LAUSR

In the present study, a very thorough and in-depth three-dimensional quantitative structure-toxicity relationship (3D-QSTR) analysis has been implemented to make a correlation between the structural information of the ionic liquids (ILs) and their cytotoxicity towards Leukemia rat cell line IPC-81, as one of the ILs' toxicological consequences. To do this, alignment free GRid-INdependent Descriptors (GRINDs), which were derived from molecular interaction fields (MIFs), were correlated to the cytotoxicity values by partial least squares (PLS) and support vector regression (SVR). Genetic algorithm (GA), as a powerful linear tool, was used to select the best and interpretative subset of variables for the predictive model building. The selected variables with the capability to screen the effective structural features, showed direct and inverse contribution to the cytotoxicity. In silico modeling can reduce the amount of cellular testing necessary by predicting the toxicological functions of the chemical structures. Acceptable predictions of both internal and external validation sets made it possible to develop the predictive models for a large set of 269 diverse ILs containing 9 cationic cores and 44 types of anions. The constructed 3D-QSTR models use simple and interpretable descriptors to provide an in-depth and mechanistic interpretation of structural characteristics. This helps provide a clear understanding of the cytotoxicity effects of the understudy ILs. The effects of the nature of the cations, anions, and substituents on the cytotoxicities were evaluated and discussed.