Cellranger mkref resource flags incorrect

[RPSeaman]

Description of the bug

Hey all, I am looking at the cellranger mkref module and noticed that automatic available resources detection has been added, but I think I have spotted an inconsistency in the flags used for memory and cores:

In their docs (https://www.10xgenomics.com/support/software/cell-ranger/latest/resources/cr-command-line-arguments) they specify --memgb and --nthreads as the two flags for cpu and memory, not --localmem and --localcores as specified in the module.

Let me know you further clarification is needed.

Command used and terminal output

No response

Relevant files

No response

System information

nf-core/scrnaseq 2.6.0

[grst]

--localmem and --localcore are global options passed to the martian runtime (martian is actually a workflow manager similar to nextflow). They should apply to all cellranger tools that are based on martian and mkref was changed from a simple script to the martian runtime in the last release.

https://www.10xgenomics.com/support/software/cell-ranger/latest/resources/cr-command-line-arguments

However, given the documentation you point at I'm wondering if we should additionally pass --nthreads because otherwise star index generation likely only runs on a single core which takes forever.

If you could update the module on nf-core/modules that would be fantastic!

[grst]

Fixed in nf-core/modules#6136

[grst]

fixed in dev