Biomart MCP

🚀 Biomart MCP Server

An MCP server integrated with Biomart

The Model Context Protocol (MCP) is an open protocol designed to standardize the way applications provide context for Anthropic's LLMs. We've created an MCP server using the MCP Python SDK, which interacts with Biomart via the pybiomart package.

There's a short demo video that shows the MCP server in action within Claude Desktop.

🚀 Quick Start

📦 Installation

Clone the repository

git clone https://github.com/jzinno/biomart-mcp.git
cd biomart-mcp

Run on Claude Desktop

uv run --with mcp[cli] mcp install --with pybiomart biomart-mcp.py

Use in Cursor

Other models can leverage the MCP server through Cursor's proxy mode, such as those from OpenAI or DeepSeek. Click the gear icon in Cursor settings, navigate to MCP, and then add the MCP server in the global configuration or project scope. For example, create a .cursor/mcp.json file in the project directory:

{
    "mcpServers": {
        "Biomart": {
            "command": "uv",
            "args": [
                "run",
                "--with",
                "mcp[cli]",
                "--with",
                "pybiomart",
                "mcp",
                "run",
                "/your/path/to/biomart-mcp.py"
            ]
        }
    }
}

Use in Glama

Development

# Create a virtual environment
uv venv

# MacOS/Linux
source .venv/bin/activate

# Windows
.venv\Scripts\activate

uv sync # or uv add mcp[cli] pybiomart

# Run the server in development mode
mcp dev biomart-mcp.py

✨ Features

Biomart-MCP offers various tools to interact with the Biomart database:

• Database and Service Discovery: List available Marts and datasets, and explore the Biomart database structure.
• Attribute and Filter Exploration: View general or all available attributes and filters for a specific dataset.
• Data Retrieval: Query Biomart for biological data using specific attributes and filters.
• Identifier Conversion: Convert between different biological identifiers (e.g., convert gene symbols to Ensembl IDs).

🔧 Contribution

Pull requests are welcome! Here are some development notes:

• We only use @mcp.tool() here to maximize compatibility with MCP-supporting clients, as described in the documentation.
• We use @lru_cache to cache the results of functions that are computationally expensive or call external APIs.
• Be mindful not to exceed the model's context window size, e.g., maintain proper formatting and conciseness in code blocks.

🚧 Future Possible Features

• Extended Attribute Support: Increase support for more biomedical-related attributes.
• Advanced Filtering: Develop more complex query logic to meet advanced user requirements.
• Result Visualization: Integrate data visualization tools to help users better understand analysis results.