张晋 曹晓宁 田翔 刘思辰 秦慧彬 乔治军

摘要 [目的]建立一种藜麦粗蛋白含量快速、无损、简便的测定方法，为藜麦的资源评价和品质育种提供技术支持。[方法]以100份藜麦种质资源为材料，其中80%为校正集，20%为验证集，扫描得到藜麦近红外原始光谱，利用OPUS/QUAN T5.5光谱定量分析软件建立藜麦蛋白质含量的快速检测模型。[结果]采用一阶导数+矢量归一化光谱方法进行预处理，结合化学方法测定数据建立藜麦粗蛋白近红外定量模型，校正和预测效果最好，藜麦粗蛋白近红外定量模型的交叉验证决定系数为0.918 2，外部验证决定系数为0.915 1。[结论]基于近红外光谱法（NIRS）测定藜麦籽粒的蛋白含量是完全可行的。

关键词 藜麦;蛋白;含量;近红外光谱技术

中图分类号 TS210.7 文献标识码 A

文章编号 0517-6611（2021）09-0175-02

doi：10.3969/j.issn.0517-6611.2021.09.047

Abstract [Objectives]In order to provide technical support for quinoa resource evaluation and quality breeding，and establish a fast，non-destructive，and simple method for determining quinoa protein content.[Method]A total of 100 quinoa varieties were selected，which 80% were used as the modeling set and 20% were used as the verification set.We collected 100 the near infrared spectra of quinoa，the OPUS/QUAN T5.5 software were used to preprocess the original near-infrared spectra data after scanning，and established the quantitative prediction model of quinoa kernel protein content.[Result]Through the preprocessing of first derivative and vector normalization spectral method，and set up the near infrared quantitative model of quinoa protein combining with the data of chemical method，whose calibration and prediction effect was best.Cross validation decision coefficient and external validation decision coefficient of protein by near infrared quantitative model were 0.918 2 and 0.915 1.[Conclusion]It is completely feasible to determine the protein content of quinoa grains based on near-infrared spectroscopy （NIRS）.

Key words Quinoa;Protein;Content;Near infrared spectroscopy

藜麥（Chenopodium quinoa Willd.）原产于南美洲安第斯山地区，有 5 000～7 000 年种植历史[1-3]，具有耐寒、抗旱耐逆、耐盐碱等特性[4]。由于其营养的均衡性[5]，被联合国粮农组织（ FAO） 推荐为适宜人类食用的“全营养食品”，在国内受到了越来越多的关注[6-10]。藜麦籽粒蛋白质含量高[11]，溶解性好，容易被人体吸收利用，属于优质蛋白质[12]，富含多种氨基酸，尤其是植物蛋白中所缺乏的赖氨酸、色氨酸等[12-13]。因此蛋白质含量作为评价藜麦品质的重要指标，对藜麦资源评价、育种材料的筛选具有重要作用。蛋白质含量的测定一般使用凯式定氮法、高效液相色谱法[14]，这些方法步骤烦琐、速度慢、费用高、周期长，不适宜批量对藜麦资源品质评价、育种过程中早代材料的快速、无损检测。……