Application of machine learning in drug side effect prediction: databases, methods, and challenges

• DOI: 10.1007/s11704-024-31063-0
• Publication Date: 2024-11-22T13:58:50Z
• Authors: Haochen Zhao; Jian Zhong; Xiao Liang; Chenliang Xie; Shaokai Wang
• Keywords: 0301 basic medicine; 03 medical and health sciences

Abstract Drug side effects have become paramount concerns in drug safety research, ranking as the fourth leading cause of mortality following cardiovascular diseases, cancer, and infectious diseases. Simultaneously, widespread use multiple prescription over-the-counter medications by many patients their daily lives has heightened occurrence resulting from Drug-Drug Interactions (DDIs). Traditionally, assessments relied on resource-intensive time-consuming laboratory experiments. However, recent advancements bioinformatics rapid evolution artificial intelligence technology led to accumulation extensive biomedical data. Based this foundation, researchers developed diverse machine learning methods for discovering detecting effects. This paper provides a comprehensive overview predicting effects, encompassing entire spectrum biological data acquisition development sophisticated models. The review commences elucidating widely recognized datasets Web servers relevant field effect prediction. Subsequently, study delves into customized binary, multi-class, multi-label classification tasks associated with These are applied variety representative computational models designed identifying induced single drugs DDIs. Finally, outlines challenges encountered using approaches concludes illuminating important future research directions dynamic field.