BCP Analysis Pipeline

A Nextflow-based pipeline for single-cell RNA sequencing data preprocessing, specifically designed for BCP (Billion Cell Program) analysis. The pipeline performs comprehensive single-cell data processing including alignment, ambient RNA removal, doublet detection, and basic quality control metrics.

Features

• Single-cell alignment using STAR
• Ambient RNA removal for cleaner expression profiles
• Doublet detection to filter multiplets
• Comprehensive QC metrics with MultiQC reporting
• Embedded QC plots included in MultiQC report
• Automated preprocessing with minimal manual intervention

Future Development

Parameters will be optimized and made configurable for different species and organ types to enhance pipeline flexibility and accuracy.

Environment Configuration

Current Setup

The pipeline currently requires manual conda environment configuration for testing and development purposes.

Planned Updates

Singularity container support will be implemented upon completion of the development phase to ensure better reproducibility and easier deployment across different computing environments.

Installation

# Clone the repository
git clone https://github.com/PhrenoVermouth/BCP_analysis.git
cd BCP_analysis

# Create conda environment
mamba env create -f bin/environment.yml

Usage

# Run the pipeline
nextflow run ~/BCP_analysis/main.nf -profile standard

Run modes

• Genefull only (default):

  nextflow run ~/BCP_analysis/main.nf --run_mode genefull

• Velocity using prior GeneFull outputs: rerun in the same project directory after a completed GeneFull run. The pipeline will reuse ${outdir}/soupx/<sample>/<sample>_corrected.h5ad by default.

  nextflow run ~/BCP_analysis/main.nf --run_mode velocity

  If GeneFull outputs live elsewhere, optionally add a counts_h5ad column in samples.csv to point to custom .h5ad locations.

Scrublet threshold override

• By default, Scrublet uses its automatic threshold.

• To force a manual cutoff, pass --scrublet_manual_threshold when launching Nextflow; this value is forwarded to run_scrublet.py --manual_threshold.

  nextflow run ~/BCP_analysis/main.nf --scrublet_manual_threshold 0.25

Mitochondrial filtering

• Global threshold (default): --max_mito sets a single mitochondrial percentage cutoff for all samples (default: 0.2).

• Per-sample overrides: provide --mito_max_map /path/to/file where each line maps one or more sample IDs to a cutoff using the format sample1, sample2 = value. Blank lines and lines starting with # are ignored. Any sample not listed will fall back to --max_mito.

  Example (resource/AC.mito):

  efm, em, fatfm, fatm, fbfm, fbm, hbfm, hbm, kdfm, kdm, lvfm, lvm, mbfm, mbm = 0.2
  hfm, hm, ifm, im, pcf, pcm, skfm, skm = 0.6
  lf, lm, smf, smm, spm, spfm = 0.4
  

Output

The pipeline generates:

• Processed single-cell count matrices
• Quality control reports via MultiQC
• Doublet detection results
• Ambient RNA removal metrics