(418g) Toxicity Prediction of Small Molecules Using Artificial Intelligence

Usually, in-vitro and in-vivo tests are carried out to investigate the safety of chemical products or drugs. However, these tests could be time-consuming, expensive, and not directly translatable from animal models to human models. With the aid of computational methods, drug discovery, and development can be performed at a higher speed at a lower cost.

The Tox21 dataset comprises 12,000 environmental chemicals and drugs which were measured for 12 different toxic effects by specially designed assays. In this study, we used low-code Machine Learning and Deep Learning approaches with MATLAB to predict the toxicity of molecules. These low-code and app-based approaches help researchers with minimal coding background to get a head-start to use advanced AI techniques in their research. Although there are recent advancements in artificial intelligence, the speed of adoption and application of these advancements in different domains may depend on the programming fluency of the researchers. We aim to address this challenge by demonstrating user-friendly workflows in MATLAB to enable researchers to use AI in their research.