The risk assessment of an expanding array of emerging contaminants in aquatic ecosystems and the establishment of water quality criteria rely on species sensitivity distribution (SSD), necessitating ample multi-trophic toxicity data. Computational methods, such as quantitative structure-activity relationship (QSAR), enable the prediction of specific toxicity data, thus mitigating the need for costly experimental testing and exposure risk assessment. In this study, robust QSAR models for four aquatic species (Rana pipiens, Crassostrea virginica, Asellus aquaticus, and Lepomis macrochirus) were developed using leave-one-out (LOO) screening variables and the partial least squares algorithm to predict toxicity data for paraquat, bisphenol A, and carbamazepine. These predicted data can be integrated with experimental data to construct SSD models and derive hazardous concentration for 5 % of species (HC5) for the criterion maximum concentration. The chronic water quality criterion for paraquat, bisphenol A, and carbamazepine were determined at 6.7, 11.1, and 3.5 μg/L, respectively. The QSAR-SSD approach presents a viable and cost-effective method for deriving water quality criteria for other emerging contaminants.
Keywords: DRAGON software; Emerging contaminants; Pesticide; Risk assessment; Toxicity.
Copyright © 2024 Elsevier B.V. All rights reserved.