HIT (Hypergraph Interaction Transformer)

AI-driven therapeutic gene target prediction

An Explainable AI model for fast and precise identification of therapeutic gene candidates by integrating complex disease-gene relationships and ontology information.

🚀 Key Features

✅ Hypergraph-based modeling: Captures many-to-many relationships between diseases and genes.

✅ Ontology integration: Utilizes disease and gene ontology information for enhanced representation.

✅ Explainable AI: Provides interpretable insights into model decision-making.

✅ Scalable implementation: Built on PyTorch, designed for large-scale biomedical datasets.

📂 Project Structure

HIT/
├── datasets/          # Original datasets
├── models/            # Model implementation
├── exp.py             # Main execution script
├── dataset.py         # Dataset construction script
├── trainer.py         # Model trainer
├── utils.py           # utils
├── requirements.txt   # Python dependencies
└── README.md

⚙️ Installation

1. Clone this repository:

$ git clone https://github.com/tigerkey10/HIT.git
$ cd HIT

2. Install required dependencies:

$ pip install -r requirements.txt

We used NVIDIA RTX A6000 GPU with CUDA version 11.7.

▶️ Usage

Run the model:

$ python exp.py 

Run with custom arguments:

# Example
$ python exp.py --epochs 50 --lr 1e-3

🌐 Webserver Access

You can access the deployed HIT webserver interface via the link below:

🔗 HIT Webserver

Webserver interface

📖User Guide

🧪Predictor

💻Computation Example

📮Feedback

ℹ️About

📖 Reference

Kim, Kibeom, et al. "Therapeutic gene target prediction using novel deep hypergraph representation learning." Briefings in Bioinformatics 26.1 (Jan 2025). 🔗 Paper

💡 If you use this code for research, please cite the above paper.

📜 License

This project is licensed under the MIT License.