Context
We presented the wiki/tutorial at my lab meeting on 12-1-25 and received valuable feedback. The consensus was to expand this workflow to cover all major preprocessing steps between 10X sequencing and downstream analysis in Seurat/Scanpy. The main value proposition is getting to data analysis as fast as possible.
Proposed additions
- Genetic demultiplexing
- Implement demultiplexing (e.g., Demuxlet, cellsnp-lite + vireo)
- Support for both genotype-based and genotype-free approaches
- Doublet detection
- Integrate computational doublet detection (e.g., DoubletFinder, Scrublet, or scDblFinder)
- Automated cell type annotation
- Add cell type annotation step using reference-based methods (e.g., Azimuth, SingleR, or CellTypist)
- Allow users to specify custom reference datasets
Updated workflow structure:
10X FASTQ → Cell Ranger count → Cell Ranger aggregate → Demultiplexing → Doublet detection → Cell type annotation → Seurat/Scanpy-ready object
Open questions
- Which specific tools should we prioritize for demultiplexing/doublet detection/annotation?
- Do we want to output both .h5ad (Scanpy) and .rds (Seurat) formats?
Context
We presented the wiki/tutorial at my lab meeting on 12-1-25 and received valuable feedback. The consensus was to expand this workflow to cover all major preprocessing steps between 10X sequencing and downstream analysis in Seurat/Scanpy. The main value proposition is getting to data analysis as fast as possible.
Proposed additions
Updated workflow structure:
10X FASTQ → Cell Ranger count → Cell Ranger aggregate → Demultiplexing → Doublet detection → Cell type annotation → Seurat/Scanpy-ready object
Open questions