BDB-Genomics RNA-seq Pipeline

A modular, industry-standard Snakemake workflow for paired-end bulk RNA-seq processing — from raw FASTQ files to quantified gene expression, QC gating, and comprehensive reporting.

Status

• Fully resolved DAG with validated I/O contracts
• 4 new rules added: markduplicates, rseqc, preseq, deseq2_prep
• Ready for integration testing and iteration toward industry-standard parity

Repository Layout

.
├── Snakefile                 # Main Snakemake entry point
├── config.yaml               # Central configuration (all rules)
├── rules/
│   ├── *.smk                # Modular rule files
│   ├── envs/                 # Per-rule Conda environments
│   └── scripts/              # Helper scripts (validate_config.py, etc.)
├── profiles/
│   ├── local/               # Local execution profile
│   └── slurm/                # SLURM cluster profile
├── scripts/
│   ├── run_pipeline.sh      # Wrapper script for execution
│   └── directory_structure.sh # Create expected directory layout
└── data/
    └── fastp/samples.tsv     # Sample sheet

Inputs

The pipeline expects:

• config.yaml — Central configuration file
• data/fastp/samples.tsv — Sample sheet with FASTQ paths and metadata
• Reference files — STAR genome index, GTF annotation, RSeQC refGene BED (configure in config.yaml)

Sample Sheet Format

data/fastp/samples.tsv must contain:

Column	Description
sample	Unique sample identifier
fastq_r1	Path to R1 FASTQ
fastq_r2	Path to R2 FASTQ
replicate	Replicate number (positive integer)
condition	Experimental condition

Optional: control — path to control sample or NONE

Example:

sample	fastq_r1	fastq_r2	replicate	condition
sample1	data/fastq/sample1_R1.fastq.gz	data/fastq/sample1_R2.fastq.gz	1	control
sample2	data/fastq/sample2_R1.fastq.gz	data/fastq/sample2_R2.fastq.gz	1	treatment

Quick Start

1. Configure

Edit config.yaml and set:

global:
  samples: "data/fastp/samples.tsv"
  index: "/path/to/star/index"        # Generate with STAR --runMode genomeGenerate
  gtf: "/path/to/annotation.gtf"      # GTF annotation file
  refgene: "/path/to/refgene.bed"     # RSeQC reference gene BED

star:
  params:
    index: "/path/to/star/index"
    sjdbOverhang: 100                  # Read length - 1

2. Dry Run

scripts/run_pipeline.sh --dry-run

3. Run

scripts/run_pipeline.sh --cores 8

4. SLURM Cluster

snakemake --profile profiles/slurm --configfile config.yaml

Configuration

config.yaml is organized by rule. Key sections:

Section	Purpose
global	Samples, references (STAR index, GTF, refGene)
fastp	Trimming parameters
fastqc	Post-trim QC
star	Splice-aware alignment
samtools_sort	BAM sorting
markduplicates	PCR duplicate marking
samtools_index	BAM indexing
samtools_stats	Alignment statistics
rseqc	RNA-seq QC metrics
preseq	Library complexity estimation
featurecounts	Gene quantification
deseq2_prep	Normalized counts, PCA, correlation
qc_gate	Threshold-based QC filtering
multiqc	Aggregate QC reporting

Outputs

Directory	Contents
results/fastp/	Trimmed FASTQs, JSON/HTML reports
results/fastqc/	FastQC HTML/ZIP reports
results/star/	Aligned BAMs, alignment logs
results/markduplicates/	Dup-marked BAMs, metrics
results/samtools_index/	BAM indices
results/samtools_stats/	Alignment statistics
results/rseqc/	Gene body coverage, strand inference, read distribution
results/preseq/	Library complexity curves
results/featurecounts/	Raw counts matrix
results/deseq2/	VST-normalized counts, PCA, correlation
results/qc_gate/	QC pass/fail per sample
results/multiqc/	Aggregated QC report

Requirements

• Snakemake ≥7.0
• Conda (for --use-conda)
• Python ≥3.9
• Reference files (STAR index, GTF, refGene BED)

License

See LICENSE.

---

Built by BDB-Genomics — advancing discoveries through modular, industry-standard computational tools.