目的 中药多基原物种间的效用差异性受到历代医家的高度重视，联用傅里叶变换近红外光谱（fourier transform near infrared spectroscopy，FT-NIR）和傅里叶变换中红外光谱（fourier transform mid-infrared spectroscopy，FT-MIR）技术，考察光谱化学表征技术应用于多基原黄连共有物质基础测定和品质辨识的可行性。方法 以黄连的4种基原（黄连Coptis chinensis、三角叶黄连C. deltoidea、峨眉野连C. omeiensis和云南黄连C. teeta）的共有物质基础小檗碱、黄连碱、木兰花碱、非洲防己碱和巴马汀为研究对象，基于光谱矩阵的优化和特征学习算法，串联近红外和中红外光谱特征构建偏最小二乘回归（partial least squares regression，PLSR）和支持向量回归（support vector regression，SVR）光谱化学表征模型以测定药材中共有活性成分的含量，辨识其品质差异。结果SVR模型对小檗碱的测定效果最优，剩余预测偏差（residual predictive deviation，RPD）值高达4.842，模型对黄连碱、巴马汀和木兰花碱含量的预测RPD值均大于2；PCA结果表明所建立的模型能有效鉴别多基原黄连共有成分含量的差异，为其品质辨识提供依据。结论 多元光谱技术串联应用可有效表征多基原黄连药材中共有成分的含量差异，提高多基原中药品质辨识的效率。

Objective The difference in the utility of the multi-source species of traditional Chinese medicine has been paid great attention by the doctors in the past dynasties. In this study, FT-NIR and FT-MIR technologies were combined to investigate the feasibility of applying spectrochemical characterization techniques to the determination and quality identification of the common material basis of multi-source Huanglian (Coptidis Rhizoma). Methods Taking the common material basis of four Coptidis Rhizoma, coptisine, magnoflorine, berberine, columbamine, and palmatine as the research objects, optimization and feature learning algorithms based on spectral matrix, combining near-infrared and mid-infrared spectroscopy characteristics to construct PLSR and SVR spectrochemical characterization models to determine the content of common active ingredients in medicinal materials and identify their quality differences. Results The SVR model had the best effect on the determination of berberine, with an RPD value of 4.842. The predicted RPD values of the model for berberine, palmatine and magnoflorine content were all higher than 2; PCA results showed that the model could effectively identify the differences in the common components of multi-source Coptidis Rhizoma, and provide a basis for its quality identification. Conclusion The serial application of multiple spectrum technology can effectively characterize the difference in the content of the common components of the multi-based original Coptidis Rhizoma, and improve the efficiency of quality identification of the multi-based original Chinese medicine.

国家重要野生植物种质资源共享平台