多维标度法(MDS)在构效关系分析中的应用  

The application of multi-dimensional scaling method in structure-activity relationship research

作　　者：陆忠华[1] 胡永宏[2] 沈斌[1] 迟学斌[1] 

机构地区：[1]中国科学院计算机网络信息中心超级计算中心,北京100080 [2]中央财经大学统计学院

出　　处：《计算机与应用化学》2007年第3期285-290,共6页Computers and Applied Chemistry

基　　金：国家自然科学基金(60673064;60373060);国家科技部资助项目(2005DKA64002)

摘　　要：尝试在构效关系分析中引入新的数量化方法:林知己夫的准记量性多维标度(multi-dimensional scaling,MDS)法。并以芳氧基乙酰肼衍生物抑制单胺氧化酶的效率为例,说明该法的实现步骤和结果分析。该法的结果直观、生动,可以将样本分成明显的高活性区与低活性区,作为判别标准,可以在试验之前预测新样本化合物的活性,以指导新化合物的合成,为在药物分子设计中判断未知化合物的生物活性提供新思路。A numerical method, multi-dimensional scaling （MDS） method, has been introduced in the research of structure-activity relationship （SAR）. The Monoamine Oxidase＇s inhibitors, i.e. aryloxylacetyl hydrazine derivatives, are taken as the example of calculating the training set, in order to illustrate the implementation steps of the MDS method and the analysis of the results. The results obtained by using the MDS method were intuitionistic and vivid, which can separate the training set by significant high activity region and lower activity region so as to be taken as the criterions for judging and predicting the activities of new specimen before the specimen be synthesized in lab for guiding the synthesis. To use MDS method in SAR research provides new idea for predicting unknown compounds in computer aided drug design.

关 键 词：数量化方法 构效关系 药物设计 多维标度法 

分 类 号：R914[医药卫生—药物化学] TQ46[医药卫生—药学]