人工智能算法用于药物研发的研究进展  

Research progress on artificial intelligence algorithms for drug discovery

作　　者：杨双萌 于江[2] 侯文彬 赵倩 李祎亮[2] YANG Shuang-meng;YU Jiang;HOU Wen-bin;ZHAO Qian;LI Yi-liang(Tianjin University of Traditional Chinese Medicine,Tianjin 301617,China;Institute of Radiation Medicine Chinese Academy of Medical Sciences&Peking Union Medical College,Tianjin 300192,China;Tianjin Nankai District Wangdingdi Hospital,Tianjin 300190,China)

机构地区：[1]天津中医药大学,天津301617 [2]中国医学科学院北京协和医学院放射医学研究所,天津300192 [3]天津市南开区王顶堤医院,天津300190

出　　处：《现代药物与临床》2023年第12期3150-3160,共11页Drugs & Clinic

基　　金：国家自然科学基金项目资助(82104012,82202950,82303681)。

摘　　要：人工智能算法包含机器学习算法和深度学习算法,可应用于药物靶标发现、先导化合物的发现与优化、候选药物的确定、成药性优化。人工智能算法通过丰富的大数据系统学习可以实现模型的建立和高通量虚拟计算,应用于药物研发中能够在一定程度上缩短研发周期、降低投入成本,进而提高研发成功率。对机器学习算法、深度学习算法应用于药物研发中的研究进展进行阐述,以期为人工智能技术与药物研发相结合的进一步发展提供参考。Artificial intelligence algorithms include machine learning algorithms and deep learning algorithms,which can be applied to discover drug targets,discover and optimize lead compounds,determine candidate drugs,and optimize druggability.Artificial intelligence algorithms can achieve model building and high-throughput virtual computing through complex big data system learning.When applied in drug research and development,it can shorten the research and development cycle to a certain extent,reduce input costs,and thereby improve the success rate of research and development.In this article,the research progress of machine learning algorithms and deep learning algorithms applied in drug development was reviewed,in order to provide reference for the further development of the combination of artificial intelligence technology and drug development.

关 键 词：药物研发 人工智能算法 机器学习算法 深度学习算法 药物靶标 先导化合物 候选药物 成药性 

分 类 号：R95[医药卫生—药学]