Bff

[LuisHeinzlmeier]

• Generating cell hashing calls from a matrix of count data by calling GenerateCellHashingCalls
• This modules was created to be included in the hadge pipeline
• Currently the module do not work for CLUSTER because the test data is probably too small

[LuisHeinzlmeier]

@nictru, I don't understand why I get this error:

ERROR ~ No such file or directory: https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/homo_sapiens/10xgenomics/cellranger/hashing_demultiplexing/hto

[itrujnara]

Looks good. Just a few minor remarks to consider.

[LuisHeinzlmeier]

@nictru, why are htodemux changes also included here?

[nictru]

@nictru, why are htodemux changes also included here?

This was because the branch in your fork was not up-to-date with the main branch of the nf-core repo - as you can see it is fixed now, as the branch was updated

[nictru]

@LuisHeinzlmeier your No such file or directory is because URLs can only ever point to files, never to directories. While nextflow can handle directories as inputs, it cannot resolve them when referenced via URL.

The same issue was already encountered and fixed in one of the other demultiplexing modules that reference the same data, as you can see here.

Btw, thanks @itrujnara for the review

@LuisHeinzlmeier, if you wonder what Igor means with Harshil alignment, you can read about it here

[LuisHeinzlmeier]

@itrujnara , thanks for the review! I adopted the feedback.

And, @nictru, thank you very much for the help! I can add the directory-url now as input and the test is running locally (snapshots are up to date). However only for docker I have an error. I think this error might arise from Seurat v4 was just loaded with SeuratObject v5 because according to the container build of seqera Seurat v5 is incompatible with cellhashr.

[itrujnara]

@itrujnara , thanks for the review! I adopted the feedback.

And, @nictru, thank you very much for the help! I can add the directory-url now as input and the test is running locally (snapshots are up to date). However only for docker I have an error. I think this error might arise from Seurat v4 was just loaded with SeuratObject v5 because according to the container build of seqera Seurat v5 is incompatible with cellhashr.

The error originates from the OS (it's a SIGKILL), so it is probably not a Seurat issue. It is more likely that the script OOMs or executes illegal instructions, but just from the CI output it's hard to tell which is the case.

[itrujnara]

I've found minor issues in the meta. Once they are resolved, I'll approve the PR so you can merge as soon as you fix the CI issues.

[nictru]

For testing purposes I increased the memory limit for all processes from 4GB to 6GB. It seems to work now, so this means that running the BFF tests in docker uses between 4 and 6 GB of memory, while singularity and conda stay below the 4GB limit.

As increasing the memory limit is probably not allowed, we will have to revert this. However, we now know what the problem is and need to reduce the memory footprint of the tests

[itrujnara]

Everything looks fine now. Feel free to merge once you fix the CI issues.