Approaches to quantitative structure–enantioselectivity relationship modeling of chiral separations using capillary electrophoresis

Quantitative structure–enantioselectivity relationships (QSERs) have been developed to describe the resolution of a series of chiral arylpropionic acids using capillary electrophoresis. Native β-cyclodextrin and two derivatized forms are used as the chiral resolving agents. The QSER models are developed using the results of molecular mechanics calculations as input to multivariate linear regression and also to neural networks. Single models are developed to predict the optimum cyclodextrin to resolve a given analyte, the migration order, and the magnitude of the separation. Models are also developed to predict only the optimum cyclodextrin.