First public release.
This version introduces major structural updates, improved performance, and clearer input handling.
π New
- New step:
CONCAT_FASTQS
Collapses lane-split FASTQs into single R1/R2 files per sample.- Skips samples that already have a single pair.
- Standardizes naming for clean downstream matching.
β¨ Improvements
- Inputs & Groups
- Restructured input style: added clearer group ID handling in sample maps.
- Whitelist CSV is now stored directly in the repository.
- FIND_LARRY_SEQS
- Underlying
CBUtoolsand Dockerfile changes enable parallelisation, yielding >4Γ faster runtime.
- Underlying
- LARRY_QC
- Added more logging between filters for transparency.
- Fixed QC plots and text in the PDF report.
- Whitelist directory now defaults to the repo path.
- ASSIGN_BARCODES (formerly
ASSIGN_CLONES)- Renamed for clarity and consistency with the poster/README.
- Refactored to support the new input structure.
- Removed
combine_samplesfunctionality (now handled via group IDs).
- MATCH_GEX
- Added support for Cell Ranger outputs (in addition to STARsolo).
- Refactored to support the new input structure.
π Internal
- Updated Dockerfile: streamlined how
CBUtoolsis installed and used. - Modified process resources and updated Singularity image.
- Cleaned up example files/scripts.
- Removed old git submodule.