Institute of Chinese Materia Medica, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, PR China; School of Pharmacy, Chengdu University of Traditional Chinese Medicine, Chengdu 611137, PR China.
Hospital of Chengdu University of Traditional Chinese Medicine, Chengdu 610072, PR China.
Spectrochim Acta A Mol Biomol Spectrosc. 2022 Jan 15;265:120325. doi: 10.1016/j.saa.2021.120325. Epub 2021 Aug 28.
This paper mainly focuses on the feasibility of rapidly identifying Fritillariae cirrhosae varieties, distinguishing its authenticity and detecting its components by using a portable near infrared (NIR) spectrometer. Five different varieties of Fritillariae cirrhosae, five common counterfeits and two main components (ethanol-soluble extractives and total alkaloids) were studied. The reference values of ethanol-soluble extractives were determined by hot dip method and the reference value of total alkaloid was determined by ultraviolet-visible spectrophotometry (UV-Vis). Linear discriminant analysis (LDA) algorithm was used to identify the sources of different varieties of Fritillariae cirrhosae and the common counterfeits of Fritillariae cirrhosae, respectively. As a result, the best models seemed to be effective, with accuracy of the two models' prediction sets reaches 83.33% and 90.91%, respectively. The partial least squares regression (PLSR) algorithm was used to relate the sample spectra with the reference values of ethanol-soluble extractives and total alkaloid content. Coefficient of determination of prediction (Rp) and root mean square errors of prediction (RMSEP) obtained were 0.8562 and 0.3911; 0.6917 and 0.0117, for ethanol-soluble extractives and total alkaloid content, respectively. The results showed that the portable NIR spectrometer could evaluate the quality of Fritillariae cirrhosae with high efficiency and practicability.
本文主要研究了利用便携式近红外（NIR）光谱仪快速鉴别贝母品种、鉴别其真伪、检测其成分的可行性。研究了 5 种不同的贝母品种、5 种常见的伪品和 2 种主要成分（醇溶性浸出物和总生物碱）。采用热浸法测定醇溶性浸出物的参考值，采用紫外-可见分光光度法（UV-Vis）测定总生物碱的参考值。分别采用线性判别分析（LDA）算法和偏最小二乘回归（PLSR）算法对不同贝母品种和常见伪品的来源进行识别和定量分析。结果表明，两个模型的预测集准确率分别达到 83.33%和 90.91%，最佳模型似乎有效。PLSR 算法分别与醇溶性浸出物和总生物碱含量的参考值相关联，得到预测的决定系数（Rp）和预测均方根误差（RMSEP）分别为 0.8562 和 0.3911；0.6917 和 0.0117。结果表明，便携式 NIR 光谱仪可高效、实用地评价贝母的质量。
