氨基酸含量与寒热药性的多元统计分析

Application of Multivariate Statistical Analysis in Amino Acids Compositions and Cold-Heat Nature of Traditional Chinese Medicine

目的:探讨中药寒热药性与18种氨基酸含量的相关性。方法:应用柱前衍生化反应和HPLC法测定17种氨基酸含量,并采用UV法测定色氨酸含量;利用Fisher线性判别分析对18种氨基酸含量与中药寒热药性的相关性进行分析;比较确定最佳的统计识别模型,并进一步识别氨基酸的寒热表征。结果:建立基于18种氨基酸含量的寒热药性数学Fisher判别函数,60种中药的寒热药性判别正确率达到82%。以SVM为最佳模型得出氨基酸HPLC分析的寒性标记为:Glu、Gly、Arg、Thr、Ala、Tyr、Val、Ile和lys;热性标记为:Asp、Ser、His、Pro、Met、Cys、Leu、Phe和Trp。结论:18种氨基酸含量与中药寒热药性具有一定的相关性。

This study was aimed to find the correlation between amino acid compositions and Cold-Heat Nature of traditional Chinese medicine（TCM） in order to provide basis for the research of TCM natural theory.A total of 17 kinds of amino acid were determined by application before column derivatization reaction and high-perfor mance liquid chromatography（HPLC） and the tryptophan was determined by UV method.The data were collected for analysis by Fisher method in PAST software.The best statistical identification model was determined.And the Cold-Hot medicine property markers（CHMP-markers） were determined.The results showed that the discriminant function established by Fisher method based on 18 kinds of amino acid contents has good identification ability,and the accuracy of the Fisher discriminant analysis is 82%.Support vector machine（SVM） is the best statistical identification model.The ＂cold＂ and ＂heat＂ markers were analyzed by SVM.The cold nature material bases include Glu,Gly,Arg,Thr,Ala,Tyr,Val,Ile and lys.And the heat nature material bases contain Asp,Ser,His,Pro,Met,Cys,Leu,Phe and Trp.It was concluded that there is relationship between 18 kinds of amino acid contents and the Cold-Heat nature of TCM.