基于近红外光谱技术绿色快速预测前胡中白花前胡甲素含量  

Rapid prediction of praeruptorin A in Peucedani Radix based on NIRS

作　　者：贾彬 王芳 宋向文[2,3] 陈抒语 罗红尘 颜辉 陈存武[2,3] 韩邦兴[2,3] JIA Bin;WANG Fang;SONG Xiangwen;CHEN Yuyu;LUO Hongchen;YAN Hui;CHEN Cunwu;HAN Bangxing(College of Pharmacy,Anhui University of Chinese Medicine,Hefei Anhui 230031;College of Biological and Pharmaceutical Engineering,West Anhui University,Lu′an Anhui 237012;Anhui Engineering Laboratory of Conservation and Sustainable Utilization of Traditional Chinese Medicine Resources,Lu′an Anhui 237012;College of Biotechnolgy,Jiangsu University of Science and Technology,Zhenjiang Jiangsu 212003,China)

机构地区：[1]安徽中医药大学药学院,安徽合肥230031 [2]皖西学院生物与制药工程学院,安徽六安237012 [3]安徽省中药资源保护与持续利用工程实验室,安徽六安237012 [4]江苏科技大学生物技术学院,江苏镇江212003

出　　处：《蚌埠医学院学报》2024年第4期510-514,共5页Journal of Bengbu Medical College

基　　金：国家重点研发计划(2023YFC3503804);财政部和农业农村部国家现代化农业产业技术体系建设专项(CARS-21);安徽省自然科学杰出青年项目(1808085J17);安徽省大别山中医药研究院开放课题(TCMADM-2023-04)。

摘　　要：目的:基于近红外光谱技术(NIRS)建立前胡中白花前胡甲素的定量模型,实现对前胡质量的绿色快速评价。方法:运用高效液相色谱法(HPLC法)测定130份前胡样品的白花前胡甲素含量并采集其漫反射NIRS光谱,采用标准正态变换(SNV)法预处理光谱,并基于最小偏二乘回归法(PLS)建立前胡中白花前胡甲素的定量校正模型。结果:此模型校正集的均方根误差和相关系数分别为0.1047和0.9380,交叉验证的均方根误差和相关系数分别为0.1230和0.9180,预测集的均方根误差和相关系数分别为0.1251和0.8815。从此模型的校正集、预测集以及交叉验证结果来看,模型具有很高的预测精度。结论:此研究使用NIRS技术所建前胡的定量模型稳定可靠,可实现对市场前胡的质量的绿色快速鉴别,并为维护消费者利益和临床使用安全性,提供了科学依据。Objective:Near infrared spectroscopy(NIRS)was used to establish the quantitative detection model of the praeruptorin A in Peucedani Radix,so as to realize the rapid evaluation of the quality of Peucedani Radix.Methods:The sum of praeruptorin A in 130 Peucedani Radix samples were determined by high-performance liquid chromatography(HPLC),and the NIRS diagram was collected.The quantitative correction model of praeruptorin A in Peucedani Radix was established by least partial multiplication(PLS).Results:The root mean square error and correlation coefficient of the calibration set of this model were 0.1047 and 0.9380,respectively,the root mean square error and correlation coefficient of cross validation were 0.1230 and 0.9180,respectively.The root mean square error and correlation coefficient of the prediction set were 0.1251 and 0.8815,respectively.From the correction set,prediction set and cross validation results of praeruptorin A content models,the model had high prediction accuracy.Conclusions:The quantitative detection model of Peucedani Radix established by using NIRS is stable and reliable,which can realize the rapid eevaluation of the quality of Peucedani Radix in the market,and provide scientific basis for safeguarding consumers′interests and clinical safety.

关 键 词：前胡 白花前胡甲素 高效液相色谱法 近红外光谱法 

分 类 号：R927[医药卫生—药学]