RESUMEN

A prediction of quantitative structure-property relationships is developed to model the polarity parameter of a set of 146 organic compounds in acetonitrile in reversed-phase liquid chromatography. Enhanced replacement method and support vector machine regressions were employed to build prediction models based on molecular descriptors calculated from the structure alone. The correlation coefficients between experimental and predicted values of polarity parameter for the test set by enhanced replacement method and support vector machine were 0.970 and 0.993, respectively. The obtained results demonstrated that the support vector machine model is more reliable and has a better prediction performance than the enhanced replacement method.

RESUMEN

In the present study a quantitative structure-activity relationship (QSAR) technique was developed to investigate the blood-to-brain barrier partitioning behavior (log BB) for various drugs and organic compounds. Important descriptors were selected by genetic algorithm-partial least square (GA-PLS) methods. Partial least squares (PLS) and support vector machine (SVM) methods were employed to construct linear and non-linear models, respectively. The results showed that, the log BB values calculated by SVM were in good agreement with the experimental data, and the performance of the SVM model was superior to the PLS model. The study provided a novel and effective method for predicting blood-to-brain barrier penetration of drugs, and disclosed that SVM can be used as a powerful chemometrics tool for QSAR studies.

Asunto(s)

RESUMEN

Support vector machine (SVM) was used to develop a quantitative structure property relationship (QSPR) model that correlates molecular structures to their bovine serum albumin water partition coefficients (KBSA/W ). The performance and predictive aptitude of SVM are considered and compared with other methods such as multiple linear regression (MLR) and artificial neural network (ANN) methods. A set of 83 natural organic compounds and drugs were selected and suitable sets of molecular descriptors were calculated. Genetic algorithm (GA) was used to select important molecular descriptors, and linear and nonlinear models were applied to correlate the selected descriptors with the experimental values of log KBSA/W . The correlation coefficients, R, between experimental and predicted log KBSA/W for the validation set by MLR, ANN and SVM are 0.951, 0.986 and 0.991, respectively. Results obtained document the reliability and good predictability of the nonlinear QSPR model to predict partition coefficients of organic compounds. Comparison between the values of statistical parameters demonstrates that the predictive ability of the SVM model is comparable or superior to those obtained by MLR and ANN.

RESUMEN

In this study, a quantitative structureproperty relationship (QSPR) study is developed for the prediction of gas to dimethyl sulfoxide solvation enthalpy (ΔHSolv ) of organic compounds based on molecular descriptors calculated solely from molecular structure considerations. Diverse types of molecular descriptors were calculated to represent the molecular structures of the various compounds studied. Multiple linear regression (MLR) was employed to select an optimal subset of descriptors that have significant contributions to the ΔHSolv overall property. Our investigation revealed that the dependence of physicochemical properties on solvation enthalpy is a nonlinear observable fact and that MLR method is unable to model the solvation enthalpy accurately. It has been observed that support vector machine (SVM) and artificial neural network (ANN) demonstrates better performance compared with MLR. The standard error value of the test set for SVM is 1.731 kJ mol(-1) , while it is 2.303 kJ mol(-1) and 5.146 kJ mol(-1) for ANN and MLR, respectively. The results showed that the calculated ΔHSolv values by SVM were in good agreement with the experimental data, and the performance of the SVM model was superior to those of MLR and ANN ones.

RESUMEN

The main aim of this study was the development of a quantitative structure-property relationship method using an artificial neural network (ANN) for predicting the water-to-wet butyl acetate partition coefficients of organic solutes. As a first step, a genetic algorithm-multiple linear regression model was developed; the descriptors appearing in this model were considered as inputs for the ANN. These descriptors are principal moment of inertia C (I(C)), area-weighted surface charge of hydrogen-bonding donor atoms (HACA-2), Kier and Hall index (order 2) ((2)χ), Balaban index (J), minimum bond order of a C atom (P(C)) and relative negative-charged SA (RNCS). Then a 6-4-1 neural network was generated for the prediction of water-to-wet butyl acetate partition coefficients of 76 organic solutes. By comparing the results obtained from multiple linear regression and ANN models, it can be seen that statistical parameters (Fisher ratio, correlation coefficient and standard error) of the ANN model are better than that regression model, which indicates that nonlinear model can simulate the relationship between the structural descriptors and the partition coefficients of the investigated molecules more accurately.

RESUMEN

In this work a quantitative structure-activity relationship (QSAR) technique was developed to investigate the air to liver partition coefficient (log Kliver) for volatile organic compounds (VOCs). Suitable set of molecular descriptors was calculated and the important descriptors were selected by GA-PLS methods. These variables were served as inputs to generate neural networks. After optimization and training of the networks, they were used for the calculation of log Kliver for the validation set. The root mean square errors for the neural network calculated log Kliver of training, test, and validation sets are 0.100, 0.091, and 0.112, respectively. Results obtained reveal the reliability and good predictivity of neural network for the prediction of air to liver partition coefficient for volatile organic compounds.

Asunto(s)

RESUMEN

The main aim of the present work was development of a quantitative structure-property relationship method using an artificial neural network (ANN) for predicting gas-to-olive oil partition coefficients of organic compounds. As a first step, a multiple linear regression (MLR) model was developed; the descriptors appearing in this model were considered as inputs for the ANN. These descriptors are: solvation connectivity index chi(-1), hydrophilic factor, conventional bond-order ID number, dipole moment and a total size index/weighted by atomic masses. Then a 5-5-1 neural network was generated for the prediction of gas-to-olive oil partition coefficients of 179 organic compounds including hydrocarbons, alkyl halides, alcohols, ethers, esters, ketones and benzene derivatives. The values of standard error for training, test and validation sets are 0.127, 0.122 and 0.162, respectively for ANN model. Comparisons between these values and other obtained statistical values reveal the superiority of the ANN model over the MLR one.

Asunto(s)