Evolved Sprint-Hackathon

A production-ready platform for therapeutic binder optimization with selectivity guidance. This project provides a complete pipeline for protein design, including a web-based UI for pipeline configuration and job management, and a backend API for running BoltzGen-based design pipelines.

📚 Documentation

Frontend (UI)

• UI README - Next.js frontend setup, features, and usage guide
  • Pipeline editor with multiple operating modes (Standard Boltzgen, Binder Optimization, Dual Target)
  • YAML editor with schema validation
  • Job results viewers (CSV tables, YAML, PDF, Mol* structure visualization)

Backend (API)

• Backend README - Backend setup, configuration, and development guide

  • Sanic-based REST API
  • PostgreSQL and Redis integration
  • Docker Compose setup with GPU support
  • Database migrations and configuration

• API Documentation - Complete API reference

  • File upload endpoints
  • Design job management
  • File serving endpoints
  • Request/response examples with cURL and Python

• Hackathon README - Fine-grained binder design pipeline guide

  • Pipeline overview and architecture
  • Configuration guide for YAML files
  • Parameter tuning recommendations
  • Output interpretation
  • Examples and best practices

🚀 Quick Start

Prerequisites

• Frontend: Node.js 18.18+ (Node 20+ recommended), pnpm 8+
• Backend: Python 3.12+, uv, Docker & Docker Compose, PostgreSQL 17, Redis 7, NVIDIA GPU with CUDA 12.6+

Frontend Setup

cd ui
pnpm install
pnpm dev
# App runs on http://localhost:3000

See UI README for detailed setup instructions.

Backend Setup

cd backend
uv sync
export ENV_FOR_DYNACONF=local
docker-compose up -d db redis
uv run alembic upgrade head
uv run sanic --dev server:create_app --host 0.0.0.0 --port 8000
# API runs on http://localhost:8000

See Backend README for detailed setup instructions.

🏗️ Project Structure

evolved_sprint_hackathon_2025/
├── ui/                    # Next.js frontend application
│   ├── src/              # Source code
│   ├── public/           # Static assets
│   └── README.md         # UI documentation
├── backend/              # Sanic backend API
│   ├── api/              # API routes
│   ├── boltzgen/         # BoltzGen ML pipeline
│   ├── core/             # Core utilities
│   ├── API.md            # API documentation
│   ├── README.md         # Backend documentation
│   └── README_HACKATHON.md  # Pipeline guide
└── README.md             # This file

🎯 Features

Pipeline Modes

• Standard Boltzgen: Standard protein design pipeline
• Binder Optimization: Optimize existing binders with selectivity scoring
• Dual Target: Design for existing + unknown targets

File Viewers

• CSV Tables: Interactive tables with column toggling and resizing
• YAML Editor: Monaco-based editor with syntax highlighting
• PDF Viewer: Full PDF viewing with page navigation
• Mol Viewer*: 3D protein structure visualization
• Image Viewer: Support for PNG, JPG, SVG, and other image formats

📖 Getting Started

1. Set up the backend following the Backend README
2. Set up the frontend following the UI README
3. Configure your pipeline using the web UI or YAML configuration files
4. Run design jobs and view results in the job results page

For detailed information on:

• API endpoints: See API.md
• Pipeline configuration: See README_HACKATHON.md
• UI features: See ui/README.md

🔗 Links

• UI Documentation
• Backend Documentation
• API Reference
• Pipeline Guide