Three-Dimensional Quantitative Structure- Activity Relationship (3D-QSAR) Model for Predicting Toxicity of Ionic Liquids Towards Leukemia rat cell line IPC-81

Ionic liquids (ILs) are generally known as designer solvents due to their unique physicochemical properties by various combination of organic and inorganic cations and anions which caused that they have been applied in multifarious fields such as analytical chemistry [1], chromatography [2], electrochemistry [3], separation and extraction techniques [4], drug design [5] etc. But it is essential to note to their potential to produce toxic effects on living ecosystems like other chemicals. Therefore, it is required to understand ILs’ toxicological consequence to design ILs which are structurally safe for environment. In this study, multivariate linear and nonlinear analysis such as partial least squares (PLS) and support vector machines (SVMs) were applied for constructing 3D-QSAR model using extracted alignment independent descriptors derived from GRID molecular interaction fields (MIFs) [6] in order to relate the structure of ILs to their cytotoxicity towards Leukemia rat cell line IPC-81. Genetic Algorithm as a powerful tool was used in order to select the best and interpretative subset of variables for building predictive models. These QSAR models provided acceptable predictions in both internal and external validation sets and made it possible to develop the predictive model for a large set of 269 ILs containing 9 cationic core and 44 types of anions. The effect of nature of the cation, anion and substituents on the subsequent toxicity were evaluated and discussed.

Ionic liquids, Cytotoxicity, Alignment independent 3D-QSAR, PLS, SVM