Prediction of Bovine Serum Albumin-Water Partition Coefficients of a Wide Variety of Neutral Organic Compounds by Means of Support Vector Machine

Using machine learning and quantum chemistry descriptors to predict the toxicity of ionic liquids

Large-scale application of ionic liquids (ILs) hinges on the advancement of designable and eco-friendly nature. Research of the potential toxicity of ILs towards different organisms and trophic levels is insufficient. Quantitative structure-activity relationships (QSAR) model is applied to evaluate the toxicity of ILs towards the leukemia rat cell line (ICP-81). The structures of 57 cations and 21 anions were optimized by quantum chemistry. The electrostatic potential surface area (S_EP) and charge distribution area (S_σ-profile) descriptors are calculated and used to predict the toxicity of ILs. The performance and predictive aptitude of extreme learning machine (ELM) model are analyzed and compared with those of multiple linear regression (MLR) and support vector machine (SVM) models. The highest R² and the lowest AARD% and RMSE of the training set, test set and total set for the ELM are observed, which validates the superior performance of the ELM than that of obtained by the MLR and SVM. The applicability domain of the model is assessed by the Williams plot.

Development of bovine serum albumin-water partition coefficients predictive models for ionogenic organic chemicals based on chemical form adjusted descriptors

The partition coefficients between bovine serum albumin (BSA) and water (K_BSA/w) for ionogenic organic chemicals (IOCs) were different greatly from those of neutral organic chemicals (NOCs). For NOCs, several excellent models were developed to predict their logK_BSA/w. However, it was found that the conventional descriptors are inappropriate for modeling logK_BSA/w of IOCs. Thus, alternative approaches are urgently needed to develop predictive models for K_BSA/w of IOCs. In this study, molecular descriptors that can be used to characterize the ionization effects (e.g. chemical form adjusted descriptors) were calculated and used to develop predictive models for logK_BSA/w of IOCs. The models developed had high goodness-of-fit, robustness, and predictive ability. The predictor variables selected to construct the models included the chemical form adjusted averages of the negative potentials on the molecular surface (V_s-adj^-), the chemical form adjusted molecular dipole moment (dipolemoment_adj), the logarithm of the n-octanol/water distribution coefficient (logD). As these molecular descriptors can be calculated from their molecular structures directly, the developed model can be easily used to fill the logK_BSA/w data gap for other IOCs within the applicability domain. Furthermore, the chemical form adjusted descriptors calculated in this study also could be used to construct predictive models on other endpoints of IOCs.

Development and evaluation of predictive model for bovine serum albumin-water partition coefficients of neutral organic chemicals

The binding of organic chemicals to serum albumin can significantly reduce their unbound concentration in blood and affect their biological reactions. In this study, we developed a new QSAR model for bovine serum albumin (BSA) - water partition coefficients (K_BSA/W) of neutral organic chemicals with large structural variance, logK_BSA/W values covering 3.5 orders of magnitude (1.19-4.76). All chemical geometries were optimized by semi-empirical PM6 algorithm. Several quantum chemical parameters that reflect various intermolecular interactions as well as hydrophobicity were selected to develop QSAR model. The result indicates the regression model derived from logK_ow, the most positive net atomic charges on an atom, Connolly solvent excluded volume, polarizability, and Abraham acidity could explain the partitioning mechanism of organic chemicals between BSA and water. The simulated external validation and cross validation verifies the developed model has good statistical robustness and predictive ability, thus can be used to estimate the logK_BSA/W values for chemicals in application domain, accordingly to provide basic data for the toxicity assessment of the chemicals.

QSPR studies for predicting polarity parameter of organic compounds in methanol using support vector machine and enhanced replacement method

In the present work, enhanced replacement method (ERM) and support vector machine (SVM) were used for quantitative structure-property relationship (QSPR) studies of polarity parameter (p) of various organic compounds in methanol in reversed phase liquid chromatography based on molecular descriptors calculated from the optimized structures. Diverse kinds of molecular descriptors were calculated to encode the molecular structures of compounds, such as geometric, thermodynamic, electrostatic and quantum mechanical descriptors. The variable selection method of ERM was employed to select an optimum subset of descriptors. The five descriptors selected using ERM were used as inputs of SVM to predict the polarity parameter of organic compounds in methanol. The coefficient of determination, r², between experimental and predicted polarity parameters for the prediction set by ERM and SVM were 0.952 and 0.982, respectively. Acceptable results specified that the ERM approach is a very effective method for variable selection and the predictive aptitude of the SVM model is superior to those obtained by ERM. The obtained results demonstrate that SVM can be used as a substitute influential modeling tool for QSPR studies.

2D and 3D QSAR models for identifying diphenylpyridylethanamine based inhibitors against cholesteryl ester transfer protein

Cholesteryl ester transfer protein (CETP) inhibitors hold promise as new agents against coronary heart disease. Molecular modeling techniques such as 2D-QSAR and 3D-QSAR analysis were applied to establish models to distinguish potent and weak CETP inhibitors. 2D and 3D QSAR models-based a series of diphenylpyridylethanamine (DPPE) derivatives (newly identified as CETP inhibitors) were then performed to elucidate structural and physicochemical requirements for higher CETP inhibitory activity. The linear and spline 2D-QSAR models were developed through multiple linear regression (MLR) and support vector machine (SVM) methods. The best 2D-QSAR model obtained by SVM gave a high predictive ability (R(2)train=0.929, R(2)test=0.826, Q(2)LOO=0.780). Also, the 2D-QSAR models uncovered that SlogP_VSA0, E_sol and Vsurf_DW23 were important features in defining activity. In addition, the best 3D-QSAR model presented higher predictive ability (R(2)train=0.958, R(2)test=0.852, Q(2)LOO=0.734) based on comparative molecular field analysis (CoMFA). Meanwhile, the derived contour maps from 3D-QSAR model revealed the significant structural features (steric and electronic effects) required for improving CETP inhibitory activity. Consequently, twelve newly designed DPPE derivatives were proposed to be robust and potent CETP inhibitors. Overall, these derived models may help to design novel DPPE derivatives with better CETP inhibitory activity.