Prediction of anticancer activities of cynaroside and quercetin in leaf of plants Cynara scolymus L and Artocarpus incisa L using structure-activity relationship

Natural products from plants are an alternative resource in the search for anti-cancer drugs and can have a direct impact on eliminating cancer cells and also reduce cancer side effects. Recently, we have isolated a few flavonoid quercetin and cynaroside from leaf of cynara scolymus L and artocarpus incisa L in Vietnam, with cytotoxic activity relatively strong in Hela cancer cells. The flavonoid compound is a search target, research and development of anti-cancer agents in clinical use. To clarify the important nature of the activity, the subject QSAR studies on cancer Hela cell line use the multiple linear regression (MLR) gradually, partial least square regression (PLS) and artificial neural network. The MLR and PLS models showed good correlation values of R-2=0.938, R-pred(2)=0.903, and R-2=0.943, R-pred(2)=0.912, respectively. The MLR model shows the level of importance of atomic charge descriptors. Also, artificial neural network architecture I(6)-HL(4)-O(1) is built with RMSE=0.00345, R-2=0.993, R-pred(2)=0.971 using the atomic charge descriptors selected in the MLR model such as neurons of input layer and the anti-cancer activity such as neuron of output layer. The anti-cancer activities of the flavonoids and isoflavonoids in the test group and compounds quercetin and cynaroside isolated from cynara scolymus L and artocarpus incisa L are compared with experimental data and those from references.