Application of combined QSAR-ICE models in calculation of hazardous concentrations for linear alkylbenzene sulfonate

Abstract

Linear alkylbenzene sulfonate (LAS) is a widely used anionic surfactant that exists as a mixture of various homologous structures in water environment. In the calculation of hazardous concentrations of LAS, cross-taxonomies toxicity estimation was often used instead of species-level-specific estimation for the normalization of toxicity data, which led to substantial uncertainties. In this study, combined quantitative structure–activity relationship (QSAR) and interspecific relationship estimation (ICE) models were developed to normalize the alkyl chain length of toxicity data and calculate the 5th percentile hazard concentrations (HC5s) of LAS. Using seven acute QSAR models based on measured data and 29 acute QSAR-ICE models derived from them, the acute HC5s of LAS were calculated as 2.09–3.67 mg/L. Furthermore, species- and family-level-specific QSAR and QSAR-ICE models were used to calculate chronic HC5s (0.19–0.38 mg/L). Additionally, the sensitivity of biological toxicity to the hydrophobicity of LAS, represented by the slope of the QSAR models, had a significant correlation with the taxa of the species. Further risk assessment based on chronic HC5s showed potential ecological risks in the Dianchi Lake basin and Haihe River basin in China, which should cause concern.