The Next Generation of Machine Learning in DDIs Prediction.

Abstract

Drug-drug interactions may occur when combining two or more drugs may cause some adverse events such as cardiotoxicity, central neurotoxicity, hepatotoxicity, etc. However, a large number of researchers who are proficient in pharmacokinetics and pharmacodynamics have been engaged in drug assays and trying to find out the side effects of all kinds of drug combinations. However, at the same time, the number of new drugs is increasing dramatically, and the drug assay is an expensive and time-consuming process. It is impossible to find all the adverse reactions through drug experiments. Therefore, new attempts have been made in using computational techniques to deal with this problem. In this review, we conduct a review of the literature on applying the computational method for predicting drug-drug interactions. We first briefly introduce the widely used data sets. After that, we elaborate on the existing state-of-art deep learning models for drug-drug interactions prediction. We also discussed the challenges and opportunities of applying the computational method in drug-drug interactions prediction.