The applicability domain of EPI Suite™ for screening phytotoxins for potential to contaminate source water for drinking

Abstract Background Toxins produced by plants constitute a potential threat to water supplies in Europe, but have not been widely considered systematic risk assessments. One way begin address risks of phytotoxins is conduct screening-level assessments known for their contaminate source drinking due persistence (P) and mobility (M). Chemical properties relevant such an assessment (octanol–water partition coefficient K OW biodegradation half-life) can be estimated from the structure with quantitative structure–property relationship (QSPR) models found United States Environmental Protection Agency’s Estimation Program Interface (EPI Suite ™ ) software, predictions must critically since these developed using data anthropogenic chemicals many could lie outside applicability domain. Results We analyzed two EPI models—KOWWIN BIOWIN5—by evaluating quality property validation sets as function Euclidean distances d E centroid descriptor space models’ training sets. identified model-specific domain boundaries local maxima plots difference between root mean square error (∆RMSE) modeled values set compounds within defined continuum possible boundaries. And, we also evaluated four generic that suggested literature. The ∆RMSE inside had positive all one considered, indicating were better predicted. With our proposed boundaries, 21% 1586 Switzerland out KOWWIN, 46% BIOWIN5. Conclusions Estimates Log half-life QSPR should viewed extrapolations are subject unquantified potentially large errors. Phytotoxins prioritized measurements basis expand support hazard identification management Europe.