This repository is the Pytorch implementation of our ISMB 2020 paper 'Toward heterogeneous information fusion: bipartite graph convolutional networks for in silico drug repurposing'.

BiFusion is a bipartite graph convolution network model for drug repurposing through heterogeneous information fusion. Our approach combines insights of multi-scale pharmaceutical information by constructing a multi-relational graph of drug–protein, disease-protein and protein–protein interactions.

Contains the code for dataloader.

Contains the code for model components.

Contains the code for BiFusion model

Run the code as following:

$ python main.py

BiFusion is tested to work under Python 3.6. The required dependencies are:

PyTorch==1.2.0  
PyTorch-Geometric==1.4.1  
numpy==1.16.0  
scikit-learn==0.21.3

If this repository is useful for your research, please consider citing this paper:

@article{wang2020toward,
  title={Toward heterogeneous information fusion: bipartite graph convolutional networks for in silico drug repurposing},
  author={Wang, Zichen and Zhou, Mu and Arnold, Corey},
  journal={Bioinformatics},
  volume={36},
  number={Supplement\_1},
  pages={i525--i533},
  year={2020},
  publisher={Oxford University Press}
}

Please send any questions you might have about this repository to zcwang0702@ucla.edu