Best Practices for the Development and Application of Bioinformatics Approaches for Species Extrapolation in Chemical Safety Assessment

Abstract It is logistically impossible to perform toxicity testing of all chemicals in all species. Therefore, chemical safety evaluations must find ways to make inferences about toxicity in untested species based on observations from model species. Development of bioinformatics tools that create scientific lines of evidence for extrapolation of toxicity knowledge across species are gaining traction in decision-making surrounding chemical safety. These approaches take advantage of computational power and allow for incorporation of large-scale omics data to infer conservation of biological pathways across species to enable predictions of chemical susceptibility and further understand potential toxicity. As comparative toxicity data from such approaches are integrated into regulatory decisions to support surrogacy challenges it is important that the tools being developed for and applied in regulation meet necessary standards identified by hazard and risk assessment practitioners. The International Consortium to Advance Cross Species Extrapolation in Regulation (ICACSER), which consists of researchers, tool developers, and regulatory decision-makers seeks to highlight the foundational role of bioinformatics for enhanced scientific evidence to inform chemical safety decisions across species. To articulate essential standards and define best practices in development and application of bioinformatics approaches specific to this context, three existing tools are examined, and their key attributes are highlighted and connections to the adverse outcome pathway framework are discussed. These practices are intended to act as a guide for current and future tool development.