Assessing comparable bioconcentration potentials for nanoparticles in aquatic organisms via combined utilization of machine learning and toxicokinetic models

Abstract The toxicokinetic (TK) model‐derived kinetic bioconcentration factor (BCF k ) provides a quantitatively comparable index to estimate the bioaccumulation potential of nanoparticles (NPs) that barely reach thermodynamic equilibrium in aquatic organisms, but experimental data are limited for various NPs. In present study, machine learning model was applied offer reliable silico predictions dynamic body burden diverse NPs derive corresponding parameters TK model. developed eXtreme Gradient Boosting‐derived (XGB‐TK) predict BCF results broad range metallic or carbonaceous NPs, with an appreciable prediction R 2 0.96. values were predicted based on random combination selected variable features, revealing their showed overall negative correlation NP density organism size. By applying importance analysis and partial dependence plots, size revealed be top essential features impact potential. conjunctively used XGB‐TK enabled prior comparison straightforward derivation dependency which could also guide mechanism exploration condition formulation.