Deep Learning Prediction of Adverse Drug Reactions Using Open TG-GATEs and FAERS Databases

With the advancements in Artificial intelligence (AI) and the accumulation of healthrelated big data, it has become increasingly feasible and commonplace to leverage machine learning technologies to analyze clinical and omics metadata to assess the possibility of adverse drug reactions or events (ADRs) in the course of drug discovery. Here, we have described a novel approach that combined drug-induced gene expression profile from Open TG-GATEs (Toxicogenomics Project-Genomics Assisted Toxicity Evaluation Systems) and ADR occurrence information from FAERS (FDA [Food and Drug Administration] Adverse Events Reporting System) database to predict the likelihood of ADRs. We generated a total of 14 models using Deep Neural Networks (DNN) to predict different ADRs; in the validation tests, our models achieved a mean accuracy of 85.71%, indicating that our approach successfully and consistently predicted ADRs for a wide range of drugs. As an example, we have described the ADR model in the context of Duodenal ulcer. We believe that our models will help predict the likelihood of ADRs while testing novel pharmaceutical compounds, and will be useful for researchers in drug discovery.