MHTAN-DTI

This repository provides a reference implementation of MHTAN-DTI as described in this paper:

MHTAN-DTI: Metapath-based Hierarchical Transformer and Attention Network for Drug-Target Interaction Prediction
Ran Zhang, Zhanjie Wang, Xuezhi Wang, Zhen Meng, Wenjuan Cui.

Dependencies

Recent versions of the following packages for Python 3.7 are required:

• PyTorch 1.12.1
• DGL 0.6.1
• NetworkX 2.6.3
• NumPy 1.21.6
• Scikit-learn 1.0.2

Dataset

The preprocessed Luo et al. dataset is available at -> https://cstr.cn/31253.11.sciencedb.01726

• Luo et al. dataset contains 708 drugs, 1512 proteins, 5603 diseases and 4192 side effects.

Usage

1. Create checkpoint/ and DTI_data/ directories
2. Download the dataset from the section above to DTI_data/
3. Execute the following command from the project home directory to train and test:
   • python run_DTI.py