RNAseq-3.14.0 Pipeline fails with aligner star_rsem

[GerardTromp]

Description of the bug

Pipeline fails when using aligner star_rsem because of missing bowtie in container singularity-mulled-v2-cf0123ef83b3c38c13e3b0696a3f285d3f20f15b-64aad4a4e144878400649e71f42105311be7ed87-0.img.

rsem-calculate-expression relies on bowtie or bowtie2. As currently configured ...:QUANTIFY_RSEM:RSEM_CALCULATEEXPRESSION is invoked before STAR_ALIGN or rather STAR_ALIGN is not in the execution plan. Thus, the step relies on executing RSEM with its default aligner, which is not present in any of the 34 containers.

The task main.nf (modules/nf-core/rsem/calculateexpression/main.nf) also specifies clearly that the container is one that should contain rsem and star. (conda "bioconda::rsem=1.3.3 bioconda::star=2.7.10a")

The pipeline aligns the data and proceeds to multiqc without error when specifying the aligner as star_salmon

Command used and terminal output

command: `nextflow run main.nf -resume -c params.config`

A key parameter and value defined in the params.config is `aligner = star_rsem`

Relevant files

nfcore-star_rsem.zip
ZIP Manifest

File name	Description
nextflow.log	renamed to remove leading period
star_rsem-params.config	originally params.config
terminal_output.log	Terminal output
work_SRR7448503.log	File from work directory, application log file
work_command.begin	Work directory timestamp
work_command.err	NF process err file
work_command.log	NF process log
work_command.out	NF process out
work_command.run	NF execution specification
work_command.sh	Command executed by .run
work_command.trace	NF trace file

System information

Nextflow version: 24.04.4
Hardware: PowerEdge R650, 2 X Xeon(R) Gold 6354 CPU @ 3.00GHz, 1.5 TB RAM
Executor: local
Engine: Singularity v 3.8.7
nf-core/rnaseq v 3.14.0

[abhi18av]

Hi rnaseq team 👋

I'm adding further context here as @GerardTromp and I had a brief discussion about this yesterday (in-person) about why it hasn't been caught with test dataset and indeed when I tried to reproduce this using test profile it doesn't fail with this error.

Therefore the test profile doesn't seem to be triggering this particular pathway 🤔

$ nextflow run nf-core/rnaseq -r 3.14.0 -profile test,singularity -params-file params.rnaseq.yaml -resum

I've used the following parameters

outdir : 'test-profile-output'
aligner : 'star_rsem'
skip_biotype_qc : true
skip_rseqc : true
featurecounts_feature_type : 'transcript'
featurecounts_group_type : 'product'
gencode : false
igenomes_ignore : true

[pinin4fjords]

I think there may be some confusion in the original post here. There is no dependency between the software environments of different steps. RSEM uses STAR internally, completely independent of the STAR processes in this workflow.

I'm also confident that the RSEM process works effectively on singularity, since that is tested as part of the modules infrastructure in nf-core.

@abhi18av are you saying you did, or did not replicate the error with the command you ran?

[abhi18av]

@pinin4fjords , in my test run, no I didn't reproduce the original problem.

But maybe I missed out on some parameter used by @GerardTromp 🤔

[pinin4fjords]

@GerardTromp if you can give me a reproducible example I can look into it. Otherwise I suspect some misconfiguration on your side I will find hard to determine.