Hello everyone, I am trying to run the rnaseq pipeline on HPC with Slurm executor. I read the documentation but I am really (like really) confused about the configuration of the pipeline. What I am sure about is that I need to use Singularity for containerized software, but I don't get if `nextflow run nf-core/rnaseq -profile singularity`is enough by itself. I think because I need Singularity and to specify Slurm as executor, it's better to use a config from https://github.com/nf-core/configs, but how do I know which one is for me? What is typical/standard/safe? (I am a junior so forgive me 🙂)

hi, can you first check if your institution maybe already has a config? otherwise I would start from a simple one like https://nf-co.re/configs/adcra/ and adapt it to your needs (you then don't need to specify

-profile singularity

because it is already enabled in the example profile.

(duplicating @mashehu mostly, but I started so I'll finish!) Using the singularity profile will define how software is delivered to your processes, but by default it won't affect your executor so all jobs will run on your 'local' machine. At a minimum you'll have to also set

process.executor

to

slurm

in your config, but you will likely also need some pieces specific to settings to your local compute cluster- like the example above.

Hi, nope, they don't have a config! Thank you! Does adapting it to my needs mean I let the pipeline run with this profile for example, and if it fails I adjust cpus/memory/time?

Ideally you'd talk to someone from your HPC team and they'd tell you what settings to use.

Got it, thanks!

See here for some more reading material regarding configs. Investing a bit of time to get this right can safe you a lot of problems and pay of in efficiency gains.

Thank you!!