Search

• We predicted the acute toxicity of organophosphates based on chemical structure. • Large data were collected to develop quantitative structure-toxicity relationship models. • The random forest algorithm was used to develop these models. • Acute toxicity data from different administration routes could not be combined. • QSTR models achieve high predictive performance despite large variance in datasets. Organophosphates (OPs) are highly toxic compounds, with widespread application in agricultural and chemical industries, whose introduction into the environment poses serious hazards to humans and ecological systems. To assess and ultimately mitigate these hazards, this study predicted the acute toxicity of OPs according to their chemical structure and administration route. The acute toxicity data of 161 OPs in two species via six different administration routes were manually collected and used to develop a series of quantitative structure–toxicity relationship (QSTR) models with robust and practical predictive abilities. The random forest algorithm was used to develop the models, employing both quantum chemical and two-dimensional descriptors according to OECD guidelines. Correlation results and feature similarities indicated that whereas acute toxicity data from rats and mice via the same administration route were combinable for modeling, data from different routes were not. Six QSTR models for each route in a single species and two QSTR models for a single route in the two species were constructed, achieving practical predictive performance. Despite significant variances in their datasets, the prediction models could predict the acute toxicity of novel or unknown OPs, realize rapid assessment, and provide guidance for regulatory decisions to reduce the hazards of OPs. [ABSTRACT FROM AUTHOR]

[Display omitted] • This study used sesquiterpenoids with 97 carbon skeletons for model construction. • The scheme based on the DFT method combined with triple correction can accurately calculate Δ c H of sesquiterpenoid HEDFs. • We developed the MLR equation with only 4 features for predicting Δ c H for sesquiterpenoid HEDFs. • The proposed DFT-QSPR scheme can effectively guide the design of novel sesquiterpenoid HDEFs. Sesquiterpenoids that are renewable and have high energy density and low freezing point are promising biomass high-energy–density fuels (HEDFs). The enthalpy of combustion (Δ c H) is the key characteristic to measure the heat energy content of HEDFs. Its accurate theoretical prediction and evaluation is essential for designing novel sesquiterpenoid HEDFs. Combining the screened and optimized density functional theory (DFT) method with triple computational correction, we presented a standardized calculation scheme for Δ c H with an average absolute error of only 2.6 % by 295 structurally diverse sesquiterpenoid HEDFs. Then, the extreme gradient boosting (XGBoost) algorithm was used to determine the correlation between the calculated Δ c H and 54 quantum chemical (QC) descriptors and the shapley additive explanation (SHAP) method to elucidate the key factors affecting Δ c H. Lastly, a quantitative structure–property relationship (QSPR) prediction model with only four QC features was constructed by using the SHAP results via the multiple linear regression (MLR) algorithms. Notably, the model has excellent prediction performance, decreased complexity, and improved applicability. The coefficient of determination (R 2) of the internal training set and external test set of the model are 0.957 and 0.956, while the root mean square error (RMSE) are 8.626 kcal/mol and 9.012 kcal/mol, respectively. The proposed DFT-QSPR (the ingenious combination of DFT and QSPR) scheme can effectively guide the design of Δ c H of novel sesquiterpenoid HDEFs. [ABSTRACT FROM AUTHOR]