运用线性和非线性的方法预测烷基苯的沸点和摩尔体积(英文)  

QSAR study of cinnamylphenol derivatives as EBV-EA inhibitors

作　　者：张文军[1] 张运陶[1] 赵鑫[1] 

机构地区：[1]西华师范大学应用化学研究所,四川南充637002

出　　处：《计算机与应用化学》2009年第12期1598-1602,共5页Computers and Applied Chemistry

摘　　要：以2D-autocorrelation描述符为结构参数,采用PSO和逐步回归的方法进行变量筛选,再结合SVM等机器学习算法对28种苯丙烯盐类化合物对EBV-EA病毒的抑制性活性进行定量构效关系(QSAR)研究.研究结果表明,PSO-v-SVM模型具有最优的模型稳健性和预测效果.由PSO选入的构成该模型的5个2D-autocorrelation描述符为ATS5v,ATS6e,ATS8e,ATS3p,GATS5p;该模型对训练集的拟合和留一法交叉验证结果的相关系数R^2和q_(cv)~2分别为0.986和0.930,对测试集预测结果的相关系数R^2_(ext)达0.955.对5个变量的理化意义的分析表明,极化率、Van der Waals体积和电负性对苯丙烯盐类化合物的抑制性活性影响分别约占57.13%、15.90%和26.97%.In this work some chemo metrics methods were applied for modeling and predicting the inhibitory activity of cinnamylphenol derivatives with 2D-autocorrelation descriptors calculated from the molecular structure alone for the first time. The stepwise multiple linear regression （Stepwise -MLR） and particle swarm optimization （PSO） methods were used to select descriptors which are responsible for the inhibitory activity of these compounds. Mathematical models are obtained by support vector machine （SVM）, least squares support vector machine regression （LSSVM） and multiple linear regression （MLR）. The square of the correlation coefficient （R^2=0.990）, the square of correlation coefficients （Rext^2） of predicting set （Rext^2=0.955）, and the obtained statistical parameter of ＇leave-one-out＇ （LOO） on PSO-v-SVM model was 0.930, which revealed the reliability of the model. Our best QSAR model illustrates the importance of an adequate distribution of atomic properties represented in topological frames and reveals the polarizability, van der Waals volumes, Sanderson and electron negativities as the most influencing atomic properties in the structures of the cinnamylphenol derivatives. A comparison with other approaches such as the Randid molecular profiles, Geometrical, 3D-MoRSE, Quantum chemical parameters and RDF descriptors were also carried out.

关 键 词：定量结构活性关系 支持向量机 2D自相关描述符 微粒群 抑制性活 

分 类 号：O62[理学—有机化学] TP181[理学—化学]