摘要
采用本实验室提出的三维全息原子场作用矢量(3D-HoVAIF)对34个环己烯神经氨酸酶抑制剂抗禽流感药物结构并与其活性建立定量构效关系模型,然后采用逐步回归对变量进行筛选后,运用多元线性回归建立3D-HoVAIF描述子与神经氨酸酶抑制剂活性之间的QSAR模型.其复相关系数(R),交互校验的复相关系数(RCV)和模型的标准偏差分别R为0.947,SD为0.466,RCV为0.912,SDCV为0.536,模型具有良好的稳定性和预测能力.表明三维全息原子场作用矢量能较好表征该类分子结构信息值得进一步推广应用.
The three-dimensional holographic vector of atomic interaction field （3D-HoVAIF） was adopted to describe the chemical structure of 34 influenza neuraminidase inhibitors used as anti bird-flu drugs. After the structural characterization, the descriptors obtained were screened by stepwise multiple regression （SMR）. The QSAR models are built by MLR with the correlation coefficients （R） and the Cross-Validation correlation coefficient（RcV） of molecular modeling being 0. 947and 0. 912. While the Cross-validation of the models, which contain selected vectors, are performed by Leave-One-Out procedure （LOO） , with the standard deviation（SD） and SDcv being 0. 466 and 0. 536, respectively. The result showed that the model had excellent stability and good prediction capability and the 3D-HoVAIF was worth popuparizing in the molecular structural characterization and biological activity prediction.
出处
《内江师范学院学报》
2009年第6期33-36,共4页
Journal of Neijiang Normal University
关键词
神经氨酸酶抑制剂
三维原子场全息作用矢量
定量构效关系
influenza neuraminidase inhibitors, three-dimensional holographic vector of atomic interaction field （3D- HoVAIF）, quantitative structure-activity relationship（QSAR）.
