GeneAgent: Self-verification Language Agent for Gene Set Analysis

Quick Start

# Clone and setup
git clone https://github.com/Uoghluvm/GeneAgent-openai-1.x.git
cd GeneAgent
conda env create -f environment.yml -y
conda activate geneagent

# Configure API (edit .env file with your credentials)
cp ".env copy" .env

# Run
python main_cascade.py

Workflows

GeneAgent provides three workflows for different analysis needs:

• English Tutorial | 中文教程
• Cascade (main_cascade.py) - High-accuracy with self-verification
• Chain-of-Thought (main_CoT.py) - Fast structured reasoning
• Summary (main_summary.py) - Enrichment analysis

Example Output

Process: MAPK Signaling Pathway
The proteins ERBB2, ERBB4, FGFR2, FGFR4, HRAS, and KRAS are integral components of the MAPK signaling pathway, crucial for cell growth, differentiation, and survival...

Evaluation: Use evaluate.ipynb for result analysis

Demo Website

Live Demo: https://www.ncbi.nlm.nih.gov/CBBresearch/Lu/Demo/GeneAgent/

Datasets

• Gene Ontology: 1000 gene sets from GO:BP branch
• MsigDB: 56 hallmark gene sets
• NeST: 50 gene sets from human cancer proteomics

Source data: LLM Evaluation | Talisman

---

Citation: DOI: 10.5281/zenodo.15008591

Funding: NIH National Library of Medicine Intramural Research Programs

Disclaimer: Research tool - not for clinical diagnosis. Consult healthcare professionals for medical decisions.