Quantitative structure–activity relationship study on antitumour activity of a series of flavonoids

Abstract

Flavonoids have been recognised as having antitumour activity for years, although the precise mechanism of this activity is not known. In this study, we used the quantitative structure–activity relationship (QSAR), a method of computer simulation, to investigate this mechanism. The heuristic method (HM), support vector machine (SVM) and gene expression programming (GEP) were employed to build the QSAR models between antitumour activity (i.e. survival percentage) of a series of flavonoids and five molecular descriptors derived from the CODESSA. The GEP, which was used for the first time to predict the antitumour activity of flavonoids, was found much more accurate than the HM and the SVM. The descriptors involved in the models are mainly electrical related, and the quantum chemical factors relevant to the oxygen atom may play an important role in antitumour activity of flavonoids. These models may provide helpful information for designing novel antitumoural compounds with higher bioactivity.

Keywords:

• flavonoids
• gene expression programming
• heuristic method
• quantitative structure–activity relationship
• support vector machine

Acknowledgement

This research was supported in part by the Program for New Century Excellent Talents in University (NCET-06-0116).