Module/scdblfinder

[KurayiChawatama]

PR checklist

• This comment contains a description of changes (with reason).

• added an adaptation of my implementation of scDblFinder 's regular random sampling-based doublet detection method as a module to the pipeline. (If this works fine, we can add the clustering-dependent method too as soon as tomorrow)

• It can be run using the settings --doublet_detection scdblfinder

• It also takes a new optional parameter doublet_rate via a columnin the samplesheet with an expected float doublet rate between 0 and 1. If not provided, scDblFinder estimates it internally.

• If you've fixed a bug or added code that should be tested, add tests!
  Super basic tests to run without errors were added to match the other R-based modules

• If you've added a new tool - have you followed the pipeline conventions in the contribution docs
  Done!

• If necessary, also make a PR on the nf-core/scdownstream branch on the nf-core/test-datasets repository.
  No need, works with the usual fitered testdataset that other doublet detection tools use

• Make sure your code lints (nf-core pipelines lint).
  It does

• Ensure the test suite passes (nextflow run . -profile test,docker --outdir <OUTDIR>).
  **My module runs fine within the pipeline tests, quit after I saw it complete, retest needed on your end @nictru as you mentioned in the Slack thread **

• Check for unexpected warnings in debug mode (nextflow run . -profile debug,test,docker --outdir <OUTDIR>).
  **Didn't bother with this either as per the same thread in Slack by @nictru **

• Usage Documentation in docs/usage.md is updated.

• Output Documentation in docs/output.md is updated.

• CHANGELOG.md is updated.

• README.md is updated (including new tool citations and authors/contributors).

[KurayiChawatama]

@nf-core-bot fix linting

[Copilot]

Pull request overview

Adds a new doublet-detection option to the nf-core/scdownstream pipeline by integrating an scDblFinder-based method as a local module, exposing it via the existing --doublet_detection parameter and wiring it into the doublet detection subworkflow.

Changes:

• Added a new local Nextflow module SCDBLFINDER (conda env, container, R implementation, nf-test coverage).
• Integrated scdblfinder into the DOUBLET_DETECTION subworkflow and pipeline test configs.
• Updated user-facing documentation and citations to include scDblFinder.

Reviewed changes

Copilot reviewed 16 out of 16 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
subworkflows/local/doublet_detection/main.nf	Adds SCDBLFINDER execution path and mixes its outputs into predictions/versions channels.
nextflow_schema.json	Extends doublet_detection allowed methods and help text to include scdblfinder.
modules/local/doublet_detection/scdblfinder/main.nf	New Nextflow process definition for the scDblFinder module.
modules/local/doublet_detection/scdblfinder/templates/scdblfinder.R	Implements scDblFinder execution, multiplet-rate estimation, and prediction export.
modules/local/doublet_detection/scdblfinder/environment.yml	Defines conda dependencies for the new module.
modules/local/doublet_detection/scdblfinder/meta.yml	Adds module metadata, I/O definitions, and tool citation details.
modules/local/doublet_detection/scdblfinder/tests/main.nf.test	Adds nf-test coverage for real and stub runs.
modules/local/doublet_detection/scdblfinder/tests/main.nf.test.snap	Adds test snapshots for the module outputs.
conf/modules.config	Adds publishDir/prefix handling for SCDBLFINDER outputs.
conf/test.config	Enables scdblfinder in the test profile’s doublet_detection methods.
conf/test_full.config	Enables scdblfinder in the full test profile’s doublet_detection methods.
docs/output.md	Documents scDblFinder as a supported doublet detection tool and output folder.
README.md	Lists scDblFinder among supported doublet detection tools.
CITATIONS.md	Adds scDblFinder citation entry.
CHANGELOG.md	Notes addition of scDblFinder module.
ro-crate-metadata.json	Updates embedded project description to include scDblFinder in the tool list.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[nictru]

Looks good in general. I did not look through the R code in detail yet (will do that in the evening)

[KurayiChawatama]

@nictru want me to change all the doublet detection tools to use mix in a seperate PR then?

[nictru]

You can also do it in this PR, I am not very strict about this

[nictru]

Also please include it here

[KurayiChawatama]

@nf-core-bot fix linting

[Copilot]

Pull request overview

Copilot reviewed 19 out of 19 changed files in this pull request and generated 4 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[nictru]

I think things start to clear up. For you to better understand my rationale:

Generally all modules in the pipeline follow the same schema:

1. Read input
2. Apply function
3. Write output

So that the pipeline modules are simple wrappers around the R/python functions.
Deviations from this schema (additional manipulations etc) are sometimes necessary but require clear justification.

If you look at how complicated the R script was in the beginning vs what it is now, this becomes apparent. Helps with keeping overview, finding bugs, and everything else

[nictru]

I found that scDblFinder has an optional samples argument, which is highly relevant here. You should pass the batch_col to it, which you can get via the ch_batch_col in the doublet detection subworkflow

[nictru]

I think this is also unnecessary because the sce is created from the anndata in a way that is perfectly prepared for reversing the process