摘要
将6种基团参数用于硝基苯化合物取代基结构参数化,应用林知己夫多维标度法,从一组对象之间的相似性度量或非相似性度量出发,求出这组对象在某低维空间中的标度对它们进行了定量构效关系研究,同时将所得的结果映射到二维空间以得到更直观的图像.研究结果表明,该方法能将样本数分成明显的高活性区、中活性区及低活性区三个部分,与实验结果完全吻合.
The multi-dimensional scaling（MDS）method was applied to systematically study the quantitative structure-activity relationship of nitrobenzene derivatives which is based on their substituted group structure descriptor characterized with six group parameters. It begins with the data of similarity measurement or non-similarity measurement between the objects of a group. Then,the scales of the group objects in a low dimension space is proposed. The results indicate that the method could separate the training subset from high activity region,middle and lower activity region by mapping these multiple descriptor values into a 2D plane. Furthermore,distribution rule of biological toxicity for nitrobenzene derivatives was obtained in good accordance with experimental database.
出处
《长沙理工大学学报（自然科学版）》
CAS
2008年第1期93-97,共5页
Journal of Changsha University of Science and Technology:Natural Science
基金
湖南省自然科学基金资助项目(05JJ30198)
关键词
多维标度法
硝基苯化合物
基团参数
生物毒性
定量结构-活性关系
multi-dimensional scaling （MDS） method
nitrobenzene derivatives
group parameter
biological toxicity
quantitative structure-activity relationship（QSAR）
