Application Of Chromatography And Spectroscopy Combined With Chemometrics In Quality Analysis For Some Foods And Traditional Chinese Medicines

The intrinsic characteristic fingerprints of complex materials, such as traditional Chinese medicines(TCM) and foods, are established by chromatography and spectroscopy in this article. The quality analysisand control for foods and TCM was accomplished with the aid of chemometrics, at the same time, the activepharmaceutical compositions were detected in these complex materials. The fingerprints could reflect on theintrinsic features of chemical components in complex materials and apply to quantitative analysis, qualitycontrol and evaluation, which is a comprehensive and quantitative characterization method. Fingerprintpossesses advantages for the study objects, such as, comprehensiveness, relevance, hierarchy and dynamic.Chemometrics method is a powerful supplementary means and could extract and process effectiveinformation about quality of TCMs and foods. In this work, chromatography and spectroscopy with the aidof chemoetrics were applied to analysis of Forsythia suspense, Cortex moutan, and Chinese yam. The contentof the main research work and the results are as follows:1. The fingerprints of Forsythia suspensa of raw (Laoqiao) and ripe (Qingqiao) fruit were obtained byHPLC-DAD and the EIS-MS techniques, it was founded that the enormous differences are existed betweenthe Laoqiao and Qingqiao samples in intensity and number of fingerprint peaks. Seventeen peaks weredetected, and in which, twelve were identified. Most were related to the glucopyranoside molecularfragment. Samples collected from three geographical areas (Shanxi, Henan and Shandong provinces), werediscriminated with the use of hierarchical clustering analysis (HCA), discriminant analysis (DA) andprincipal component analysis (PCA) models, but only PCA provided extra information about therelationships between objects and loadings. Thus, the Qingqiao ripe fruit were discriminated on the basis oftheir places of origins, because there are dramatic variations existed in Forsythia suspensas from differentorigins.2. Cortex Moutan (CM) samples from different origin were used as the study objects. The application ofhigh performance liquid chromatography-diode array detector (HPLC-DAD) and inductively coupledplasma-mass spectrometry in determination of organic components and metal elements in Cortex Moutan(CM) samples. The HPLC peaks (organic components) of the CM samples, and the ICP-MS measurements(metal elements) were investigated with PCA and linear discriminant analysis (LDA) methods; essentiallyqualitative results suggested that discrimination of the CM samples from three different origins were possiblewith two-way matrix and producing better results. Another three methods were applied for the classificationand prediction of the samples. Again the two-way data matrix analysed by the K-nearst neighbour method(KNN) produced best results (100%correct for prediction set). Additionally, multiple linear regression(MLR) was utilized to explore any relationships between the organic constituents and the metal elements ofthe CM samples; the extracted linear regression equations showed that both essential metals as well as somemetallic pollutants were related to the organic compounds on the basis of their concentrations.3.108Chinese yam samples from four geographical regions were regarded as the research objects. Nearinfrared spectroscopy (NIRS,800-2500nm) and middle infrared spectroscopy (MIRS,400-4000cm-1) wereused for the discrimination of Chinese yam from four geographical regions, and for quantifying theconcentration of the total sugar and flavoinds. PCA was applied to classify the Chinese yam powderaccording to their origins on the basis of NIR and MIR data, respectively, and satisfactory classificationresults were obtained. The prediction results of two pattern recognition methods, partial least-squares (PLS),and least squares-.support vector machine (LS-SVM) for geographical regions and chemical compounds(total sugar and flavonoids) were compared with each other. These models could have better performancebased on the NIR and MIR data.