异恶唑姜黄素类似物抗结核分枝杆菌活性定量构效关系的研究

Study on the QSAR of Isoxazole Analogs of Curcuminoids with Anti-Mycobacterial Activity

本文分别采用启发式(HM)和径向基函数神经网络方法(RBFNN),研究了异恶唑姜黄素类似物抗结核分枝杆菌活性的定量构效关系。结果显示,运用HM方法建立的线性模型预测得到的判定系数R2=0.882 6,运用RBFNN方法建立的非线性模型预测得到的判定系数R2=0.915 2。表明RBFNN建立的构效关系模型具有更好的预测能力,能够很好地指出姜黄素类似物结构与其抗结核分枝杆菌活性之间的定量关系,指导新药物的合成。

Quantitative structure-activity relationships of isoxazole curcumin analogues against Mycobacterium tuberculosis were studied by using heuristic(HM)and radial basis function neural network(RBFNN).The linear model was built through the method of heuristic method(HM),and the predicted linear model coefficient of determination is R^2=0.8826.The nonlinear model was built through the method of radial basis function neural network(RBFNN),and the predicted model coefficient of determination is R^2=0.9152.The result showed that the radial basis neural network method had high predictive ability,and more suitable for building the Quantitative structure-activity elationship research model of Isoxazole analogs of curcuminoids.