DeepDrug: A general graph‐based deep learning framework for drug‐drug interactions and drug‐target interactions prediction

• DOI: 10.15302/j-qb-022-0320
• Publication Date: 2023-03-22T08:27:02Z
• Authors: Qijin Yin; Rui Fan; Xusheng Cao; Qiao Liu; Rui Jiang; Wanwen Zeng
• Keywords: Drug repositioning; Drug target; Binary classification

Computational methods for DDIs and DTIs prediction are essential accelerating the drug discovery process. We proposed a novel deep learning method DeepDrug, to tackle these two problems within unified framework. DeepDrug is capable of extracting comprehensive features both target protein, thus demonstrating superior performance in series experiments. The downstream applications show that useful facilitating repositioning discovering potential against specific disease. Background approaches accurate interactions, such as drug‐drug interactions (DDIs) drug‐target (DTIs), highly demanded biochemical researchers. Despite fact many have been developed predict respectively, their success still limited due lack systematic evaluation intrinsic properties embedded corresponding chemical structure. Methods In this paper, we develop framework overcoming above limitation by using residual graph convolutional networks (Res‐GCNs) (CNNs) learn structure‐ sequence‐based representations drugs proteins. Results outperforms state‐of‐the‐art experiments, including binary‐class DDIs, multi‐class/multi‐label classification regression tasks. Furthermore, visualize structural learned Res‐GCN module, which displays compatible accordant patterns categories, providing additional evidence support strong predictive power DeepDrug. Ultimately, apply perform on whole DrugBank database discover candidates SARS‐CoV‐2, where 7 out 10 top‐ranked reported be repurposed potentially treat coronavirus disease 2019 (COVID‐19). Conclusions To sum up, believe an efficient tool provides promising insight understanding underlying mechanism relations.