fig1
From: Artificial intelligence for toxicokinetic parameterization in environmental exposure assessment
Figure 1. General workflow of AI/ML-based ADME prediction, including (1) chemical data curation; (2) molecular feature generation; (3) AI/ML model development; and (4) rigorous model validation [Created in BioRender. Zhang, Z. (2026) https://BioRender.com/86v7eo1]. AL/ML: Artificial intelligence and machine learning; ADME: absorption, distribution, metabolism, and excretion; SMILES: simplified molecular-input line-entry system; MW: molecular weight; TPSA: topological polar surface area; ECFP: extended-connectivity fingerprint; MACCS: molecular access system; GNN: graph neural network; OECD: Organisation for Economic Co-operation and Development.
