rnaseq-de

name	rnaseq-de
description	Differential expression analysis for bulk RNA-seq and pseudo-bulk count matrices with QC, PCA, and contrast testing.
version	0.1.0
tags	["rna-seq","differential expression","bulk","pseudo-bulk","transcriptomics","DESeq2","PyDESeq2","QC","PCA"]
trigger_keywords	["rna-seq","differential expression","bulk RNA","pseudo-bulk","volcano plot","ma plot","count matrix","DESeq2","pydeseq2"]
metadata	{"openclaw":{"requires":{"bins":["python3"],"env":[],"config":[]},"always":false,"emoji":"🧬","homepage":"https://github.com/ClawBio/ClawBio","os":["darwin","linux"],"install":[{"kind":"uv","package":"pandas","bins":[]},{"kind":"uv","package":"numpy","bins":[]},{"kind":"uv","package":"matplotlib","bins":[]},{"kind":"uv","package":"scikit-learn","bins":[]}]}}

🧬 RNA-seq Differential Expression

This skill performs differential expression on bulk RNA-seq or pseudo-bulk count matrices.

Core Capabilities

1. Input validation for count matrix and sample metadata
2. Pre-DE QC (library size, detected genes, low-count filtering)
3. PCA visualisation on normalized expression
4. Differential expression from formula + contrast
5. Volcano and MA plots
6. Markdown report with reproducibility files

Input Contract

• Count matrix (.csv or .tsv): rows are genes, columns are samples, first column is gene identifier
• Metadata table (.csv or .tsv): one row per sample, must include sample_id
• Formula: e.g. ~ condition or ~ batch + condition
• Contrast: factor,numerator,denominator (e.g. condition,treated,control)

Output Structure

rnaseq_de_report/
├── report.md
├── figures/
│   ├── pca.png
│   ├── volcano.png
│   └── ma_plot.png
├── tables/
│   ├── qc_summary.csv
│   ├── normalized_counts.csv
│   └── de_results.csv
└── reproducibility/
    ├── commands.sh
    ├── environment.yml
    └── checksums.sha256

Usage

python rnaseq_de.py \
  --counts counts.csv \
  --metadata metadata.csv \
  --formula "~ batch + condition" \
  --contrast "condition,treated,control" \
  --output report_dir

Safety

• Local-only processing
• Warn before overwriting existing output
• Report-level disclaimer required