Network-based method to infer the contributions of proteins to the etiology of drug side effects

作     者：Li, Rui Chen, Ting Li, Shao 

作者机构：MOE Key Laboratory of Bioinformatics and Bioinformatics Division TNLIST Department of Automation Tsinghua University Beijing China Program in Computational Biology and Bioinformatics University of Southern California Los Angeles USA 

出 版 物：《Frontiers of Electrical and Electronic Engineering in China》 (中国电气与电子工程前沿（英文版）)

年 卷 期：2015年第10卷第3期

页      面：124-134页

核心收录：

学科分类：0710[理学-生物学] 1007[医学-药学(可授医学、理学学位)] 071010[理学-生物化学与分子生物学] 081704[工学-应用化学] 07[理学] 08[工学] 0817[工学-化学工程与技术] 10[医学] 

基　　金：This work is supported by National Natural Science Foundation of China (Nos. 81225025 and 91229201)  and Tsinghua National Laboratory of Information Science and Technology (TNLIST) Big Data Grant 

主　　题：network pharmacology drug targets side effects triptolide 

摘      要：Studying the molecular mechanisms that underlie the relationship between drugs and the side effects they produce is critical for drug discovery and drug development. Currently, however, computational methods are still unavailable to assess drug-protein interactions with the aim of globally inferring the contributions of various classes of proteins toward the etiology of side effects. In this work, we integrated data reflecting drug-side effect relationships, drug- target relationships, and protein-protein interactions to develop a novel network-based probabilistic model, SidePro, to evaluate the contributions of proteins toward the etiology of side effects. For a given side effect, the method applies an expectation--maximization algorithm and a diffusion kernel-based approach to estimate each protein＇s contribution. We applied this method to a wide range of side effects and validated the results using cross-validation and records from the Side Effect Resource database. We also studied a specific side effect, nephrotoxicity, which is known to be associated with the irrational use of the Chinese herbal compound triptolide, a diterpenoid epoxide in the Thunder of God Vine, Tripterygium wilfordii （Lei-Gong-Teng）. Using triptolide as an example, we scored the target proteins of triptolide using our model and investigated the high-scoring proteins and their related biological processes. The results demonstrated that our model could differentiate between the potential side effect targets and therapeutic targets of triptolide. Overall, the proposed model could accurately pinpoint the molecular mechanisms of drug side effects, thus making contribution to safe and effective drug development.