分子电性距离矢量预测金线莲挥发油成分结构表征及色谱保留值

Structural Description and Prediction of Retention Times for Essential Oil Components of Anoectochilus Roxburghii with MEDV Method

采用分子电性距离矢量(molecular electro-negativity distance vector,MEDV)对金线莲挥发油中分离的72种化合物进行了结构表征,并采用SPSS统计处理软件对变量进行了逐步回归(SMR)分析,建立了4变量的定量结构-色谱保留(quantitative structure retention relationship,QSRR)相关模型,同时采用内部及外部双重验证的办法对所建模型稳定性进行了分析和验证,建模计算值、留一法(leave-one-out,LOO)交互检验预测值和外部样本预测值的复相关系数R、RLOO、Qext分别为0.959,0.947,0.913。结果表明,MEDV能较好地表征非同类化合物的分子结构信息,所建模型具有良好的稳定性和预测能力,将其用于中药化学成分分离、分析、结构表征及其药用价值的有效开发,具有重要的理论指导意义。

The MEDV descriptor has been used to characterize the structure of 72 components from Anoectochilus roxburghii essential oil and the model of the quantitative structure retention relationship of the above mentioned 72 components has also been built.Four variables were selected to build the model through the method of stepwise multiple regression（SMR） using the SPSS software.The estimation stability and generalization ability of the model was strictly analyzed by both the internal and external validation.The correlation coefficient（R） for the established muhiple linear regression （MLR）, leave-one-out（ LOO ）, cross validation （ CV ） and Qext were 0. 959,0. 947,0. 913, respectively. The result showed that the MEDV can be used for the characterization of the structure information for inhomogeneous compounds. The model is good of stability and predictability. The method plays an important role in the theoretical gu!dance, such as separating, analysising and confirming the structure of chemical composition from the Chinese medicine and exploiting their medical value.