人工智能在药物再利用治疗新型冠状病毒感染研究中的应用及问题分析  

Application and analysis of problems of artificial intelligence in drug repurposing for Corona Virus Disease 2019 (COVID-19)

作　　者：单士喆 文博 乔天慈 单光存 SHAN Shizhe;WEN Bo;QIAO Tianci;SHAN Guangcun(Guang′anmen Hospital,China Academy of Chinese Medical Sciences,Beijing 100032,China;Hong Kong Metropolitan University,Hong Kong 999077,China;Shanghai Yueyang Integrated TCM and Western Medicine Hospital Affiliated to Shanghai University of TCM,Shanghai 200437,China;School of Instrument Science and Optoelectronic Engineering,Beihang University/Beijing Advanced Innovation Center for Big Data-Based Precision Medicine,Beijing 100191,China)

机构地区：[1]中国中医科学院广安门医院,北京100032 [2]香港都会大学,中国香港999077 [3]上海中医药大学附属岳阳中西医结合医院,上海200437 [4]北京航空航天大学仪器科学与光电工程学院/北京大数据与精准医疗高精尖创新中心,北京100191

出　　处：《中国药理学与毒理学杂志》2024年第4期294-303,共10页Chinese Journal of Pharmacology and Toxicology

基　　金：中央高校基本科研业务费(理工医科类)(YWF-22-BJ-J-313);北京市自然基金-昌平创新联合基金(L234003)。

摘　　要：近年来,为应对新型冠状病毒感染(COVID-19)的暴发,药物再利用成为寻找COVID-19治疗药物的有效策略。人工智能(AI)能够快速计算筛选大量药物数据库以获取候选药物,在药物再利用领域得到广泛应用。根据算法设计原理,AI应用于药物再利用治疗COVID-19研究的方法可分为3类:①基于网络的模型,强调药物与疾病间关联性的识别,以揭示药物的潜在治疗机制;②基于结构的方法,通过药物和靶点间结构相互作用的分析实现精确筛选;③机器学习/深度学习方法,利用复杂非线性数据的多维度处理进行候选药物预测。尽管AI在药物再利用中发挥了重要作用,但数据的质量和数量对AI计算结果影响显著;实验研究无法全面模拟人体的复杂生理环境,从而可能限制候选药物在非临床研究阶段的精确验证;而且针对原始适应证的药物优化可能影响候选药物在治疗COVID-19中的有效性,治疗时机和个体差异也可能对临床效果产生影响。本文对AI在药物再利用治疗COVID-19研究中的应用和挑战进行综述,以期为将AI技术进一步应用于治疗COVID-19药物研究提供参考。In recent years, drug repurposing has emerged as an effective strategy for identifying potential treatments for the outbreak of the Corona Virus Disease 2019 (COVID-19). Artificial intelli⁃ gence (AI) has been widely employed in the field of drug repurposing, enabling rapid computation and screening of extensive drug databases. Based on different algorithm design principles, AI methodolo⁃ gies for drug repurposing in the context of COVID-19 can be categorized into three types:① network-based models, which emphasize the identification of associations between drugs and diseases to reveal potential therapeutic mechanisms;② structure-based methods, which employ the analysis of structural interactions between drugs and targets for precise screening;and ③ machine learning/deep learning approaches, which utilize multidimensional processing of complex nonlinear data for candidate drug prediction. Despite the significant role of AI in drug repurposing, the quality and quantity of data have a notable impact on the computational results of AI. Experimental studies alone cannot fully simulate the complex physiological environment of the human body, which may limit the precise validation of candi⁃ date drugs in the preclinical stage. Optimization of drugs originally indicated for other conditions may also affect the effectiveness of candidate drugs for COVID-19. Moreover, treatment timing and individual differences may influence clinical outcomes. This review provides an overview of the application and challenges of AI in the field of drug repurposing for COVID-19 in order to provide reference for wider use of AI technology in COVID-19 treatment.

关 键 词：人工智能 药物再利用 新型冠状病毒感染 

分 类 号：R978.7[医药卫生—药品] R965[医药卫生—药学]