Bridging the Gap Between Nanotoxicological Data and the Critical Structure-Activity Relationships

The rapid development of nanotoxicology research has led to an exponential increase in data being accumulated and the urgent need of developing computational methods for extracting and processing critical nanostructure-activity relationships from large data sets. During the past several years, artificial intelligence, especially deep learning, has emerged as a powerful method to mine useful information from complex big data, which has been widely used for face recognition, autonomous driving, and medical diagnosis. Inspired by these successes, researchers have successfully applied these technologies in the areas of toxicology. Compared with small molecules, the complexity and diversity of nanostructures lead to many challenges in the application of artificial intelligence to nanotoxicology. Here, we focus on the current status of nanomaterial databases and the applications of artificial intelligence to nanotoxicology research and provide future perspectives on developments that are likely or need to occur in the near future that allow big data and artificial intelligence to make a deeper contribution to nanosafety.