FTIR结合化学计量学快速预测铁皮石斛中总多糖含量  

Rapid Forecast of Total Polysaccharide Content in Dendrobium officinale by FTIR Combined with Chemometrics

作　　者：曾贞 孙恒[1] 陈骏飞[1] 徐天才[1] 王宇萍[1] 李坤 康平德[1] 胡强[3] 

机构地区：[1]云南省农业科学院高山经济植物研究所,云南丽江674199 [2]云南省农业科学院热区生态农业研究所,云南元谋651300 [3]云南省农业科学院,云南昆明650000

出　　处：《江西农业学报》2018年第1期117-121,共5页Acta Agriculturae Jiangxi

摘　　要：有效成分含量检测是评价药用植物质量的主要手段,采集时间不同对药用植物有效成分含量有显著影响。通过傅里叶变换红外光谱结合化学计量学,建立了快速预测不同采收时间铁皮石斛中总多糖含量的方法。采收2014年1~12月的样品干燥粉碎;以苯酚硫酸法测定铁皮石斛中总多糖含量,分析了不同采收时间铁皮石斛中总多糖随时间的累积规律;采集样品红外光谱信息,归属红外光谱吸收峰,拟合光谱数据和总多糖含量数据,结合一阶导数、二阶导数、标准正态变量、多元散射校正、正交信号校正等对数据进行预处理,建立偏最小二乘回归模型预测样品中总多糖含量。结果显示:(1)样品和标准品葡萄糖均在488 nm附近有共有吸收峰,以488 nm为总多糖定量波长,标准曲线为y=0.0079x+0.027,相关系数R2=0.9995,线性关系良好;精密度、重现性和稳定性相对标准偏差分别为0.2%、0.31%和1.3%,该方法稳定可靠;(2)总多糖含量随时间变化趋势为先升高后降低,1~4月和12月样品含量较高,平均含量大于0.284 g/g;(3)铁皮石斛红外光谱数据与总多糖含量拟合后进行一阶导数、二阶导数、标准正态变量、多元散射校正、正交信号校正等组合处理,用PLSR模型预测铁皮石斛的总多糖含量,结果最佳预处理方式为2D+OSC+MSC+SNV+SG5-PLSR,训练集和验证集R2分别为0.843和0.782,验证均方根误差(RMSEE)和预测均方根误差(RMSEP)分别为2.8904和3.5513,铁皮石斛中总多糖含量预测值与测量值较接近,表明PLSR模型可用于总多糖含量的快速预测。傅里叶变换红外光谱结合化学计量学能实现铁皮石斛中总多糖含量准确预测,为铁皮石斛质量评价提供快速、有效的方法。The detection of active ingredient contents is the main means of evaluating the quality of medicinal plants,and different harvest time has a significant influence on the content of effective components in medicinal plants.In this study,Fourier transform infrared spectroscopy（FTIR） combined with chemometrics was used to establish a method for the rapid forecast of total polysaccharide content in Dendrobium officinale harvested in different periods.The D.officinale samples harvested from January to December of 2014 were dried and crushed,their total polysaccharide contents were determined by using phenol-sulfuric acid method,and the accumulation regularity of total polysaccharides in D.officinale was analyzed.The infrared spectra of these samples were recorded,and their absorption peaks were identified.According to the obtained data of infrared spectra and total polysaccharide content,the original data were preprocessed by simulation combined with the first derivative（1 D）,the second derivative（2 D）,standard normal variable（SNV）,multiplicative scatter correction（MSC）,and orthogonal signal correction（OSC）;then a partial least squares regression（PLSR） model was established to predict the total polysaccharide content in samples.The results showed that：（1） Both the sample and the standard substance glucose had a common absorption peak near 488 nm; thus,in this experiment,the total polysaccharide content was measured at 488 nm.The standard curve equation of total polysaccharidecontent was y = 0.0079 x ＋ 0.027,and its correlation coefficient R2 was 0.9995; the precision,reproducibility and stability（relative standard deviation） of this method were 0.2%,0.31% and 1.3%,respectively,suggesting this method was stable and reliable.（2） Along with the delaying of harvest time,the content of total polysaccharide in samples exhibited the change trend of ＂ rising-decreasing-rising＂; the total polysaccharide content in samples harvested in January,February,March,April and December was

关 键 词：傅里叶变换红外光谱 吸光度 铁皮石斛 多糖 模型 

分 类 号：R284.2[医药卫生—中药学]