RepurposAI is a modular, open-source web app for drug repurposing and target discovery.
It integrates ML, cheminformatics, and the Open Targets Platform.
Hackathon issue: Open Targets Hackathon — Project #5
- Install dependencies:
pip install -r requirements.txt - Run the Streamlit app:
streamlit run app/main.py - Explore modules: similarity search, target prediction, pathway mapping
Here’s the project structure:
RepurposAI/
├── README.md
├── requirements.txt
├── .gitignore
├── app/
│ ├── main.py # Streamlit main app
│ ├── dashboard.py # Dashboard assembly module
│ ├── visualization.py # Plotly/Seaborn visualization module
│ └── utils/
│ ├── api_integration.py # Open Targets & KEGG API helpers
│ ├── similarity.py # RDKit similarity functions
│ └── ml_model.py # Placeholder for target prediction ML
├── data/ # Sample data or placeholder CSVs
├── docs/ # README, usage guide, hackathon slides
├── tests/ # Basic test scripts for functions
└── notebooks/ # Jupyter notebooks for testing ML/API modules
└── example_notebook.ipynb
- Fork the repository and pick a task from the GitHub Epic.
- Implement the placeholder modules in
/appand/app/utils. - Add sample Jupyter notebooks in
/notebooksto test modules. - Start building Streamlit pages in
/app/main.pyand/app/dashboard.py. - Commit your work regularly and create issues/sub-issues for new features or bugs.
- Tag issues with relevant labels:
backend,frontend,api,ml,docs,streamlit,rdkit.
This project is released under the Apache 2.0 License, in line with the Open Targets Hackathon requirements.
Developed by contributors for the Open Targets Hackathon 2025.
Thanks to the Open Targets team and all hackathon participants for their collaboration and data infrastructure.
