Application Of Chemometrics In The Metabolism Of Traditional Chinese Medicine

There are many differences between Chinese medicine and Western medicine in composition,function and so on.The active components of traditional Chinese medicine are complex and the function of each component will affect each other,toxicity evaluation of traditional Chinese medicine is the urgent problem of the toxicology.Using metabolomics methods of traditional Chinese medicine toxicity conduct a comprehensive analysis and using ~1H NMR for Wistar male rats urine comprehensive testing.~1H NMR spectrum contains all of the metabolites of its urine,but the overall information of the spectrum is very complex.It has become a very important part of modern research to use the methods of the chemometrics to deal with metabolic data.But metabolic data having the number of variables is much larger than the number of samples of features,direct pattern recognition algorithms analyze such data will determine the metabolic bias the results,for large data,if you do not choose a variable,you can not eliminate the interference of other unrelated factors,Therefore,how to accurately and quickly selected variables become the key to an accurate analysis of the metabolic data.This study proposes a new method of pattern recognition,namely ABB-SVM method,and compared with the PCA,SVM,ABB-Fisher three methods,ABB-SVM method of screening variables improved branch and bound method,this method Principal component analysis method compared to more explicit physical meaning,and the improved branch and bound algorithm compared to traditional variable selection method has faster search speed;by determining the effect of a line chart to solve specific sample can be determined from the results of FIG visually see the specific sample;while using support vector machines to solve nonlinear small sample and data classification,good data classification capability,have broad applicability,and has great advantages in terms of visibility..This study on the administration of high and low doses of arsenic and administered six drug model rat urine ~1H NMR data were analyzed,respectively,with PCA and SVM,ABB-Fisher and ABB-SVM four kinds of pattern recognition method established six models of toxicity of arsenic toxicity effect,the discriminant results in visually display the relationship between arsenic dose time toxic.This study on the high and low doses of arsenic and six drug models'rat urine ~1H NMR data were analyzed,respectively,with PCA and SVM,ABB-Fisher and ABB-SVM pattern recognition methods established six models to analyze the toxicity of arsenic toxicity effect,the discriminant results in visually display the relationship of dose-time-toxic.Results show that the varying degrees damage of liver and kidney in high and low dose arsenic,which the toxic effects of arsenic belongs entirely to ANIT,BEA,CCl4,Na2CrO4 models,absolutely does not belong to HYD,HgCl2 models.The study of administration of high,after low dose of body self-recovery process is studied.The model is built using the SVM,ABB-Fisher and ABB-SVM method.Research results show that both high dose and low dose arsenic,after administration,organisms,changes with time and gradually restored to the level before the treatment.As a result,the arsenic further development and utilization of the medicinal value and provide the basis.