4-X-N-Y-6-氮杂雄-4-烯-3-酮衍生物的QSAR研究

分别采用人工神经网络BP算法（网络结构为3-9-1）和线性回归分析方法对17个4-X-N-Y-6-氮杂雄-4-烯-3-酮衍生物在4℃时与小牛子宫雌激素受体的亲合力参数Iog1/K（iKi为衍生物对3BHSD的抑制常数）与分子的范德华体积V、最高被占据分子轨道能量EHOMO和9号碳原子的净电荷Q之间建立了QSAR模型,ANN模型的相关系数R=0.9999,标准偏差SD=0.0014。MLR模型的相关系数R=0.9470,标准偏差SD=0.4459。结果表明人工神经网络是一种比较精密的拟合方法,具有良好的预测效果。

Study on the QSAR of Diphenylethylene Derivatives

The antiestrogen activity of 4-X-N-Y-6 azaandrost-4-en-3-ones(logl/Ki) are highly correlated with van der Waals volume(V),the highest occupied orbital energy(EHOMO) and net charge of C9(Q).For 17 4-X-N-Y-6 azaandrost-4-en-3-ones,using these parameters the QSAR models were set up respectively with ANN and MLR methods.The correlation coefficient R and the standard error SD of the model are 0.9999 and 0.0014 respectively.The linear regression equation was obtained with MLR method and its correlation coefficient R is equal to 0.9470 and the standard error SD to 0.4459.The result showed that the fitted performance of neural network method was comparatively precise and it has a preferable predicted effect.