偏最小二乘法和近红外光谱技术快速测定发酵黄芪茎叶中黄酮含量及抗氧化活性  

Rapid determination of flavonoids and antioxidant activity of fermented Astragalus membranaceus stems and leaves by partial least square method and near infrared spectroscopy technique

作　　者：张燕[1,2] 贾阳 杜涓 刘娜 王园[1,2] 齐景伟 安晓萍 ZHANG Yan;JIA Yang;DU Juan;LIU Na;WANG Yuan;QI Jing-wei;AN Xiao-ping

机构地区：[1]内蒙古农业大学动物科学学院,内蒙古呼和浩特010018 [2]智慧畜牧自治区高等学校重点实验室内蒙古自治区高校智慧畜牧集成攻关大平台、内蒙古自治区草食家畜饲料技术研究中心、国家乳业技术创新中心、奶牛繁育与养殖技术研究中心,内蒙古呼和浩特010018 [3]包头北辰饲料科技股份有限公司,内蒙古包头014010

出　　处：《饲料研究》2024年第18期70-75,共6页Feed Research

基　　金：内蒙古自治区科技重大专项(项目编号:2021ZD0023-3、2021ZD0046-4);国家乳业技术创新中心创新能力建设重点项目(项目编号:2022-科研攻关-2);内蒙古自治区科技计划项目(项目编号:2022YFHH0072)。

摘　　要：试验测定发酵前、后黄芪茎叶黄酮含量和抗氧化活性,并应用近红外光谱技术(NIR)结合化学计量学法快速预测发酵黄芪茎叶黄酮含量,为定量检测发酵黄芪茎叶黄酮含量提供理论依据以及为黄芪茎叶深加工利用提供技术支持。试验采用硝酸铝-亚硝酸钠比色法测定发酵黄芪茎叶黄酮的含量,测定黄酮对1,1-二苯基-2-三硝基苯肼(DPPH)自由基清除能力和羟基自由基的清除能力,利用偏最小二乘方法构建发酵黄芪茎叶黄酮的快速预测模型。结果表明,发酵黄芪茎叶黄酮抗氧化能力强于发酵前,构建快速预测模型采用了二阶导数(SD)+正态变量变换(SNV)+去趋势算法(Detred)的方法对全谱图进行预处理的效果较好,优化后的模型决定系数(R^(2))、校正均方根误差(RMSEC)、校准标准差(SEC)分别为0.92、0.24、0.24,其相对分析误差(PRD)为3.46,外部验证R2为0.88。研究表明,构建的NIR模型校正和交互验证决定系数均较大,PRD均大于2,说明模型预测性能较好,建立的模型有助于检测发酵黄芪茎叶黄酮含量活性成分。The flavonoid content and antioxidant activity of Astragalus membranaceus stems and leaves before and after fermentation were determined,and the near infrared spectroscopy(NIR)combined with stoichiometry was further applied to rapidly predict the flavonoid content of Astragalus membranaceus stems and leaves in fermentation,in order to provide theoretical basis for quantitative detection of flavonoid content of Astragalus membranaceus stems and leaves and provide technical support for deep processing and utilization of Astragalus membranaceus stems and leaves.The experiment used the aluminum nitrate-nitrite colorimetric method to determine the content of flavonoids in fermented Astragalus membranaceus stems and leaves,measured the 1,1-diphenyl-2-picrylhydrazyl(DPPH)free radical scavenging ability and hydroxyl radical scavenging ability of flavonoids,and constructed a rapid prediction model for flavonoids in fermented Astragalus membranaceus stems and leaves using the partial least squares method.The results showed that the antioxidant activity of Astragalus membranaceus stems and leaves flavonoids after fermentation was stronger than that before fermentation.The method of second derivative(SD)+normal variable transformation(SNV)+detrend algorithm(Detred)was used to construct the fast prediction model,and the effect of pre-processing the whole spectrum was good.After optimization,the coefficient of determination(R^(2)),corrected root mean square error(RMSEC)and calibration standard deviation(SEC)were 0.92,0.24,and 0.24,respectively,and the relative analysis error(PRD)was 3.46.The coefficient of determination for external R2 was 0.88.The study indicates that the NIR model established in this study has a large coefficient of determination for correction and cross-validation,and the PRD is greater than 2,indicating that the model has good prediction performance,and the established model is helpful for detecting the active components of flavonoid content in fermented Astragalus membranaceus stems and leaves.

关 键 词：发酵黄芪茎叶 黄酮 抗氧化性 近红外光谱技术 偏最小二乘法 

分 类 号：S816.7[农业科学—饲料科学]