Insight QSDAR models for prediction of anticancer activity on Hela cell line of new flavonoid isolating from rhizome Zingiber zerumbet SM in Viet Nam

Abstract

This research predicts the anticancer activity on the Hela cell line of new flavonoid kaempferol-3-O-methyl ether isolating from rhizome Zingiber zerumbet SM by using the spectrum data activity relationship (QSDAR) models. This model has been developed for a set of 3-aminoflavonoids based on the simulated-spectral data ¹³C NMR and ¹⁵O NMR resulting from the semi-empirical quantum chemical calculations TNDO/2 SCF. The atomic sites O₁, O₁₁, C₂, C₃, C₆, C_7, and C_2’ in the QSDAR models significantly contribute to anticancer activity resulting from the Genetic algorithm (GA). The best regression model QSDAR_MLR with the values R²_train of 0.9057 and R²_test of 0.7137, and the neural network model QSDAR_ANN I(7)-HL(9)-O(1) with the values R²_train of 0.993 and R²_pred of 0.971 have been explored to predict the anticancer activities on Hela cell line for new flavonoid kaempferol-3-O-methyl ether from rhizome Zingiber zerumbet SM in Viet Nam.