Human Gut Microbiome AI Analysis Project

This project focuses on analyzing data from the Human Gut Microbiome Project and training AI models to extract insights and make predictions based on microbiome data.

Project Structure

├── data/
│   ├── raw/                  # Raw data from the Human Gut Microbiome Project
│   └── processed/            # Processed data ready for analysis and modeling
├── notebooks/                # Jupyter notebooks for exploration and visualization
├── src/
│   ├── data/                 # Scripts for data downloading and processing
│   ├── eda/                  # Scripts for exploratory data analysis
│   ├── features/             # Scripts for feature engineering and preprocessing
│   ├── models/               # Scripts for model training and evaluation
│   └── utils/                # Utility functions used across the project
└── requirements.txt          # Python dependencies

Setup

1. Create a virtual environment:

   python -m venv venv
   venv\Scripts\activate  # Windows
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Download the data (see data/README.md for details)

Data Sources

The Human Gut Microbiome Project (HMP) provides a wealth of data on the microorganisms that live in the human gut. This project uses data from:

• NIH Human Microbiome Project
• Integrative HMP (iHMP)

Analysis Workflow

1. Exploratory Data Analysis (EDA)
2. Feature Engineering and Preprocessing
3. Model Training and Evaluation
4. Interpretation and Visualization

Models

This project explores various AI models for microbiome data analysis:

• Classification models for disease prediction
• Regression models for quantitative trait prediction
• Clustering models for microbiome community identification
• Deep learning models for complex pattern recognition

BioMed Research Assistant

A Streamlit chat application that allows users to ask questions about biomedicine and gut microbiome research. The application uses SmolAgents to create an intelligent agent that can answer questions directly or search PubMed for relevant research papers.

Features

• Interactive chat interface for asking questions about biomedicine and gut microbiome research
• Intelligent agent that decides whether to answer directly or search for research papers
• PubMed integration for searching and retrieving scientific articles
• Ability to extract article abstracts, citation counts, and links
• Balanced presentation of research that both supports and challenges user hypotheses

Installation

1. Clone this repository
2. Install the required dependencies:

pip install -r requirements.txt

3. Update the email address in pubmed_tools.py with your valid email (required by NCBI):

Entrez.email = "your_email@example.com"  # Replace with a valid email

Usage

1. Run the Streamlit app:

streamlit run app.py

2. Open your web browser and navigate to the URL displayed in the terminal (usually http://localhost:8501)

3. Start asking questions about biomedicine and gut microbiome research!

Example Queries

• "What is the gut microbiome?"
• "How does the gut microbiome affect mental health?"
• "What are the latest research findings on probiotics and gut health?"
• "Is there evidence linking gut bacteria to autoimmune diseases?"
• "What research exists on the relationship between diet and microbiome diversity?"

Architecture

The application is built with the following components:

• Streamlit: Provides the web interface and chat functionality
• SmolAgents: Framework for creating intelligent agents with tool-using capabilities
• BiomedicalAgent: Custom agent implementation that specializes in biomedical research
• PubMed Tools: Custom tools for searching PubMed and extracting article information

Dependencies

• streamlit
• smol-agents
• biopython
• requests
• beautifulsoup4
• lxml

License

MIT License