Teaching computational toxicology as new approach methodologies in the era of next-generation risk assessment

The global initiative toward non-animal chemical testing methods has recently gained significant momentum, requiring developing new approach methodologies (NAMs) as transformative solutions for next-generation risk assessment (NGRA). NAMs refer to any in chemico, in silico, and in vitro technology, methodology, or combination that do not rely on animals. A range of computational toxicology methods are becoming critical components within the NGRA framework across multiple testing tiers. These approaches are needed to support hazard characterization and prioritization for further testing, calculate points of departure, identify modes of action, and bridge the data gaps between in vitro assays and in vivo exposures as well as apical health endpoints. The core technologies of computational toxicology include (i) quantitative structure activity relationship, (ii) benchmark dose modeling, (iii) bioinformatic analysis of omics data, (iv) physiologically based kinetic modeling, (v) quantitative adverse outcome pathway modeling, (vi) molecular docking and dynamics, and (vii) more recently, machine learning and artificial intelligence tools. The NAM community, in academia, industry, and regulatory agencies, are facing increasing demand for training current and future professionals to master these skills. In this article, we described our perspective, experiences, and challenges in disseminating these computational toxicology techniques. Unlike established disciplines, the curriculum of computational toxicology and associated software tools are constantly evolving, requiring instructors to adapt in a timely manner. Challenges abound, we are embraced with an unprecedented opportunity that educating next-generation toxicologists is a critical effort of this revolutionary movement of chemical toxicity testing and safety assessment.