MRHGNN: Enhanced Multimodal Relational Hypergraph Neural Network for Synergistic Drug Combination Forecasting

Overview

This repository contains Python codes and datasets necessary to run the MRHGNN model. MRHGNN is a novel framework for predicting synergistic drug combinations. Specifically, we design a dual-channel architecture to capture the physicochemical attributes of drugs and their interactive synergies, thereby facilitating the generation of multimodal drug representations. To obtain comprehensive representations of drugs, we utilize an attention mechanism to explore complementarity among multimodal drug embeddings. Additionally, the unified framework jointly learns the primary and self-supervised learning tasks, fostering a robust predictive capability. Please take a look at our paper for more details on the method.

Install the Environment

OS Requirements

The package development version is tested on Linux (Ubuntu 20.04) operating systems with CUDA 11.7.

Python Dependencies

MRHGNN is tested under Python == 3.10.14.

We provide a txt file containing the necessary packages for MRHGNN. All the required basic packages can be installed using the following command:

pip install -r requirements.txt

Usage

To train and evaluate the model, you could run the following command.

• O'Neil Dataset

python MRHGNN.py --dataset 'ONEIL' --threshold 30 --alpha 0.2 --mask_ratio 0.2 --learning_rate 1e-2 --weight_decay 1e-4 --epochs 2000

• NCI-ALMANAC Dataset

python MRHGNN.py --dataset 'ALMANAC' --threshold 10 --alpha 0.3 --mask_ratio 0.3 --learning_rate 1e-2 --weight_decay 1e-4 --epochs 2000

Citation

Please kindly cite this paper if you find it useful for your research. Thanks!

@article{10950093,
  author={Chen, Mengjie and Zhang, Ming and Yan, Guiying and Wang, Guanghui and Qu, Cunquan},
  journal={IEEE Transactions on Neural Networks and Learning Systems}, 
  title={MRHGNN: Enhanced Multimodal Relational Hypergraph Neural Network for Synergistic Drug Combination Forecasting}, 
  year={2025},
  volume={},
  number={},
  pages={1-13},
  doi={10.1109/TNNLS.2025.3553385}}