🧬 cBioPortal MCP Server
A high-performance, production-ready Model Context Protocol (MCP) server that enables AI assistants to seamlessly interact with cancer genomics data from cBioPortal. Built with modern async Python architecture, enterprise-grade modular design, and BaseEndpoint pattern for maximum reliability, maintainability, and 4.5x faster performance.
🌟 Overview & Key Features
- ⚡ 4.5x Performance Boost: Full async implementation with concurrent API operations
- 🏗️ Enterprise Architecture: BaseEndpoint pattern with 60% code duplication elimination
- 📐 Modular Design: Professional structure with 71% code reduction (1,357 → 396 lines)
- 📦 Modern Package Management: uv-based workflow with pyproject.toml
- 🔄 Concurrent Operations: Bulk fetching of studies and genes with automatic batching
🔧 Enterprise Features
- ⚙️ Multi-layer Configuration: CLI args → Environment variables → YAML config → Defaults
- 📋 Comprehensive Testing: 93 tests across 8 organized test suites with full coverage
- 🛡️ Input Validation: Robust parameter validation and error handling
- 📊 Pagination Support: Efficient data retrieval with automatic pagination
- 🔧 Code Quality: Ruff linting, formatting, and comprehensive code quality checks
- ⚡ Configurable Performance: Adjustable batch sizes and performance tuning
🧬 Cancer Genomics Capabilities
- 🔍 Study Management: Browse, search, and analyze cancer studies
- 🧪 Molecular Data: Access mutations, clinical data, and molecular profiles
- 📈 Bulk Operations: Concurrent fetching of multiple entities
- 🔎 Advanced Search: Keyword-based discovery across studies and genes
🎆 Recent Quality & Architecture Improvements
🚀 Major Refactoring Achievements (2025)
- 🏗️ BaseEndpoint Architecture: Eliminated ~60% code duplication through inheritance-based design
- 📝 Code Quality Excellence: Comprehensive external review integration with modern linting (Ruff)
- ⚙️ Enhanced Configurability: Gene batch sizes, retry logic, and performance tuning now configurable
- 🛡️ Robust Validation: Decorator-based parameter validation and error handling
- 🧪 Testing Maturity: 93 comprehensive tests with zero regressions through major refactoring
📈 Production-Ready Status
- ✅ External Code Review: Professional code quality validation and improvements implemented
- 🔧 Modern Python Practices: Type checking, linting, formatting, and best practice adherence
- 🏗️ Enterprise Architecture: Modular design with clear separation of concerns
- 🚀 Performance Optimized: 4.5x async improvements with configurable batch processing
🧠🤖 AI-Collaborative Development
This project demonstrates cutting-edge human-AI collaboration in bioinformatics software development:
- 🧠 Domain Expertise: 20+ years cancer research experience guided architecture and feature requirements
- 🤖 AI Implementation: Advanced code generation, API design, and performance optimization through systematic LLM collaboration
- 🔄 Quality Assurance: Iterative refinement ensuring professional standards and production reliability
- 🏗️ Architectural Evolution: BaseEndpoint pattern and 60% code duplication elimination through AI-guided refactoring
- 📈 Innovation Approach: Showcases how domain experts can effectively leverage AI tools to build enterprise-grade bioinformatics platforms
Recent Achievements: External code review integration with comprehensive quality improvements including Ruff configuration, configurable performance settings, and modern Python best practices.
Methodology: This collaborative approach combines deep biological domain knowledge with AI-powered development capabilities, accelerating innovation while maintaining rigorous code quality and scientific accuracy.
🚀 Quick Start
Prerequisites
- Python 3.10+ 🐍
- uv (modern package manager) - recommended 📦
- Git (optional, for cloning)
⚡ Installation & Launch
# Install uv if needed
pipx install uv
# Clone and setup
git clone https://github.com/yourusername/cbioportal-mcp.git
cd cbioportal-mcp
uv sync
# Launch server
uv run cbioportal-mcp
That's it! 🎉 Your server is running and ready for AI assistant connections.
