Facing small and biased data dilemma in drug discovery with enhanced federated learning approaches

doi:10.1007/s11427-021-1946-0

	Facing small and biased data dilemma in drug discovery with enhanced federated learning approaches
	Xiong, Zhaoping 2,3,4,5; Cheng, Ziqiang 2,3,6; Lin, Xinyuan 1; Xu, Chi 1; Liu, Xiaohong 2,3,4; Wang, Dingyan 4,5; Luo, Xiaomin 4; Zhang, Yong 1; Jiang, Hualiang 2,3,4; Qiao, Nan 1
刊名	SCIENCE CHINA-LIFE SCIENCES
	2021-07-26
页码	11
关键词	federated learning drug discovery FedAMP Non-IID data
ISSN号	1674-7305
DOI	10.1007/s11427-021-1946-0
通讯作者	Jiang, Hualiang(hljiang@simm.ac.cn) ; Qiao, Nan(qiaonan3@huawei.com) ; Zheng, Mingyue(myzheng@simm.ac.cn)
英文摘要	Artificial intelligence (AI) models usually require large amounts of high-quality training data, which is in striking contrast to the situation of small and biased data faced by current drug discovery pipelines. The concept of federated learning has been proposed to utilize distributed data from different sources without leaking sensitive information of the data. This emerging decentralized machine learning paradigm is expected to dramatically improve the success rate of AI-powered drug discovery. Here, we simulated the federated learning process with different property and activity datasets from different sources, among which overlapping molecules with high or low biases exist in the recorded values. Beyond the benefit of gaining more data, we also demonstrated that federated training has a regularization effect superior to centralized training on the pooled datasets with high biases. Moreover, different network architectures for clients and aggregation algorithms for coordinators have been compared on the performance of federated learning, where personalized federated learning shows promising results. Our work demonstrates the applicability of federated learning in predicting drug-related properties and highlights its promising role in addressing the small and biased data dilemma in drug discovery.
资助项目	Shanghai Municipal Science and Technology Major Project ; National Natural Science Foundation of China[81773634] ; National Science and Technology Major Project of the Ministry of Science and Technology of China[2018ZX09711002] ; Strategic Priority Research Program of the Chinese Academy of Sciences[XDA12050201] ; Strategic Priority Research Program of the Chinese Academy of Sciences[XDA12020368]
WOS关键词	AQUEOUS SOLUBILITY ; HERG ; PREDICTION ; MODEL ; CLASSIFICATION ; RESISTANCE ; INHIBITORS ; MECHANISM ; COMPOUND ; SECURE
WOS研究方向	Life Sciences & Biomedicine - Other Topics
语种	英语
出版者	SCIENCE PRESS
WOS记录号	WOS:000678456300002
内容类型	期刊论文
源URL	[http://119.78.100.183/handle/2S10ELR8/298174]
专题	中国科学院上海药物研究所
通讯作者	Jiang, Hualiang; Qiao, Nan; Zheng, Mingyue
作者单位	1.Huawei Technol Co Ltd, Lab Hlth Intelligence, Shenzhen 518100, Peoples R China 2.ShanghaiTech Univ, Shanghai Inst Adv Immunochem Studies, Shanghai 200031, Peoples R China 3.ShanghaiTech Univ, Sch Life Sci & Technol, Shanghai 200031, Peoples R China 4.Chinese Acad Sci, Shanghai Inst Mat Med, Drug Discovery & Design Ctr, State Key Lab Drug Res, Shanghai 201203, Peoples R China 5.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 6.Univ Sci & Technol China, Sch Informat Sci & Technol, Hefei 230000, Peoples R China
推荐引用方式 GB/T 7714	Xiong, Zhaoping,Cheng, Ziqiang,Lin, Xinyuan,et al. Facing small and biased data dilemma in drug discovery with enhanced federated learning approaches[J]. SCIENCE CHINA-LIFE SCIENCES,2021:11.
APA	Xiong, Zhaoping.,Cheng, Ziqiang.,Lin, Xinyuan.,Xu, Chi.,Liu, Xiaohong.,...&Zheng, Mingyue.(2021).Facing small and biased data dilemma in drug discovery with enhanced federated learning approaches.SCIENCE CHINA-LIFE SCIENCES,11.
MLA	Xiong, Zhaoping,et al."Facing small and biased data dilemma in drug discovery with enhanced federated learning approaches".SCIENCE CHINA-LIFE SCIENCES (2021):11.