Error executing nf-core/metaboigniter pipeline

Question

[eesha28112001]

I ran this command: export NXF_VER=22.10.8; nextflow run nf-core/metaboigniter -profile test

The error obtained:

Error executing process > 'get_software_versions'

Caused by:
  Process `get_software_versions` terminated with an error exit status (127)

Command executed:

  echo 1.0.1 > v_pipeline.txt
  echo 22.10.8 > v_nextflow.txt
  Rscript -e "cat(as.character(packageVersion('CAMERA')),'\n')" &> v_camera.txt
  Rscript -e "cat(as.character(packageVersion('xcms')),'\n')" &> v_xcms.txt
  Rscript -e "cat(as.character(packageVersion('MSnbase')),'\n')" &> v_msnbase.txt
  Rscript -e "cat(as.character(packageVersion('IPO')),'\n')" &> v_ipo.txt
  OpenMSInfo |  grep -oP -m 1 '([0-9][.][0-9][.][0-9])' &> v_openms.txt
  sh /usr/bin/CSI/bin/sirius.sh --loglevel=OFF --version 2>1 | grep -oP -m 1 '([0-9][.][0-9][.][0-9])' &> v_sirius.txt

  scrape_software_versions.py &> software_versions_mqc.yaml

Command exit status:
  127

Command output:
  (empty)

Work dir:
  /home/eesha/metaboigniter/work/6f/76127fb7388723e1a4c8b08c4e543d

Tip: you can replicate the issue by changing to the process work dir and entering the command `bash .command.run`

How do I solve this issue? Thanks!

[Pierre Lindenbaum]

see https://github.com/nf-core/atacseq/issues/54

---

biostats wants so text to validate my comment....

[eesha28112001]

The solution in the link you provided states that a proper profile be mentioned (docker / singularity). However I need to run it using the 'test' profile first. Is there a way to keep profile to 'test' and still run the pipeline without an error?

[Phil Ewels]

It's difficult to know without seeing the command that you used to launch the pipeline, but yes this kind of error typically means that you didn't tell Nextflow how to fetch the required software.

You can chain config profiles with commas, so for example -profile docker,test

[colindaven]

Did the container download work properly ? Are the containers present on your system ? What does this command output on your system: env | grep NXF

If using singularity containers, following can optionally be set (corrected after comment below) and containers will be downloaded to this env variable, so check this is set and writeable: NXF_SINGULARITY_CACHEDIR

On my system, I need to download containers using the helper nf-core python package https://nf-co.re/tools (you need to install and run this separately to your nextflow command). Only then will nf-core pipelines work for me when using singularity containers.

[eesha28112001]

I think my container (singularity) is working fine. The env | grep NXF command shows no output on my terminal.

[Phil Ewels]

If using singularity containers, following needs to be set and containers will be downloaded to this env variable

This is not quite correct. That environment variable will be used for that purpose, if set. But it's not a requirement. If not set then Nextflow will default to downloading them to a folder within the work directory (which is itself, by default, ./work).

This is contingent on Nextflow being told to use Singularity, with -profile singularity,test or via another config (see my comment above).

[colindaven]

Phil Ewels - thanks for the clarification, In my setup direct download via nextflow -profile singularity is blocked, which is why I use the nf-core download to obtain the images.

Is the work directory ./work still used with nf-core download, or do you need to set export NXF_SINGULARITY_CACHEDIR=/home/user/tmp etc ?

[Phil Ewels]

With nf-core download the singularity images will be moved into the target output directory if NXF_SINGULARITY_CACHEDIR is not set (alongside the pipeline code and configs).

Docs are here, but they're a bit verbose / unclear and don't explicitly describe this behaviour as far as I can see.