📦 Installation Options
🔥 Option 1: uv (Recommended)
Modern, lightning-fast package management with automatic environment handling:
# Install uv
pipx install uv
# Or with Homebrew: brew install uv
# Clone repository
git clone https://github.com/yourusername/cbioportal-mcp.git
cd cbioportal-mcp
# One-command setup (creates venv + installs dependencies)
uv sync
🐍 Option 2: pip (Traditional)
Standard Python package management approach:
# Create virtual environment
python -m venv cbioportal-mcp-env
# Activate environment
# Windows: cbioportal-mcp-env\Scripts\activate
# macOS/Linux: source cbioportal-mcp-env/bin/activate
# Install dependencies
pip install -e .
⚙️ Configuration
🎛️ Multi-Layer Configuration System
The server supports flexible configuration with priority: CLI args > Environment variables > Config file > Defaults
YAML Configuration 📄
Create config.yaml for persistent settings:
# cBioPortal MCP Server Configuration
server:
base_url: "https://www.cbioportal.org/api"
transport: "stdio"
client_timeout: 480.0
logging:
level: "INFO"
format: "%(asctime)s - %(name)s - %(levelname)s - %(message)s"
api:
rate_limit:
enabled: false
requests_per_second: 10
retry:
enabled: true
max_attempts: 3
backoff_factor: 1.0
cache:
enabled: false
ttl_seconds: 300
batch_size:
genes: 100 # Configurable gene batch size for concurrent operations
Environment Variables 🌍
export CBIOPORTAL_BASE_URL="https://custom-instance.org/api"
export CBIOPORTAL_LOG_LEVEL="DEBUG"
export CBIOPORTAL_CLIENT_TIMEOUT=600
export CBIOPORTAL_GENE_BATCH_SIZE=50 # Configure gene batch size
export CBIOPORTAL_RETRY_MAX_ATTEMPTS=5
CLI Options 💻
# Basic usage
uv run cbioportal-mcp
# Custom configuration
uv run cbioportal-mcp --config config.yaml --log-level DEBUG
# Custom API endpoint
uv run cbioportal-mcp --base-url https://custom-instance.org/api
# Generate example config
uv run cbioportal-mcp --create-example-config
🔌 Usage & Integration
🖥️ Claude Desktop Integration
Configure in your Claude Desktop MCP settings:
Option 1: Direct Script Path (Recommended)
{
"mcpServers": {
"cbioportal": {
"command": "/path/to/your/project/cbioportal_MCP/.venv/bin/cbioportal-mcp",
"env": {
"CBIOPORTAL_LOG_LEVEL": "INFO"
}
}
}
}
Option 2: uv run (Alternative)
{
"mcpServers": {
"cbioportal": {
"command": "uv",
"args": ["run", "cbioportal-mcp"],
"cwd": "/path/to/your/project/cbioportal_MCP",
"env": {
"CBIOPORTAL_LOG_LEVEL": "INFO"
}
}
}
}
Important Setup Steps:
- Replace
/path/to/your/project/cbioportal_MCP with your actual project path - Ensure the project is installed in editable mode:
uv pip install -e . - Restart Claude Desktop after updating the configuration
🔧 VS Code Integration
Add to your workspace settings:
{
"mcp.servers": {
"cbioportal": {
"command": "uv",
"args": ["run", "cbioportal-mcp"],
"cwd": "/path/to/cbioportal-mcp"
}
}
}
🏃♂️ Command Line Usage
# Development server with debug logging
uv run cbioportal-mcp --log-level DEBUG
# Production server with custom config
uv run cbioportal-mcp --config production.yaml
# Using custom cBioPortal instance
uv run cbioportal-mcp --base-url https://private-instance.org/api
🏗️ Architecture
📁 Modern Project Structure
cbioportal-mcp/
├── 📁 cbioportal_mcp/ # Main package directory
│ ├── 📊 server.py # Main MCP server implementation
│ ├── 🌐 api_client.py # Dedicated HTTP client class
│ ├── ⚙️ config.py # Multi-layer configuration system
│ ├── 📋 constants.py # Centralized constants
│ ├── 📁 endpoints/ # Domain-specific API modules
│ │ ├── 🏗️ base.py # BaseEndpoint pattern (60% duplication reduction)
