Exclusive Content:

Haiper steps out of stealth mode, secures $13.8 million seed funding for video-generative AI

Haiper Emerges from Stealth Mode with $13.8 Million Seed...

“Revealing Weak Infosec Practices that Open the Door for Cyber Criminals in Your Organization” • The Register

Warning: Stolen ChatGPT Credentials a Hot Commodity on the...

VOXI UK Launches First AI Chatbot to Support Customers

VOXI Launches AI Chatbot to Revolutionize Customer Services in...

Create a Biomedical Research Agent Using Biomni Tools and Amazon Bedrock AgentCore Gateway

Accelerating Biomedical Research: Leveraging AI Agents with Amazon Bedrock and Biomni

Introduction

Unlocking the potential of AI in the biomedical research space by integrating advanced agents with robust infrastructures for enhanced data discovery and analysis.

Bridging the Complexity Gap from Prototype to Production


1. Enterprise Security for AI Agents

Addressing the security challenges in deploying AI agents while conducting sensitive biomedical research.

2. Contextual Management in Research Workflows

Ensuring continuity and context in biomedical investigations across diverse research sessions.

3. The Role of a Scalable Tool Gateway

Facilitating efficient access and management of specialized biomedical tools for enhanced research productivity.


A Comprehensive Solution Overview

Implementing a research agent leveraging specialized tools for improved biomedical analyses and outcomes.

Step 1: Creating a Gateway for the Biomni Tools

Simplifying agent access to diverse databases and applications through a centralized gateway.

Step 2: Deploying Strands Research Agent

Integrating local tools within the research agent framework for enhanced usability.

Step 3: Building Persistent Memory for Enhanced Context

Utilizing memory services to improve contextual awareness in research.

Step 4: Deploying and Managing with AgentCore Runtime

Streamlining deployment processes for effective, scalable research capabilities.

Case Studies: Agents in Action

Exploring representative use cases showcasing the versatility of AI agents in various biomedical research scenarios.


Future Directions in Biomedical Research AI

Anticipating advancements to support ongoing tool evolution and multi-agent collaborations.

Conclusion

Demonstrating the potential of AI agents in transforming biomedical research, ensuring scalability, security, and integration.


About the Authors

Profiles of the individuals driving this transformative research initiative.

Accelerating Biomedical Research with AI Agents: The Role of Biomni and Amazon Bedrock AgentCore

This post is co-authored with the Biomni group from Stanford.

Biomedical researchers are often inundated with vast amounts of information, spending roughly 90% of their time manually processing data. A striking example is Genentech’s challenge of analyzing 38 million biomedical publications alongside numerous databases like the Human Protein Atlas. The rapid growth of specialized databases and analytical tools in genomics, proteomics, and pathology means researchers must stay updated on their capabilities, sidelining the hypothesis-driven work that spurs breakthrough discoveries.

Enter AI agents. Powered by foundation models, they offer a game-changing solution by autonomously planning, executing, and adapting complex research tasks. The Biomni project developed at Stanford embodies this potential; it is a versatile biomedical AI agent that integrates over 150 specialized tools, 105 software packages, and 59 databases to facilitate sophisticated analyses like gene prioritization and drug repurposing.

The Challenge of Production-Ready AI in Biomedical Research

Deploying such agents in a production environment demands a robust infrastructure that effectively addresses the challenges of computational workflows, user access, and data security. Amazon’s Bedrock AgentCore serves as a comprehensive solution for deploying and operating these intelligent agents with enterprise-grade security and scalability.

This post details how to construct a research agent using AgentCore, leveraging access to over 30 specialized biomedical database tools from Biomni, ultimately expediting scientific discovery while ensuring robust security and scalability.

Bridging the Prototype-to-Production Gap

Moving from a localized prototype to a production-ready system accessible by various research teams presents a range of infrastructure challenges.

Security Concerns in Agent Deployment

Enterprise security issues are paramount. These involve implementing OAuth-based authentication, secure tool sharing, and thorough observability for auditing research workflows. Many promising projects never make it to production due to the complexity of fulfilling enterprise-grade requirements while retaining the specialized expertise necessary for precise biomedical analyses.

Context Management for Research Relevance

Biomedical research often unfolds across multiple conversations, requiring an enduring memory of previous analyses, experimental parameters, and research preferences. Research agents must retain contextual knowledge while ensuring isolation between different projects in a multi-tenant environment.

Scalable Tool Gateway Implementation

Creating a reusable tool gateway is vital for accommodating concurrent requests from research agents and managing authentication effectively. The gateway not only needs to facilitate tool discovery but also must manage authentication securely, making it critical for scaling research efforts.

Overview of the Solution

This article employs Strands Agents, an open-source framework, to construct a research agent that integrates local tools for PubMed searches and extends its capabilities via Biomni’s databases.

Key Components of the AgentCore Framework

  1. AgentCore Gateway: Centralizes access to biomolecular tools through secure endpoints with semantic search capabilities.
  2. AgentCore Memory: Maintains contextual awareness across research sessions.
  3. AgentCore Identity: Manages user authentication and tool access.
  4. AgentCore Runtime: Streamlines deployment, enabling scalable, managed service configurations.
  5. AgentCore Observability: Provides critical monitoring for auditing research processes.

