What is the best way to use scanpy in slurm cluster


We have to use a very large scRNA combination of datasets (overall more than 4M of cells) and we need to use a cluster. I have not seen a clear way to use scanpy in a cluster. Is there any example or any recommendation of “best practice” ?

Thank you very much for your time!


I don’t think any special handling will be needed. You should just use scanpy as you usually would in a script, and submit that as a job. Some functions have a njobs parameter which could be useful.

Did you have anything specific you were wondering about?