Application of genetic algorithm-kernel partial least square as a novel non-linear feature selection method: partitioning of drug molecules.

Genetic algorithm (GA) and partial least squares (PLS) and kernel PLS (KPLS) techniques were used to investigate the correlation between immobilized liposome chromatography partitioning (log Ks) and descriptors for 65 drug compounds. The models were validated using leave-group-out cross validation LGO-CV. The results indicate that GA-KPLS can be used as an alternative modelling tool for quantitative structure-property relationship (QSPR) studies.

Application of artificial neural network to predict the retention time of drug metabolites in two-dimensional liquid chromatography.

Genetic algorithm and partial least square (GA-PLS) and Levenberg-Marquardt artificial neural network (L-M ANN) techniques were used to investigate the correlation between retention time and descriptors for drug metabolites which obtained by two-dimensional liquid chromatography. The applied internal (leave-group-out cross validation (LGO-CV)) and external (test set) validation methods were used for the predictive power of four models. Both methods resulted in accurate prediction whereas more accurate results were obtained by L-M ANN model. The best model obtained from L-M ANN showed a good R(2) value (determination coefficient between observed and predicted values) for all compounds, which was superior to GA-PLS models.