jjgao/cbioportal-mcp-server

cBioPortal MCP Server

A Model Context Protocol (MCP) server that provides access to cBioPortal's cancer genomics data through a standardized interface.

Features

Tools

Basic Data Access:

• get_studies: Search and retrieve cancer studies
• get_study_details: Get detailed information about specific studies
• get_samples: Retrieve sample information from studies
• get_patients: Get patient information from studies
• get_sample_lists: Get pre-defined sample lists (cohorts) for studies

Molecular Data:

• get_molecular_profiles: Get available molecular data types for studies
• get_mutations: Query mutation data for specific genes
• get_copy_number_alterations: Get copy number alteration data
• get_molecular_data: Access expression or other molecular data

Clinical Data:

• get_clinical_data: Access patient and sample clinical information
• get_clinical_attributes: Get available clinical attributes for studies

Gene & Reference Data:

• search_genes: Search for genes by symbol or keyword
• get_gene_panels: Get gene panels available for studies
• get_cancer_types: List all cancer types in cBioPortal

Advanced Analysis:

• get_significantly_mutated_genes: Get MutSig results for studies
• get_survival_data: Access survival data for patients
• get_treatment_data: Get treatment information for patients or samples

Resources

• cbioportal://studies: Browse all available cancer studies
• cbioportal://cancer-types: List all cancer types
• cbioportal://genes: Gene information and search
• cbioportal://gene-panels: Available gene panels used in sequencing
• cbioportal://molecular-profiles: All molecular profile types across studies
• cbioportal://server-info: Server status and version information

Installation

1. Clone and setup:

mkdir cbioportal-mcp-server
cd cbioportal-mcp-server
npm init -y

2. Install dependencies:

npm install @modelcontextprotocol/sdk axios
npm install -D typescript @types/node tsx

3. Copy the source code to src/index.ts

4. Build the project:

npm run build

Usage

Running the Server

npm start

Integration with Claude Desktop

Add to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "cbioportal": {
      "command": "node",
      "args": ["/path/to/your/cbioportal-mcp-server/build/index.js"]
    }
  }
}

Example Queries

Once integrated with Claude, you can ask questions like:

Basic Exploration:

• "What cancer studies are available for lung cancer?"
• "Show me details about the TCGA lung adenocarcinoma study"
• "List the sample types available in the breast cancer study"
• "What clinical attributes are collected for prostate cancer patients?"

Molecular Data Queries:

• "Get mutation data for TP53 and KRAS in the lung cancer study"
• "Show me copy number alterations for EGFR in breast cancer samples"
• "What molecular profiles are available for the brain tumor study?"
• "Get gene expression data for MYC in the liver cancer cohort"

Clinical Analysis:

• "What clinical data is available for breast cancer patients?"
• "Show me survival data for the lung adenocarcinoma study"
• "Get treatment information for colorectal cancer patients"
• "What are the significantly mutated genes in the melanoma study?"

Advanced Queries:

• "Compare mutation rates between primary and metastatic samples"
• "Find studies with both RNA-seq and protein data available"
• "Get gene panel information for targeted sequencing studies"
• "Show me cancer types with the most available studies"

API Rate Limiting

The server respects cBioPortal's API limits:

• Implements 30-second timeouts
• Handles pagination appropriately
• Includes error handling for API failures

Development

Running in Development Mode

npm run dev

Building

npm run build

Project Structure

cbioportal-mcp-server/
├── src/
│   └── index.ts          # Main server implementation
├── build/                # Compiled JavaScript (generated)
├── package.json
├── tsconfig.json
└── README.md

Contributing

This is a starting implementation. You can extend it by:

1. Adding more cBioPortal endpoints
2. Implementing caching for better performance
3. Adding authentication support for private instances
4. Enhancing error handling and logging
5. Adding data visualization tools

License

MIT License