field list in bin/sra_ids_to_runinfo.py includes:

"parent_study"

which is apparently no longer accepted by the current ENA read_run filereport API.

Because the script creates the output file and writes its header before attempting to write returned metadata records, the failed/empty ENA response results in a header-only *.runinfo.tsv.

The subsequent sra_runinfo_to_ftp.py process then iterates over zero records and does not create the expected *.runinfo_ftp.tsv, resulting in the Nextflow error:

Missing output file(s) `*.tsv`

I confirmed that removing parent_study from the requested ENA metadata fields resolves the issue.

For example, this ENA request succeeds:

curl -L -G \
  'https://www.ebi.ac.uk/ena/portal/api/filereport' \
  --data-urlencode 'accession=SRX4201949' \
  --data-urlencode 'result=read_run' \
  --data-urlencode 'fields=run_accession,experiment_accession,sample_accession,secondary_sample_accession,study_accession,secondary_study_accession,submission_accession,run_alias,experiment_alias,sample_alias,study_alias,library_layout,library_selection,library_source,library_strategy,library_name,instrument_model,instrument_platform,base_count,read_count,tax_id,scientific_name,sample_title,experiment_title,study_title,sample_description,fastq_md5,fastq_bytes,fastq_ftp,fastq_galaxy,fastq_aspera'

As a workaround, running fetchngs with an explicit --ena_metadata_fields list that excludes parent_study allows the metadata query to succeed.

It may therefore be useful to:

1. Remove or replace parent_study from the default ENA metadata field list.
2. Add validation that the ENA response contains at least one data record rather than silently creating a header-only TSV.
3. Report the ENA API error response directly so that the upstream cause is visible instead of only failing later in SRA_RUNINFO_TO_FTP.

Environment:

nf-core/fetchngs: 1.12
Executor: Slurm
System: Pawsey Setonix
Profile: pawsey_setonix / Singularity

Command used and terminal output

nextflow run nf-core/fetchngs \
    -r 1.12.0 \
    -resume \
    --input sra_ids.csv \
    --download_method sratools \
    --outdir "$MYSCRATCH/fetchngs_results" \
    -profile singularity,pawsey_setonix

Relevant files

No response

System information

No response

Start with bin/sra_ids_to_runinfo.py and inspect the default ENA metadata fields, especially parent_study, then trace how its output reaches sra_runinfo_to_ftp.py in SRA_RUNINFO_TO_FTP. Reproduce the ENA request using the example accessions and verify that the corrected workflow produces records and the expected FTP TSV instead of a header-only file.