MaBoSS MCP Server

Model Context Protocol Integration for Boolean Network Simulation

This is a 3. Calls: run_simulation() 4. Calls: get_simulation_result() to analyze dynamics 5. Interprets results: "The simulation shows bistable behavior with p53 activation leading to either DNA repair (oscillatory) or apoptosis (stable high state)"odel Context Protocol (MCP) server** that exposes MaBoSS Boolean network simulation capabilities to Large Language Models (LLMs). It enables LLMs to perform complex Boolean dynamics analysis through natural language interactions and tool chaining.

What is an MCP Server?

Model Context Protocol (MCP) is a standardized way to connect LLMs with external tools and data sources. This server:

• Exposes MaBoSS functionality as callable tools for LLMs
• Enables natural language interaction with Boolean network simulations
• Provides structured tool chaining for complex workflows
• Maintains session state across multiple LLM interactions

LLM Integration Patterns

1. Direct Tool Invocation

LLMs can directly call MaBoSS functions through natural language:

User Prompt: "Convert the BNET file to simulation format and run a Boolean dynamics analysis"

LLM Tool Chain:

1. bnet_to_bnd_and_cfg("Network.bnet")
2. build_simulation("output.bnd", "output.cfg") 
3. run_simulation()
4. get_simulation_result()

2. Exploratory Analysis Workflows

User Prompt: "I have a cancer gene network. Help me understand the steady-state behavior and pathway dynamics."

LLM Response Pattern:

• Identifies available BNET files
• Converts to simulation format
• Runs trajectory analysis
• Interprets biological significance of attractors
• Suggests parameter variations for exploration

3. Cross-Server Coordination

User Prompt: "Build a Boolean network from TP53, MYC, and BAX genes, then simulate the dynamics"

Multi-Server Tool Chain:

NeKo Server:    create_network(['TP53', 'MYC', 'BAX']) → export_network('bnet')
MaBoSS Server:  bnet_to_bnd_and_cfg() → run_simulation() → analyze_results()

Tool Categories Exposed

File Conversion Tools

• bnet_to_bnd_and_cfg() - Convert network definitions to simulation format
• Enable seamless integration with NeKo network construction

Simulation Management

• build_simulation() - Initialize Boolean dynamics engine
• run_simulation() - Execute stochastic simulations
• update_maboss_parameters() - Configure simulation parameters

Analysis Tools

• get_simulation_result() - Extract probability trajectory results
• simulate_mutation() - Test gene knockout/overexpression effects
• visualize_network_trajectories() - Generate trajectory plots

Prompt Engineering Patterns

Pattern 1: Workflow Automation

"Analyze the Boolean dynamics of [gene list] under [conditions]"
→ LLM chains: file_conversion → simulation → analysis → interpretation

Pattern 2: Parameter Exploration

"Test different time scales for this Boolean network"
→ LLM iterates: parameter_variation → simulation → comparison → optimization

Pattern 3: Biological Interpretation

"What do these Boolean attractors mean for cell fate decisions?"
→ LLM combines: simulation_results → biological_knowledge → pathway_analysis

Integration Benefits for LLMs

1. Computational Access: LLMs gain ability to perform quantitative Boolean simulations
2. Workflow Automation: Complex multi-step analyses become single natural language requests
3. Cross-Domain Integration: Combine network construction (NeKo) with simulation (MaBoSS) and multicellular modeling (PhysiCell)
4. State Persistence: Maintain simulation context across conversation turns
5. Error Recovery: Built-in guidance for parameter adjustment and troubleshooting

Example LLM Conversation Flow

User: "I need to simulate the p53 pathway under DNA damage conditions"

LLM with MCP Access:

1. Identifies need for Boolean network simulation
2. Calls: bnet_to_bnd_and_cfg("p53_network.bnet")
3. Calls: build_simulation("output.bnd", "output.cfg")
4. Calls: run_simulation()
5. Calls: get_probtraj() to analyze dynamics
6. Interprets results: "The simulation shows bistable behavior with p53 activation leading to either DNA repair (oscillatory) or apoptosis (stable high state)"

Technical Architecture

• Protocol: Model Context Protocol (MCP) standard
• Interface: JSON-RPC tool calling
• State Management: Session-based simulation context
• Error Handling: Structured error responses with recovery suggestions
• Integration: Seamless connection with NeKo and PhysiCell MCP servers

Getting Started

1. Install MCP client in your LLM environment
2. Connect to MaBoSS MCP server endpoint
3. Use natural language to request Boolean network analysis
4. Chain tools for complex workflows

Learn More About MaBoSS: MaBoSS Official Documentation

This MCP server transforms MaBoSS from a command-line tool into an LLM-accessible computational engine for Boolean network analysis, enabling natural language-driven systems biology research.