│ │ ├── 🔬 studies.py # Cancer studies & search
│ │ ├── 🧬 genes.py # Gene operations & mutations
│ │ ├── 🧪 samples.py # Sample data management
│ │ └── 📈 molecular_profiles.py # Molecular & clinical data
│ └── 📁 utils/ # Shared utilities
│ ├── 📄 pagination.py # Efficient pagination logic
│ ├── ✅ validation.py # Input validation
│ └── 📝 logging.py # Logging configuration
├── 📁 tests/ # Comprehensive test suite (93 tests)
├── 📁 docs/ # Documentation
├── 📁 scripts/ # Development utilities
└── 📄 pyproject.toml # Modern Python project config
🎯 Design Principles
- 🔧 Modular: Clear separation of concerns with domain-specific modules
- ⚡ Async-First: Full asynchronous implementation for maximum performance
- 🏗️ BaseEndpoint Pattern: Inheritance-based architecture eliminating 60% code duplication
- 🛡️ Robust: Comprehensive input validation and error handling with decorators
- 🧪 Testable: 93 tests ensuring reliability and preventing regressions
- 🔄 Maintainable: Clean code architecture with 71% reduction in complexity
- 📝 Code Quality: Ruff linting, formatting, and modern Python practices
The server provides 12 high-performance tools for AI assistants:
| 🔧 Tool | 📝 Description | ⚡ Features |
|---|
get_cancer_studies | List all available cancer studies | 📄 Pagination, 🔍 Filtering |
search_studies | Search studies by keyword | 🔎 Full-text search, 📊 Sorting |
get_study_details | Detailed study information | 📈 Comprehensive metadata |
get_samples_in_study | Samples for specific studies | 📄 Paginated results |
get_genes | Gene information by ID/symbol | 🏷️ Flexible identifiers |
search_genes | Search genes by keyword | 🔍 Symbol & name search |
get_mutations_in_gene | Gene mutations in studies | 🧬 Mutation details |
get_clinical_data | Patient clinical information | 👥 Patient-centric data |
get_molecular_profiles | Study molecular profiles | 📊 Profile metadata |
get_multiple_studies | 🚀 Concurrent study fetching | ⚡ Bulk operations |
get_multiple_genes | 🚀 Concurrent gene retrieval | 📦 Automatic batching |
get_gene_panels_for_study | Gene panels in studies | 🧬 Panel information |
- ⚡ Concurrent Operations:
get_multiple_* methods use asyncio.gather for parallel processing - 📦 Smart Batching: Automatic batching for large gene lists
- 📄 Efficient Pagination: Async generators for memory-efficient data streaming
- ⏱️ Performance Metrics: Execution timing and batch count reporting
📊 Benchmark Results
Our async implementation delivers significant performance improvements:
🏃♂️ Sequential Study Fetching: 1.31 seconds (10 studies)
⚡ Concurrent Study Fetching: 0.29 seconds (10 studies)
🎯 Performance Improvement: 4.57x faster!
🔥 Async Benefits
- 🚀 4.5x Faster: Concurrent API requests vs sequential operations
- 📦 Bulk Processing: Efficient batched operations for multiple entities
- ⏱️ Non-blocking: Asynchronous I/O prevents request blocking
- 🧮 Smart Batching: Automatic optimization for large datasets
- Use
get_multiple_studies for fetching multiple studies concurrently - Leverage
get_multiple_genes with automatic batching for gene lists - Configure
concurrent_batch_size in config for optimal performance - Monitor execution metrics included in response metadata
👨💻 Development
🔨 Development Workflow
# Setup development environment
uv sync
# Run tests
uv run pytest
# Run with coverage
uv run pytest --cov=.
# Run specific test file
uv run pytest tests/test_server_lifecycle.py
# Update snapshots
uv run pytest --snapshot-update
# Lint code
uv run ruff check .
# Format code
uv run ruff format .
