Abstract
ABSTRACT
Hexafluoropropylene oxide‐dimer acid (GenX), a prominent alternative to legacy per‐ and polyfluoroalkyl substances (PFAS), poses a significant challenge to traditional linear risk assessment models due to its ability to induce hormesis—a biphasic “low‐dose stimulation, high‐dose inhibition” response. This study established an interpretable machine learning (ML) framework to identify and predict GenX‐induced non‐monotonic dose–response (NMDR) relationships in zebrafish (
Danio rerio
). By integrating 263 independent experimental records across 11 biological categories, we benchmarked 6 mL paradigms, with the CatBoost‐based Master Model demonstrating exceptional predictive fidelity (AUC = 0.977; Accuracy = 91.25%). Mechanistic interpretation via SHAP (SHapley Additive exPlanations) analysis revealed that “Effect Measurement” and “Biological Category” are the primary determinants of hormetic responses. Specifically, the immune, endocrine, and reproductive systems, alongside neurobehavioral endpoints, exhibited the highest sensitivity to low‐dose stimulation. Structure–activity relationship (SAR) analysis identified lipophilicity and surface distribution descriptors (
MolLogP, SlogP_VSA3)
, electrostatic state and surface area descriptors (
VSA_EState
series), and topological connectivity index (
Chi1v
) as core structural drivers of hormesis. These findings suggest that GenX‐induced hormesis stems from a strategic metabolic trade‐off, where optimized electronic distribution and lipid‐mediated transport trigger adaptive overcompensation via signaling pathways like oxidative stress. This research provides a quantitative tool and a robust scientific basis for the ecological risk assessment of emerging PFAS alternatives, shifting the paradigm from empirical data fitting to mechanistic predictive generalization.