Computational Modeling in Toxicology: NAMs-based Artificial Intelligence, Mechanistic in Silico Approaches, and Software Platforms

Computational toxicology has evolved from traditional quantitative structure–activity relationship (QSAR) models to sophisticated artificial intelligence (AI)-driven and mechanistic approaches capable of predicting complex toxicological outcomes. Recent developments in deep learning, graph neural networks, toxicokinetic–toxicodynamic (TK/TD) modeling, physiologically based pharmacokinetic (PBPK) models, adverse outcome pathways (AOPs), systems toxicology and molecular simulations have expanded the ability to understand and predict chemical safety. At the same time, the availability of large toxicological databases, high-throughput screening data and omics technologies has accelerated the development of more reliable predictive models. Modern software platforms support automated data curation, model development, validation, interpretation and deployment, enabling reproducible and transparent computational workflows for safety research, industrial applications and regulatory decision-making. The increasing size of chemical space, together with regulatory and societal demands for safer chemicals and reduced animal testing, highlights the need for accurate, scalable and mechanistically interpretable approaches to toxicity assessment.This Research Topic aims to showcase recent advances in computational toxicity modeling across diverse applications, from molecular mechanisms to environmental risk assessment. We particularly encourage studies that integrate multiple computational methodologies, develop reproducible software platforms, and improve the prediction, interpretation and regulatory applicability of toxicity models. Emerging approaches such as multi-task learning, explainable AI and integrated modeling frameworks are especially welcome alongside established computational methodologies. The Research Topic also seeks to promote the development and application of computational tools that support New Approach Methodologies (NAMs), uncertainty assessment and evidence-based regulatory decision-making.To gather further insights into computational approaches for toxicity modeling, we welcome Original Research, Review Articles, Mini Reviews, Methods, Software Articles, Data Reports, Perspective Articles and Brief Research Reports addressing, but not limited to, the following themes:Artificial intelligence and machine learning for toxicity predictionQSAR / QSTR modeling and multi-task or multi-endpoint predictionExplainable and interpretable AIGraph neural networks and other advanced deep-learning approachesPBPK and TK/TD modeling in safety evaluationsSystems Toxicology and adverse outcome pathway (AOPs) modelingMolecular simulations and mechanistic toxicity modelingComputational ecotoxicology and nanotoxicologyToxicity assessment of chemical mixturesNew Approach Methodologies (NAMs) and alternatives to animal testingSoftware platforms for data curation, model development, validation, interpretation, visualization and deploymentCurated toxicological datasets and benchmark studiesReproducible computational workflows and uncertainty analysisRegulatory applications of computational toxicity modelsState-of-the-art methods in selective organ-toxicity with particular emphasis on: Neurotoxicity, Reproductive and Developmental Toxicity Topic editors: Dr. Pravin Ambure, Juan M. Ruso, Arkaprava Banerjee, Michael Gonzalez Durruthy Source: https://www.frontiersin.org/research-topics/84152/computational-modeling-in-toxicology-nams-based-artificial-intelligence-mechanistic-in-silico-approaches-and-software-platforms