Biomarker Agentic Q&A System

A comprehensive offline testing framework for multi-domain biomarker analysis using agentic workflows. This system provides mock LLM clients and demo agents to test the complete workflow without making actual API calls.

Features

• Multi-Domain Support: Methylation, Metagenomics, Proteomics, Transcriptomics, Whole Exome, and Whole Genome analysis
• Agentic Workflow: Analysis, Summary, and Recommendation agents for each domain
• Offline Testing: Complete testing suite using mock LLM clients
• Comprehensive Logging: Detailed logs for each agent and domain
• Test Data: Realistic test data for all domains
• Prompt Management: Centralized prompt management system

Project Structure

biomarker-agentic-qna/
├── agents/                          # Agent implementations
│   ├── common/                      # Common agent types
│   ├── methylation/                 # Methylation domain agents
│   ├── metagenomics/               # Metagenomics domain agents
│   ├── proteomics/                 # Proteomics domain agents
│   ├── transcriptomics/            # Transcriptomics domain agents
│   ├── whole_exome/                # Whole exome domain agents
│   ├── whole_genome/               # Whole genome domain agents
│   └── demo_agent.py               # Demo agent for testing
├── utils/                          # Utility modules
│   ├── prompt_manager.py           # Prompt management system
│   ├── mock_llm_client.py          # Mock LLM client for offline testing
│   └── ...
├── user_data/                      # Test data for each domain
│   ├── methylation/
│   ├── metagenomics/
│   ├── proteomics/
│   ├── transcriptomics/
│   ├── whole_exome/
│   └── whole_genome/
├── testing_logs/                   # Test execution logs and reports
├── test_agentic_workflow.py        # Comprehensive testing script
├── test_simple_demo.py             # Simple demo test script
└── README.md                       # This file

Installation

1. Clone the repository:

   git clone <your-repo-url>
   cd biomarker-agentic-qna

2. Install dependencies (if any):

   pip install -r requirements.txt

3. Run the tests:

   # Quick test
   python3 test_simple_demo.py
   
   # Full comprehensive test
   python3 test_agentic_workflow.py

Testing

Quick Test

Run a simple test to verify the system works:

python3 test_simple_demo.py

Comprehensive Test

Run the full test suite that tests all agents and domains:

python3 test_agentic_workflow.py

Test Results

After running tests, check the results in:

• Terminal output: Real-time test results
• testing_logs/: Detailed logs and reports
  • test_execution.log: Main test execution log
  • test_report.json: Machine-readable test results
  • Domain-specific logs: testing_logs/<domain>/<agent>_agent.log

Components

Mock LLM Client

• File: utils/mock_llm_client.py
• Purpose: Simulates LLM responses without making actual API calls
• Features: Domain-specific response templates, realistic response generation

Demo Agent

• File: agents/demo_agent.py
• Purpose: Tests the complete agentic workflow
• Features: Prompt formatting, user data loading, comprehensive logging

Prompt Manager

• File: utils/prompt_manager.py
• Purpose: Manages prompts for all domains and agents
• Features: Dynamic prompt loading, template formatting

Supported Domains

Domain	Description	Agents
Methylation	DNA methylation and epigenetic analysis	Analysis, Summary, Recommendation
Metagenomics	Gut microbiome analysis	Analysis, Summary, Recommendation
Proteomics	Protein expression analysis	Analysis, Summary, Recommendation
Transcriptomics	Gene expression analysis	Analysis, Summary, Recommendation
Whole Exome	Whole exome sequencing analysis	Analysis, Summary, Recommendation
Whole Genome	Whole genome analysis	Analysis, Summary, Recommendation

Understanding the Logs

Log Structure

Each domain has its own log directory with agent-specific logs:

testing_logs/
├── methylation/
│   ├── analysis_agent.log
│   ├── summary_agent.log
│   └── recommendation_agent.log
├── metagenomics/
│   └── ...
└── ...

Log Levels

• INFO: Successful operations, processing times, response details
• ERROR: Failed operations, missing data, formatting errors
• WARNING: Non-critical issues, fallback to mock data

Test Reports

• Success Rate: Overall test success percentage
• Processing Times: Time taken for each agent
• Response Lengths: Size of generated responses
• Error Details: Specific error messages and causes

Usage Examples

Running a Single Agent Test

from agents.demo_agent import DemoAgent

# Test methylation analysis
agent = DemoAgent(domain="methylation", agent_type="analysis", user_id="test_user_001")
result = agent.process()
print(f"Success: {result['success']}")
print(f"Response: {result['response'][:200]}...")

Testing Multiple Domains

domains = ["methylation", "metagenomics", "proteomics"]
for domain in domains:
    agent = DemoAgent(domain=domain, agent_type="analysis", user_id="test_user_001")
    result = agent.process()
    print(f"{domain}: {'✓' if result['success'] else '✗'}")

Customization

Adding New Domains

1. Create domain directory in agents/
2. Add prompts.json with agent templates
3. Create test data in user_data/<domain>/
4. Update domain mapping in demo_agent.py

Modifying Test Data

Edit files in user_data/<domain>/<user_id>_report.txt to customize test scenarios.

Customizing Mock Responses

Modify utils/mock_llm_client.py to change response templates and behavior.

Performance Metrics

The system tracks:

• Processing Time: Time taken for each agent
• Success Rate: Percentage of successful tests
• Response Quality: Length and content of responses
• Error Analysis: Detailed error tracking and categorization

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests for new functionality
5. Submit a pull request

License

[Add your license information here]

Support

For issues and questions:

1. Check the logs in testing_logs/
2. Review the test reports
3. Open an issue with detailed error information

---

Note: This is an offline testing system designed for development and validation. For production use, replace the mock LLM client with actual LLM API clients.