Practical Steps to Set Up the Research Agent

  1. Create a Centralized Tool Gateway: The AgentCore Gateway service helps centralize tools into secure endpoints, making them reusable and simplifying agent access.

    def create_gateway(gateway_name: str, api_spec: list) -> dict:
       # JWT authentication with Cognito
       auth_config = {
           "customJWTAuthorizer": {
               "allowedClients": [
                   get_ssm_parameter("/app/researchapp/agentcore/machine_client_id")
               ],
               "discoveryUrl": 
                   get_ssm_parameter("/app/researchapp/agentcore/cognito_discovery_url"),
           }
       }
    
       ...
    
    create_gateway("MyAppTemplateAgentCoreGateway", api_spec)
  2. Implement PubMed Tool in the Agent: Using the Strands framework to extend the agent’s capabilities for comprehensive biomedical literature search.

    from agent.agent_config.tools.PubMed import PubMed
    
    @tool(name="Query_pubmed")
    def query_pubmed(query: str) -> str:
       pubmed = PubMed()
       result = pubmed.run(query)
       return result
  3. Add Persistent Memory: This allows the agent to maintain context across sessions by integrating with the AgentCore Memory service.

  4. Deploy with AgentCore Runtime: Configure and launch the agent as a managed service.

    agentcore configure --entrypoint agent/main.py -er arn:aws:iam::<Account-Id>:role/<Role> --name researchapp<AgentName>
  5. Test Your Agent: Use automated scripts to verify the tool functionalities.

Illustrative Research Scenarios

  • Drug Mechanism Analysis: Analyzing trastuzumab’s mechanism and resistance mechanisms by querying relevant biomedical databases.
  • Genetic Variant Investigation: Investigating clinical significance of BRCA1 variants in breast cancer.

Scalability and Observability

As research workflows grow more unpredictable, having scalable and observable AI agents becomes crucial. Amazon Bedrock AgentCore Observability allows teams to monitor agent performance, ensuring timely issue resolution critical for maintaining research momentum.

Future Directions

The future of AI in biomedicine is promising. Further enhancements, such as integrating multi-agent collaboration and specialized research tools tailored to particular therapeutic areas, can significantly evolve research capabilities.

Conclusion

This implementation demonstrates how Amazon Bedrock AgentCore can transform biomedical research prototypes into production systems. By leveraging Biomni’s extensive toolset integrated through the AgentCore Gateway, research teams can establish a secure and scalable infrastructure. The project exemplifies how AI can catalyze breakthroughs in biomedical research, enabling researchers to focus more on hypothesis-driven work rather than data management.

To get started on building your own biomedical research agent with Biomni tools, explore our GitHub repository for complete code and documentation, and join the collaborative journey in advancing biomedical AI.

About the Authors

Hasan Poonawala, Pierre de Malliard, Necibe Ahat, and Kexin Huang are leading experts in AI solutions for healthcare, focusing on enhancing biomedical research through innovative technology.


For more insights into the implementation of AI agents in biomedical research, visit [GitHub repository link].

Latest

Chinese Astronauts Return to Earth After Being Stranded, Yet Space Junk Hazards Persist

Chinese Astronauts Safely Return to Earth Amid Ongoing Space...

Exploitation of ChatGPT via SSRF Vulnerability in Custom GPT Actions

Addressing SSRF Vulnerabilities: OpenAI's Patch and Essential Security Measures...

This Startup Is Transforming Touch Technology for VR, Robotics, and Beyond

Sensetics: Pioneering Programmable Matter to Digitize the Sense of...

Leveraging Artificial Intelligence in Education and Scientific Research

Unlocking the Future of Learning: An Overview of Humata...

Don't miss

Haiper steps out of stealth mode, secures $13.8 million seed funding for video-generative AI

Haiper Emerges from Stealth Mode with $13.8 Million Seed...

VOXI UK Launches First AI Chatbot to Support Customers

VOXI Launches AI Chatbot to Revolutionize Customer Services in...

Investing in digital infrastructure key to realizing generative AI’s potential for driving economic growth | articles

Challenges Hindering the Widescale Deployment of Generative AI: Legal,...

Microsoft launches new AI tool to assist finance teams with generative tasks

Microsoft Launches AI Copilot for Finance Teams in Microsoft...

A Comprehensive Guide to Developing AI Agents in GxP Environments

Leveraging Generative AI in GxP-Compliant Healthcare: Key Strategies and Frameworks Transforming Healthcare with AI in Regulatory Environments A Risk-Based Framework for Implementing AI Agents in GxP...

Introducing Agent-to-Agent Protocol Support in Amazon Bedrock’s AgentCore Runtime

Unlocking Seamless Collaboration: Introducing Agent-to-Agent (A2A) Protocol on Amazon Bedrock AgentCore Runtime Maximize Efficiency and Interoperability in Multi-Agent Systems Explore how Amazon Bedrock AgentCore Runtime empowers...

Optimize VLMs for Multipage Document-to-JSON Conversion Using SageMaker AI and SWIFT

Leveraging Intelligent Document Processing: Unleashing the Power of Vision Language Models for Accurate Document-to-JSON Conversion Overview of Intelligent Document Processing Challenges and Solutions Advancements in Document...