Abstract:Compared with monotherapy, drug combination therapy has demonstrated more effective and powerful therapeutic effects in cancer treatment. However, due to the large size of the combinatorial space, experimental identification of synergistic drug combinations is cost- and time-consuming, even with high-throughput screens. With the accumulation of various types of drug informatics data, artificial intelligence techniques have become applicable for the prediction of synergistic drug combinations, thereby reducing the experimental workload and accelerating the discovery of combination therapy. Currently, the emerging graph neural network (GNN) has been applied to drug combination prediction and brings new insight on the discovery of combination therapy. In this study, we proposed Drug-Target Graph based Recurrent Network (DTGRN) to predict synergistic drug combinations in various cancer cell lines. Specifically, we proposed a Graph Recurrent Network-based encoder to learn a low-dimensional representation of features for each node and then predicted the synergy score of novel drug combinations based on these representations via a cell line-specific projection matrix. In DTGRN, gating mechanism and cell-line-level states are employed to ensure each node representation more comprehensively absorbs local and global information and potential association information. In the application of the cell line-specific drug-target heterogeneous graph, DTGRN significantly outperforms the state-of-the-arts and several classical machine learning algorithms. More prominently, the novel drug combinations predicted by DTGRN have been validated by computational analysis and in vitro experiments, confirming the high predictive performance of DTGRN. These drug combinations could be further evaluated as potential combination therapy strategy. In summary, these results indicate that DTGRN can serve as a powerful tool for integrating heterogeneous graph information to predict drug combinations and provide a promising way to discover novel synergistic drug combination for cancer treatment.

MDMNI-DGD: A novel graph neural network approach for druggable gene discovery based on the integration of multi-omics data and the multi-view network

From multi-omics data to the cancer druggable gene discovery: a novel machine learning-based approach

Enhancing Cancer Driver Gene Prediction by Protein-Protein Interaction Network

AGMI - Attention-Guided Multi-omics Integration for Drug Response Prediction with Graph Neural Networks.

drGAT: Attention-Guided Gene Assessment of Drug Response Utilizing a Drug-Cell-Gene Heterogeneous Network

MODIG: Integrating Multi-Omics and Multi-Dimensional Gene Network for Cancer Driver Gene Identification Based on Graph Attention Network Model.

DGMP: Identifying Cancer Driver Genes by Jointing DGCN and MLP from Multi-omics Genomic Data

Multiplex Networks and Pan-Cancer Multiomics-Based Driver Gene Identification Using Graph Neural Networks

HMM-GDAN: Hybrid Multi-View and Multi-Scale Graph Duplex-Attention Networks for Drug Response Prediction in Cancer.

Integration of multi-source gene interaction networks and omics data with graph attention networks to identify novel disease genes

DMHGNN: Double Multi-View Heterogeneous Graph Neural Network Framework for Drug-Target Interaction Prediction

Drug-Target Graph Based Recurrent Network for Drug Combination Prediction

Interpretable Dynamic Directed Graph Convolutional Network for Multi-Relational Prediction of Missense Mutation and Drug Response

MDGAE-DTI: Drug-Target Interactions Prediction Based on Multi-information Integration and Graph Auto-Encoder

Dual-Channel Heterogeneous Graph Neural Network for Predicting Microrna-Mediated Drug Sensitivity.

Gndd: A Graph Neural Network-Based Method For Drug-Disease Association Prediction

DMGL-MDA: A dual-modal graph learning method for microbe-drug association prediction

Probabilistic graph-based model uncovers previously unseen druggable vulnerabilities in major solid cancers

HOGMMNC: A higher order graph matching with multiple network constraints model for gene-drug regulatory modules identification.

Multi-layer graph attention neural networks for accurate drug-target interaction mapping

MCDHGN: heterogeneous network-based cancer driver gene prediction and interpretability analysis