🧪 Testing
Comprehensive test suite with 93 tests across 8 categories:
- 🔄
test_server_lifecycle.py - Server startup/shutdown & tool registration - 📄
test_pagination.py - Pagination logic & edge cases - 🚀
test_multiple_entity_apis.py - Concurrent operations & bulk fetching - ✅
test_input_validation.py - Parameter validation & error handling - 📸
test_snapshot_responses.py - API response consistency (syrupy) - 💻
test_cli.py - Command-line interface & argument parsing - 🛡️
test_error_handling.py - Error scenarios & network issues - ⚙️
test_configuration.py - Configuration system validation
- 📦 uv: Modern package management (10-100x faster than pip)
- 🧪 pytest: Testing framework with async support and 93 comprehensive tests
- 📸 syrupy: Snapshot testing for API response consistency
- 🔍 Ruff: Lightning-fast linting, formatting, and code quality enforcement
- 📊 pytest-cov: Code coverage reporting and quality metrics
- 🏗️ BaseEndpoint: Inheritance pattern eliminating 60% code duplication
- ⚙️ Type Checking: Comprehensive type annotations for better code safety
- 🛡️ Validation Decorators: Automatic parameter validation and error handling
🤝 Contributing
- 🍴 Fork the repository
- 🌿 Create a feature branch (
git checkout -b feature/amazing-feature) - ✅ Test your changes (
uv run pytest) - 📝 Commit with clear messages (
git commit -m 'Add amazing feature') - 🚀 Push to branch (
git push origin feature/amazing-feature) - 🔄 Create a Pull Request
🔧 Troubleshooting
🚨 Common Issues
Server Fails to Start
# Check Python version
python --version # Should be 3.10+
# Verify dependencies
uv sync
# Check for conflicts
uv run python -c "import mcp, httpx, fastmcp; print('Dependencies OK')"
Claude Desktop Connection Issues
- ✅ Use direct script path (Option 1) for most reliable connection
- ✅ Verify paths in MCP configuration are absolute (no
~ or relative paths) - ✅ Install in editable mode: Run
uv pip install -e . in project directory - ✅ Ensure the virtual environment
.venv/bin/cbioportal-mcp script exists - ✅ For Option 2: Check that
uv is in your system PATH and cwd points to project directory - ✅ Review Claude Desktop logs for detailed errors
- 🔧 Increase
concurrent_batch_size in config - 🔧 Adjust
max_concurrent_requests for your system - 🔧 Use
get_multiple_* methods for bulk operations - 🔧 Monitor network latency to cBioPortal API
Configuration Problems
# Generate example config
uv run cbioportal-mcp --create-example-config
# Validate configuration
uv run cbioportal-mcp --config your-config.yaml --log-level DEBUG
# Check environment variables
env | grep CBIOPORTAL
🌐 API Connectivity
# Test cBioPortal API accessibility
curl https://www.cbioportal.org/api/cancer-types
# Test with custom instance
curl https://your-instance.org/api/studies
💡 Examples & Use Cases
🔍 Research Queries
"What cancer studies are available for breast cancer research?"
"Search for melanoma studies with genomic data"
"Get mutation data for TP53 in lung cancer studies"
"Find clinical data for patients in the TCGA-BRCA study"
"What molecular profiles are available for pediatric brain tumors?"
🧬 Genomic Analysis
"Compare mutation frequencies between two cancer studies"
"Get all genes in the DNA repair pathway for ovarian cancer"
"Find studies with both RNA-seq and mutation data"
"What are the most frequently mutated genes in glioblastoma?"
📊 Bulk Operations
"Fetch data for multiple cancer studies concurrently"
"Get information for a list of cancer genes efficiently"
"Compare clinical characteristics across multiple studies"
"Retrieve molecular profiles for several cancer types"
📜 License
This project is licensed under the MIT License - see the LICENSE file for details.
🙏 Acknowledgments
- 🧬 cBioPortal - Open-access cancer genomics data platform
- 🔗 Model Context Protocol - Enabling seamless AI-tool interactions
- ⚡ FastMCP - High-performance MCP server framework
- 📦 uv - Modern Python package management
- 🤖 AI Collaboration - Demonstrating the power of human-AI partnership in scientific software development
🌟 Production-ready bioinformatics platform built through innovative human-AI collaboration! 🧬✨
Demonstrating the power of domain expertise + AI-assisted development for enterprise-grade scientific software.