SmartSeq - Bugfixes

[sminot]

While using this pipeline to analyze some SmartSeq data, I ran into a couple of issues that needed to be addressed. This PR contains fixes for a small number of minor issues:

• When running the STARsolo tool, account for the case where no whitelist file is provided
• If the SmartSeq protocol is used, set soloUMIdedup to NoDedup
• If only 1 cell is detected, skip the Seurat conversion (since that is a documented limitation of Seurat)
• When running the STARsolo tool, account for case where no reads pass filtering

PR checklist

• This comment contains a description of changes (with reason).
• Make sure your code lints (nf-core pipelines lint).
• Ensure the test suite passes (nextflow run . -profile test,docker --outdir <OUTDIR>).
• Check for unexpected warnings in debug mode (nextflow run . -profile debug,test,docker --outdir <OUTDIR>).
• CHANGELOG.md is updated.

[nf-core-bot]

Warning

Newer version of the nf-core template is available.

Your pipeline is using an old version of the nf-core template: 3.2.0.
Please update your pipeline to the latest version.

For more documentation on how to update your pipeline, please see the nf-core documentation and Synchronisation documentation.

[grst]

Hi @sminot,

thanks for the PR! SmartSeq support has been frequently requested and until now we suggested to use nf-core/rnaseq instead, but I can see how the AnnData/Seurat conversion etc. is useful.

is this for SmartSeq2 specifically? Because I think SmartSeq3 also requires UMI deduplication.

[sminot]

Hi @grst,

Yes, the debugging I did here was on a SmartSeq2 dataset. But hopefully the modifications are still useful even if users provide the appropriate parameters for SmartSeq3 data as well?

[grst]

But hopefully the modifications are still useful even if users provide the appropriate parameters for SmartSeq3 data as well?

I was specifically refering to this statement. For SmartSeq3 UMI deduplication should be enabled.

[sminot]

My apologies, I should have realized that. I agree that the UMI deduplication should be able to enabled by the user. I will leave it to you whether that should be done by setting the protocol param (differentiating SmartSeq2 and SmartSeq3), or simply letting the UMI deduplication be a separate parameter. I can implement if you advise on your preferred approach.

[grst]

PR #488 has additionally just appeared, which tries to achieve something similar. I have to think how to harmonize them.