Rapid Discrimination of the Different Processed Paris polyphylla var. yunnanensis with Infrared Spectroscopy Combined with Chemometrics
Wu Zhe,Zhang Ji,Zuo Zhi-tian,Xu Fu-rong,Wang Yuan-zhong,Zhang Jin-yu
DOI: https://doi.org/10.3964/j.issn.1000-0593(2018)04-1101-06
2018-01-01
Spectroscopy and spectral analysis
Abstract:Based on the theory of Chinese medicine, the processing which improves efficacy, property and flavor of traditional Chinese medicine (TCM) is an effective way to moderate property, enhance therapeutic effect or reduce toxicity of TCM. The processing has significant influence on the chemical components, efficacy and toxicity of TCM. Therefore, it is vital to establish a system method to discriminate and evaluate different processed products of TCM, which can provide an important support for the quality and clinical medication security of TCM. In this paper, Paris polyphylla var. yunnanensis which were processed with nine different methods were conducted comparative analysis by infrared spectroscopy combined with chemometrics, and the principal component analysis-Mahalanobis distance (PCA-MD) discriminant model was established to differentiate them. The original infrared spectra data was preprocessed by automatic baseline correction and ordinate normalization, and the averaged spectra were obtained. The averaged and second derivative spectra showed that: (1) The main characteristic absorption peaks were 3 387, 2 923, 1 745, 1 463, 1 338, 1240, 1207, 1158, 1 180, 1 080, 1048, 1 020, 988, 921, 895, 859, 833, 765, 708, 572 and 529 cm(-1). (2) The peak shape of samples was almost alike, which could exhibit the infrared spectral features of processed P. yunnanensis. (3) Some differences of a few characteristic absorption peaks existed in number, position and absorption intensity, which indicated that the chemical components and content were changed after different processing. The infrared spectra data was pretreated by multiplicative scatter correction (MSC), standard normal variate (SNV), first derivative (1st Der), second derivative (2nd Der) and Savitzky-Golay (SG) smoothing. Samples were divided into calibration set and prediction set at the ratio of 3 : 1 by Kennard-Stone algorithm. Then, the optimized spectra data were used to establish the discrimination model based on PCA-MD. The results showed that the best spectral pretreatment of PCA-MD model was 1st Der+SG (11 : 3). The cumulative accounting was 88. 2%, when extracted the first five principal components. The first three principal components were selected for establishing the 3D scattered plot of PCA-DA model. It is obvious that samples with different processed methods could be grouped completely. The clustering result of P. yunnanensis I, H, G and F were better than others, and the first three (I, H and G) were nearer. It indicated that the chemical composition of processing by sun -drying and oven drying were similar to traditional processing method. Additionally, P. yunnanensis D was close to P. yunnanensis E, it conjectured that chemical compositions of processing by microwave drying and steam treatment were similar. The prediction set could accurately conform to the calibration set, and the accuracy of PCA-MD model was 100%. Infrared spectroscopy combined with PCA-MD could distinguish different processed P. yunnanensis accurately. Furthermore, it could provide references for clinical application, discriminating of processed P. yunnanensis as well as other processed TCM.