Knowledge Graph Neural Network

This is our implementation for the paper

Xuan Lin, Zhe Quan, Zhi-Jie Wang, Tengfei Ma and Xiangxiang Zeng. KGNN: Knowledge Graph Neural Network for Drug-Drug Interaction Prediction. IJCAI' 20 accepted.

Figure 1 shows the overview of KGNN. It takes the parsed DDI matrix and knowledge graph obtained from preprocessing of dataset as the input. It outputs the interaction value for the drug-drug pair.

Requirement

To run the code, you need the following dependencies:

• Python == 3.6.6
• Keras == 2.3.0
• Tensorflow == 1.13.1
• scikit-learn == 0.22

Installation

You can create a virtual environment using conda.

conda create -n kgnn python=3.6.6  
source activate kgnn  
git clone https://github.com/xzenglab/KGNN.git  
cd KGNN  
pip install -r requirement.txt  

Dataset

We just provide the preprocessed KG from KEGG-drug dataset owing to the size limited. And you can directly download the original DrugBank dataset (V5.1.4). Note that the construction of KG please refer to Bio2RDF tool in detail.

Usage

python run.py

Citation

@inproceedings{ijcai2020-380,  
  title     = {KGNN: Knowledge Graph Neural Network for Drug-Drug Interaction Prediction},  
  author    = {Lin, Xuan and Quan, Zhe and Wang, Zhi-Jie and Ma, Tengfei and Zeng, Xiangxiang},  
  booktitle = {Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, {IJCAI-20}},  
  publisher = {International Joint Conferences on Artificial Intelligence Organization},               
  editor    = {Christian Bessiere},    	
  pages     = {2739--2745},    
  year      = {2020},  
  month     = {7},  
  note      = {Main track},  
  doi       = {10.24963/ijcai.2020/380},  
  url       = {https://doi.org/10.24963/ijcai.2020/380},  
}  

For any clarification, comments, or suggestions please create an issue or contact Jacklin.