Sample Healthcare and Life Sciences Agents on AWS

Project Components

Agents catalog

Library of specialized agents for common workflows across drug research, clinical trials, and commercialization agent-catalog

Multi-agent collaboration

Framework for agent collaboration and knowledge sharing. End to end examples for cancer biomarker discovery, clinical trial protocol assistant, and competitive intelligence. multi-agent-collaboration

Evaluation

Methods for assessing agent performance and result quality. Agent task and goal metrics for cancer biomarker discovery evaluations

The key components are illustrated in the diagram below:

Prerequisites

Please complete the following steps in your AWS account before deployment.

1. Request Amazon Bedrock model access

Request access to the following Amazon Bedrock foundation models using the official documentation.

• Amazon Titan Embeddings G1 - Text
• Amazon Nova Pro
• Anthropic Claude 3.5 Sonnet
• Anthropic Claude 3.5 Sonnet v2
• Anthropic Claude 3 Sonnet

2. Request service quota increase

Request an increase of the Amazon Bedrock "Parameters per function" quota to at least 10.

Deployment

Choose Launch Stack for your preferred region to deploy the toolkit application and selected agents into your AWS account. It may take up to 30 minutes to finish the deployment.

Name	us-east-1	us-west-2
All resources

This will deploy the following agents:

• Cancer Biomarker Discovery (Multi-Agent Supervisor)
• Biomarker Database Analyst
• Clinical Evidence Researcher
• Medical Imaging Expert
• Statistician
• Clinical Trial Protocol Assistant (Multi-Agent Supervisor)
• Clinical Study Research Agent
• Clinical Trial Protocol Generator Agent
• Wiley Open Access Life Sciences Agent
• UniProt Protein Search Agent
• Competitive Intelligence (Multi-Agent Supervisor)
• SEC 10-K Search Agent
• Tavily Web Search Agent
• USPTO Search Agent

In most cases, you may leave the default template parameters unchanged. However, several parameters require new values:

• ReactAppAllowedCidr: CIDR range from where access to the React UI is allowed. Learn about best practices in the AWS VPC documentation. To restrict React App UI access to just your IP, first find your IP address through public websites such as WhatIsMyIpAddress and suffix it with 32. For example, if your ip adress is 192.0.2.0 the minimal CIDR range would be 192.0.2.0/32.

• RedshiftPassword: Password for Redshift database login. Remember and store password securely. Must be at least 8 characters long and contain at least one uppercase letter, one lowercase letter, and one number. Only printable ASCII characters, spaces, and the /, @, ", and ' values may be used.

• [Optional] TavilyApiKey: API key for the Tavily web search API. Required to deploy the Tavily Web Search and Competitive Intelligence Supervisor agents. Please visit tavily.com to create a account and generate an API key.

• [Optional] USPTOApiKey: API key for the US Patent and Trademark Office (USPTO) Open Data Portal. Required to deploy the USPTO Search and Competitive Intelligence Supervisor agents. Please visit the USPO Open Data Portal to create a account and generate an API key.

• [Optional] ExistingVpcId: To use an existing VPC to host the ReactApp, provide the VPC Id here. If you do not provide it will will create a new VPC and required public and private subnets

• [Optional] ExistingPublicSubnets: To use existing public subnets in an existing VPC to host the ReactApp, provide 2 Public Subnet Ids here (In 2 separate availability zones). If you do not provide it will will create a new VPC and required public and private subnets.

• [Optional] ExistingPrivateSubnets: To use existing private subnets in an existing VPC for VPC based resources, provide 2 Private Subnet Ids here (In 2 separate availability zones). If you do not provide it will will create a new VPC and required public and private subnets.

