QSPR Prediction for Nitrogen-Containing Heterocycles from Molecular Descriptors in HPLC

초록

Liquid chromatographic retention of 29 nitrogen containing heterocycles has been predicted by quantitative structure property relationships (QSPR) methods based on simple and multiple linear regression analysis. In order to indicate the influence of different computed molecular descriptors on retention factors and well understand the important descriptors affecting the experimental values, simple and multiple linear models derived from two groups of different molecular descriptors (structural and physicochemical) were built. The proposed simple linear models gave the following results: the square of correlation coefficient, R2, for the models physicochemical and structure molecular descriptors was in variety 0.80≤R2≤1.00. The QSPR estimations of these nitrogen containing heterocyclic compounds could be predicted with a multiple linear regression equation having the statistical index, R2=1.00. Furthermore, some molecular descriptors in these models were discussed to well understand the relationships between molecular structures and their retention parameters. The present work provides an effective method for the prediction of the retention for the nitrogen containing heterocyclic compounds. This study also showed that the utility of the QSPR treatment involving descriptors derived solely from chemical structure. The correlation equations and descriptors can be used for the prediction of the chromatographic retention for unknown structures.