The pipeline skips most steps when the sortmerna index is provided

[davidecarlson]

Description of the bug

When the --sortmerna_index flag is provided (e.g., with a directory of prebuilt rRNA indexes) along with the --remove_ribo_rna flag, the pipeline will do the initial trimming and QC steps, create the STAR index, and then immediately skip to the MultiQC step, finishing "successfully" without doing any alignment or quantification.

If the --sortmerna_index flag is not provided, then the pipeline will run normally, producing the full expected results.

Command used and terminal output

nextflow run \
nf-core/rnaseq \
--input ${SAMPLES} \
--outdir ${OUTDIR} \
--remove_ribo_rna \
--fasta ${GENOME} \
--gtf ${GTF} \
--aligner star_salmon \
-r 3.17.0 \
--trimmer fastp \
--sortmerna_index ./idx \
--min_trimmed_reads 1000 \
-profile seawulf

Relevant files

I have put a reproducible example dataset and script that will show this occurring on Google Drive here:

https://drive.google.com/file/d/1SHgCrcgdG4ux2ZLNvtdj88qTFtWSUQW_/view?usp=sharing

System information

Nextflow version: 24.09.2-edge
Hardware: HPC
Executor: slurm
Container engine: Singularity
OS: Rocky Linux 8.7
Version of nf-core/rnaseq: confirmed with 3.17.0 and dev

[davidecarlson]

I should also add this this error is similar (but presumably distinct from) to the one that was addressed with PR #1353.