Mold2 Descriptors Facilitate Development of Machine Learning and Deep Learning Models for Predicting Toxicity of Chemicals

Numerical description of chemical structures is necessary for development of machine learning and deep learning models for predicting the potential toxicity of chemicals. Mold2 is a software tool developed in C++ for fast calculating molecular descriptors from two-dimensional structures. Mold2 descriptors contain rich information and can be used to build high-performance models in computational toxicology. Multiple studies have compared Mold2 descriptors with other descriptors and fingerprints in machine learning and deep learning models for predicting the toxicity of chemicals. These studies have demonstrated that models built with Mold2 descriptors outperform models developed with other descriptors and fingerprints.