Toxicity of ionic liquids: Database and prediction via quantitative structure–activity relationship method

Yongsheng Zhao,Jihong Zhao,Ying Huang,Qing Zhou,Xiangping Zhang,Suojiang Zhang
DOI: https://doi.org/10.1016/j.jhazmat.2014.06.018
IF: 13.6
2014-08-01
Journal of Hazardous Materials
Abstract:A comprehensive database on toxicity of ionic liquids (ILs) is established. The database includes over 4000 pieces of data. Based on the database, the relationship between IL's structure and its toxicity has been analyzed qualitatively. Furthermore, Quantitative Structure-Activity relationships (QSAR) model is conducted to predict the toxicities (EC50 values) of various ILs toward the Leukemia rat cell line IPC-81. Four parameters selected by the heuristic method (HM) are used to perform the studies of multiple linear regression (MLR) and support vector machine (SVM). The squared correlation coefficient (R(2)) and the root mean square error (RMSE) of training sets by two QSAR models are 0.918 and 0.959, 0.258 and 0.179, respectively. The prediction R(2) and RMSE of QSAR test sets by MLR model are 0.892 and 0.329, by SVM model are 0.958 and 0.234, respectively. The nonlinear model developed by SVM algorithm is much outperformed MLR, which indicates that SVM model is more reliable in the prediction of toxicity of ILs. This study shows that increasing the relative number of O atoms of molecules leads to decrease in the toxicity of ILs.
environmental sciences,engineering, environmental
What problem does this paper attempt to address?