GenomeMCP
Enables AI agents to query clinical genomics databases, retrieve supporting literature, analyze population genetics, and visualize biological pathways.
README
GenomeMCP
AI-powered genomic intelligence through the Model Context Protocol
GenomeMCP is a research-grade Model Context Protocol (MCP) server that enables AI agents to query clinical genomics databases, retrieve supporting scientific literature, analyze population genetics, and visualize biological pathways — all in real-time.
🖥️ CLI Tool
GenomeMCP includes a beautiful command-line interface with rich formatting and an interactive TUI mode.
Quick Install
# Recommended (any platform with Python)
pipx install genomemcp
# macOS (Homebrew)
brew install nexisdev/tap/genomemcp
# Windows (Scoop)
scoop bucket add genomemcp https://github.com/nexisdev/scoop-genomemcp
scoop install genomemcp
# From source
git clone https://github.com/nexisdev/GenomeMCP.git
cd GenomeMCP && ./install.sh
Standalone binaries available on GitHub Releases.
CLI Commands
genomemcp search BRCA1 # 🔍 Search ClinVar
genomemcp variant 12345 # 📋 Get variant report
genomemcp gene TP53 # 🧬 Get gene info
genomemcp pathway EGFR --visualize # 🔬 Pathway analysis
genomemcp population 1-55516888-G-GA # 👥 gnomAD frequencies
genomemcp discover "Lynch Syndrome" # 🔗 Discover related genes
genomemcp tui # 🖥️ Interactive mode
Theme Options
genomemcp --theme cyberpunk search BRCA1
genomemcp --theme professional gene TP53
genomemcp --theme minimal pathway EGFR
See CLI Guide for complete documentation.
🎯 Why GenomeMCP?
| Problem | GenomeMCP Solution |
|---|---|
| AI agents lack genomic knowledge | Direct ClinVar, gnomAD, Reactome integration |
| No evidence for clinical claims | Auto-retrieves PubMed abstracts |
| Variant interpretation is complex | Population frequency + pathway context |
| Gene-disease links are opaque | Automatic relationship discovery |
🧬 Features
Core Genomics Tools
search_clinvar(term)— Query ClinVar for genes, variants, or diseasesget_variant_report(id)— Detailed clinical significance reportget_gene_info(symbol)— Gene function, location, and aliases from NCBI Geneget_supporting_literature(id)— PubMed articles linked to a variant
Population Genetics
get_population_stats(variant)— Allele frequency from gnomAD (Genome Aggregation Database)
Pathway Analysis
get_pathway_info(gene)— Reactome biological pathways for a genevisualize_pathway(gene)— Generate Mermaid.js diagrams of gene-pathway relationships
Discovery & Synthesis
find_related_genes(phenotype)— Discover genes associated with a diseaseget_genomic_context(gene, position)— Identify exon vs intron regionsget_discovery_evidence(phenotype)— Aggregate PubMed abstracts for AI reasoning
🚀 Quick Start
MCP Server Installation
# Clone the repository
git clone https://github.com/nexisdev/GenomeMCP.git
cd GenomeMCP
# Install dependencies with uv
uv sync
# Run the MCP server
uv run python src/main.py
CLI Installation
# Using the install script
./install.sh
# Or with pip
pip install genomemcp[cli]
# Or for development
./setup-dev.sh
source .venv/bin/activate
Claude Desktop Integration
Add to your claude_desktop_config.json:
{
"mcpServers": {
"genomemcp": {
"command": "uv",
"args": [
"--directory",
"/path/to/GenomeMCP",
"run",
"python",
"src/main.py"
]
}
}
}
### ☁️ Cloud Deployment (Railway)
You can deploy the GenomeMCP server to the cloud with one click. It will be exposed as an SSE (Server-Sent Events) endpoint, ready for remote agents.
1. Click the **Deploy on Railway** button above.
2. Provide your `SUPABASE_URL` and `SUPABASE_KEY` (optional, for persistence).
3. Connect your agent to the deployment URL (e.g. `https://your-app.up.railway.app/sse`).
---
## 📖 Usage Examples
### Search for a Gene Variant
User: "What variants are associated with BRCA1?" Agent uses: search_clinvar("BRCA1")
### Get Population Frequency
User: "How common is the variant 1-55516888-G-GA?" Agent uses: get_population_stats("1-55516888-G-GA") → Returns gnomAD allele frequency: 0.000123 (0.01%)
### Discover Gene-Disease Relationships
User: "What genes are linked to Lynch Syndrome?" Agent uses: find_related_genes("Lynch Syndrome") → Returns: MSH2 (12 variants), MLH1 (8 variants), PMS2 (5 variants)
### Visualize Pathways
User: "Show me the pathways for TP53" Agent uses: visualize_pathway("TP53") → Returns Mermaid diagram:
```mermaid
graph TD
TP53((TP53))
TP53 --> P_123["Transcriptional Regulation by TP53"]
TP53 --> P_456["Cell Cycle Checkpoints"]
TP53 --> P_789["DNA Damage Response"]
🔬 Data Sources
| Source | Description | API |
|---|---|---|
| ClinVar | Clinical variant interpretations | NCBI E-utilities |
| gnomAD | Population allele frequencies | gnomAD GraphQL |
| Reactome | Biological pathway database | Reactome Content Service |
| PubMed | Scientific literature | NCBI E-utilities |
| NCBI Gene | Gene annotations | NCBI E-utilities |
🏗️ Architecture
GenomeMCP/
├── src/
│ ├── main.py # MCP server & tool definitions
│ ├── clinvar.py # ClinVar & PubMed API client
│ ├── genomics.py # Exon/Intron mapping
│ ├── population.py # gnomAD integration
│ ├── pathways.py # Reactome integration
│ ├── utils.py # Shared utilities
│ └── cli/ # Command-line interface
│ ├── app.py # Typer CLI application
│ ├── formatters/ # Rich output formatters
│ ├── tui/ # Textual interactive UI
│ └── config.py # Theme configuration
├── tests/ # Unit tests
├── docs/ # Documentation
├── install.sh # Quick install script
├── setup-dev.sh # Development setup
└── pyproject.toml # Project configuration
🧪 Testing
# Run all tests
uv run pytest
# Run CLI tests
uv run pytest tests/test_cli.py -v
# Run specific test suite
uv run pytest tests/test_phase4.py tests/test_phase5.py
📚 Documentation
- CLI Guide — Command-line interface documentation
- Tool Reference — Complete API documentation
- Architecture Guide — System design
🤝 Contributing
Contributions are welcome! Please open an issue or submit a pull request.
📄 License
MIT License — see LICENSE for details.
🔗 Keywords
genomics bioinformatics clinvar gnomad mcp model-context-protocol ai-agent claude variant-interpretation population-genetics reactome pathway-analysis pubmed ncbi gene-discovery clinical-genomics precision-medicine llm-tools cli tui terminal
<p align="center"> <strong>Built for AI agents. Powered by open genomic data.</strong> </p>
<div align="center"> <img src="https://raw.githubusercontent.com/Nexis-AI/branding-assets/refs/heads/main/nex_banner.png" alt="Nexis AI Banner" width="100%" /> </div>
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。