NeoDTI: neural integration of neighbor information from a heterogeneous network for discovering new drug–target interactions

• DOI: 10.1093/bioinformatics/bty543
• Publication Date: 2018-06-29T12:11:58Z
• Authors: Fangping Wan; Lixiang Hong; An Xiao; Tao Jiang; Jianyang Zeng
• Keywords: Information integration

Abstract Motivation Accurately predicting drug–target interactions (DTIs) in silico can guide the drug discovery process and thus facilitate development. Computational approaches for DTI prediction that adopt systems biology perspective generally exploit rationale properties of drugs targets be characterized by their functional roles biological networks. Results Inspired recent advance information passing aggregation techniques generalize convolution neural networks to mine large-scale graph data greatly improve performance many network-related tasks, we develop a new nonlinear end-to-end learning model, called NeoDTI, integrates diverse from heterogeneous network automatically learns topology-preserving representations prediction. The substantial improvement over other state-of-the-art methods as well several novel predicted DTIs with evidence supports previous studies have demonstrated superior predictive power NeoDTI. In addition, NeoDTI is robust against wide range choices hyperparameters ready integrate more target related (e.g. compound–protein binding affinity data). All these results suggest offer powerful tool development repositioning. Availability implementation source code used are available at: https://github.com/FangpingWan/NeoDTI. Supplementary at Bioinformatics online.