Read-across-driven binary classification for the developmental and reproductive toxicity of organic compounds tested...

Developmental and reproductive toxicity (DART) refers to the adverse effects on sexual function, fertility, and the development of offspring resulting from exposure to toxic substances or chemicals, which may occur at various stages of the reproductive cycle. In response to the increasing volume of chemicals, regulatory bodies advocate for implementing various new approach methodologies (NAMs) as alternatives to animal testing, enabling rapid assessments of the toxic potential of numerous chemical substances. In this study, in silico methodologies were utilized to assess the DART properties of various industrial chemicals. We employed a Read-Across (RA)-based binary classification approach to evaluate the DART potential of these chemicals. The data for the binary classification have been compiled from two distinct sources: eChemPortal (https://www.echemportal.org/echemportal/) and the National Institute of Health Sciences (NIHS) databases. The information gathered from these sources encompasses two types of toxicity data: No Observed Adverse Effect Level (NOAEL) and Low Observed Adverse Effect Level (LOAEL) tested as per the Organisation for Economic Co-operation and Development Test Guidelines 421 and 422, adopting the principles of Good Laboratory Practice (GLP). The data were utilized separately for safety assessment through a binary classification-based read-across prediction, demonstrating commendable classification capabilities for new chemicals (Accuracy_test ~0.700).