Ecotoxicological risk assessment of pesticides against different aquatic and terrestrial species: using mechanistic QSTR and iQSTTR modelling approaches to fill the toxicity data gap

Owing to the continuous growth of farming activities, pesticides, a class of toxic and harmful chemicals, are widely present in the environment and thus pose a potential risk to the ecosystem and human health via biomagnification. Therefore, it is significant to conduct a more comprehensive environmental risk assessment towards pesticides in the context of green sustainable chemistry. This study collected the ecotoxicity data of hundreds of pesticides to develop four easily understandable, transferable and reproducible 2D descriptor-based quantitative structure-toxicity relationship (QSTR) models for evaluating the acute toxicity towards Colinus virginianus, Oncorhynchus mykiss, Daphnia magna, and Rat, respectively. According to the strict OECD guidelines, four individual QSTR models were obtained with excellent goodness-of-fit, robustness and predictivity. Moreover, four acceptable interspecies quantitative structure toxicity-toxicity relationship (iQSTTR) models were also developed with good statistical parameters. Applicability domain (AD) analysis indicates that the developed models have a wide application range and reliable predictive ability. Mechanistic explanation reveals that the electronegativity, lipophilicity and molecular bulk can obviously enhance the toxicity of pesticides, whereas the hydrophilic and polar features are shown to reduce the toxicity. Importantly, we predicted the toxicity of many untested pesticides in the comprehensive database for the first time using the developed models and provided the priority ranking list for the toxicity of untested pesticides. In conclusion, the established QSTR and iQSTTR models can be used for the ecotoxicological risk assessment, toxicity data gap filling and regulatory decision of pesticides, as well as aiding the design of new "green" pesticide products. The toxicity prediction for newly designed or untested pesticides will reduce unnecessary chemical synthesis and animal testing, and contribute to the design of "greener and safer" pesticide chemicals.