BioAgent — Agentic Bioinformatics Discovery Platform

BioAgent is an AI-powered research assistant that enables scientists to perform complex structural biology tasks through natural language. Unlike traditional pipelines, it uses an autonomous agent (LangGraph) to reason, plan, and execute workflows dynamically.

System Architecture

The platform connects a modern React frontend with a Python-based agentic backend, orchestrating specialized bioinformatics tools.

graph TB
    subgraph Frontend["Frontend (Vite + React + TypeScript)"]
        Chat[Chat Interface]
        NGL[NGL 3D Viewer]
        Cyto[Cytoscape.js Maps]
        Panels[Result Panels]
    end

    subgraph Backend["Backend (FastAPI + Python)"]
        API[FastAPI REST + SSE]
        Agent[LangGraph Agent]
        LLM[LLM Provider Layer]
        Tools[Bio Tool Registry]
        Eval[Eval / Trace Logger]
        DB[(SQLite)]
    end

    subgraph BioTools["Bio Tools (Local Binaries + APIs)"]
        AF[AlphaFold DB API]
        P2R[P2Rank - Pocket Detection]
        Vina[AutoDock Vina - Docking]
        PLIP_T[PLIP - Interactions]
        FS[Foldseek - Structure Search]
    end

    Chat -->|SSE stream| API
    API --> Agent
    Agent -->|tool calls| Tools
    Agent --> LLM
    Agent --> Eval
    Tools --> AF
    Tools --> P2R
    Tools --> Vina
    Tools --> PLIP_T
    Tools --> FS
    API -->|results| NGL
    API -->|results| Cyto
    API -->|results| Panels
    Agent --> DB

Loading

Key Features

• Natural Language Interface: "Find binding pockets in p53 and dock aspirin into the largest one."
• Autonomous Reasoning: The agent plans steps, recovers from errors, and chains multiple tools.
• Integrated Bio-Tools:
  • AlphaFold DB: Instant structure retrieval.
  • P2Rank: Machine learning-based pocket detection.
  • AutoDock Vina: Molecular docking simulation.
  • Foldseek: Structural similarity search (10,000x faster than TM-align).
• Visual-First: embedded NGL viewer for 3D structures and Cytoscape for interaction maps.

Repository Structure

Directory	Purpose
/backend	FastAPI server, LangGraph agent, and tool wrappers.
/frontend	React UI, visualizations, and chat logic.

🚀 Getting Started

This is a monorepo. You will need to set up both the backend and frontend.

1. Backend Setup

Go to the backend/ directory and follow the instructions to install Python dependencies and download bio-tools.

cd backend
pip install uv
uv sync

2. Frontend Setup

Go to the frontend/ directory to install Node.js dependencies and start the dev server.

cd frontend
npm install
npm run dev

📜 License

MIT License.