You may also deploy the application using the AWS CLI by first running aws cloudformation package and then aws cloudformation deploy. For example:

set -xe

export BUCKET_NAME="amzn-s3-demo-bucket"
export REGION="us-east-1"
export STACK_NAME="hcls-agents-toolkit"
export TAVILY_API_KEY="1234567890abcdef0"
export USPTO_API_KEY="1234567890abcdef0"
aws cloudformation package --template-file "infra_cfn.yaml" \
  --s3-bucket $BUCKET_NAME \
  --output-template-file "packaged_infra_cfn.yaml" \
  --region $REGION
aws cloudformation deploy --template-file "packaged_infra_cfn.yaml" \
  --s3-bucket $BUCKET_NAME \
  --capabilities CAPABILITY_IAM \
  --stack-name $STACK_NAME \
  --region $REGION \
  --disable-rollback \
  --parameter-overrides \
    RedshiftPassword="MyExamplePassword123$$" \
    ReactAppAllowedCidr="192.0.2.0/24" \
    TavilyApiKey=$TAVILY_API_KEY \
    USPTOApiKey=$USPTO_API_KEY
rm packaged_infra_cfn.yaml

Note

To use existing VPC and subnets, provide them as additional arguments in the --parameter-overrides section above. For example:

--parameter-overrides ExistingVpcId="vpc-i767asdfg8ase87f6" ExistingPublicSubnets=subnet-9867zdf897sa6fe,subnet-7896sadf786sdf ExistingPrivateSubnets=subnet-87d6cv78s6df,subnet-87s6df786asdf78

Access the toolkit application

1. Navigate to AWS CloudFormation via AWS Console search
2. Click the parent stack name that was chosen to deploy the Infra_cfn.yaml
3. In the Outputs tab, find the ReactAppExternalURL link and add 'http://' to the beginning of the URL and paste in your browser
4. You should be able to see a landing page with all (or a subset) deployed agents as shown in the video below:

Multi-agent collaboration for cancer biomarker discovery

Read more about sample agents to accelerate analysis and discovery of cancer biomarkers: https://aws.amazon.com/blogs/machine-learning/accelerate-analysis-and-discovery-of-cancer-biomarkers-with-amazon-bedrock-agents/

The multi-agent solution overview of Cancer biomarker discovery is illustrated below.

Model Context Protocol (MCP)

The Tavily web search and USPTO search tools can be added to your MCP client of choice using the AWS Lambda MCP Server. After installing one of both of them, you can add them to your MCP client configuration:

"hcls-agents": {
   "args": ["awslabs.lambda-mcp-server"],
   "env": {
         "AWS_ACCESS_KEY_ID": "<YOUR AWS ACCESS KEY ID>",
         "AWS_SECRET_ACCESS_KEY": "<YOUR AWS SECRET ACCESS KEY>",
         "AWS_SESSION_TOKEN":"<YOUR AWS SESSION TOKEN>",
         "AWS_REGION": "YOUR AWS REGION",
         "FUNCTION_TAG_KEY": "Application",
         "FUNCTION_TAG_VALUE": "HCLSAgents",
   },
   "command": "uvx"
}

Depending on your local environment, you may need to provide the full path to your uvx executable. View the installation instructions for more ways to provide AWS credentials to the tools.

Contributing Guidelines

Setting Up Your Development Environment

1. Fork the repository to your GitHub account. Ensure the fork remains public during development and testing.

2. Clone your forked repository to your local machine.

3. Add new agents to the agents_catalog folder. You may also include your new agents as stack resources in the infra_cfn.yaml file.

4. For testing purposes, deploy the infra_cfn.yaml template to AWS CloudFormation providing your forked repository and branch as the GithubLink and GitBranch parameters.

AI-Assisted Development

Please use the CONTEXT.md file to provide additional context to coding agents.

Submitting a Pull Request

**Follow the guidelines to contribute a new agent to the catalog here: add-a-new-agent

1. Ensure you have forked the main repository: amazon-bedrock-agents-healthcare-lifesciences

2. Create a new branch in your forked repository for your changes.

3. Implement your changes, following the project's coding standards and guidelines.

4. Commit your changes with clear, concise commit messages.

5. Push your branch to your forked repository on GitHub.

6. Open a pull request from your branch to the main repository's main branch.

7. Provide a clear description of your changes in the pull request, including any relevant issue numbers.

8. Be prepared to address any feedback or questions during the code review process.

License

This project is licensed under the MIT-0 License.

Legal Notes

Important: This solution is for demonstrative purposes only. It is not for clinical use and is not a substitute for professional medical advice, diagnosis, or treatment. The associated notebooks, including the trained model and sample data, are not intended for production. It is each customers’ responsibility to determine whether they are subject to HIPAA, and if so, how best to comply with HIPAA and its implementing regulations. Before using AWS in connection with protected health information, customers must enter an AWS Business Associate Addendum (BAA) and follow its configuration requirements.