Machine learning methods, databases and tools for drug combination prediction
Lianlian Wu,Yuqi Wen,Dongjin Leng,Qinglong Zhang,Chong Dai,Zhongming Wang,Ziqi Liu,Bowei Yan,Yixin Zhang,Jing Wang,Song He,Xiaochen Bo
DOI: https://doi.org/10.1093/bib/bbab355
IF: 9.5
2021-09-02
Briefings in Bioinformatics
Abstract:Abstract Combination therapy has shown an obvious efficacy on complex diseases and can greatly reduce the development of drug resistance. However, even with high-throughput screens, experimental methods are insufficient to explore novel drug combinations. In order to reduce the search space of drug combinations, there is an urgent need to develop more efficient computational methods to predict novel drug combinations. In recent decades, more and more machine learning (ML) algorithms have been applied to improve the predictive performance. The object of this study is to introduce and discuss the recent applications of ML methods and the widely used databases in drug combination prediction. In this study, we first describe the concept and controversy of synergism between drug combinations. Then, we investigate various publicly available data resources and tools for prediction tasks. Next, ML methods including classic ML and deep learning methods applied in drug combination prediction are introduced. Finally, we summarize the challenges to ML methods in prediction tasks and provide a discussion on future work.
biochemical research methods,mathematical & computational biology