傅立叶近红外漫反射光谱技术在花生脂肪酸分析中的应用

FT-NIR Diffuse Reflectance Spectroscopy to Predicate Main Fatty Acids in Intact Peanut Seeds

选取高油酸花生品系与普通油酸含量花生品种搭配的4个杂交组合共计987份F2种子,应用近红外反射光谱技术,结合偏最小二乘法,采用检验集检验,成功构建了花生油酸、亚油酸、棕榈酸、4种有害脂肪酸（棕榈酸、花生酸、山嵛酸、二十四碳烷酸）、碘值（IV）和不饱和脂肪酸/饱和脂肪酸（U/S）等6个近红外模型。各模型决定系数（R2）分别达到94.67、95.72、86.36、83.71、94.90和73.53,预测根均方差（RMSEP）分别为2.52、1.91、0.60、0.67、1.57和0.27。各模型预测偏差分别为-4.399-4.838、-2.011-1.874、-1.247-1.438、-1.634-1.420、-2.231-3.733、-0.533-1.396,预测相对误差分别为0.562-9.687、0.055-7.010、0.642-12.72、0.636-11.464、0.217-4.145、1.582-17.934。上述模型可用于花生种子脂肪酸快速分析预测,在花生脂肪酸品质育种、高油酸花生种子生产和原料花生质量控制中具有重要价值。

A total of 987 F2 seeds from 4 crosses involving 2 high-oleic and 3 normal-oleic peanut parental lines/cultivars were exploited in the development of near infrared spectroscopy calibration equations predictive of oleic, linoleic, palmitic acids and 4 harmful fatty acids （palmitic, arachidic, behenic and Lignoceric acid）, IV （iodine value） and U/S （ratio of unsaturated to saturated fatty acids） using Partial Least Square （PLS） algorithm and external validation. R2 was 94.67, 95.72, 86.36, 83.71, 94.90 and 73.53, and RMSECP was 2.52, 1.91, 0.60, 0.67, 1.57 and 0.27, respectively. The predictive error was -4. 399-4. 838, -2. 011-1. 874, -1.247-1. 438, --1. 634-1.420, --2.231- 3. 733 and -0. 533-1. 396, and predictive relative error was 0. 562-9. 687, 0. 055-7. 010, 0. 642-12.72, 0. 636-11. 464, 0. 217-4. 145 and 1. 582-17. 934 respectively. The calibration equations constructed can be used for rapid predication of main fatty acids in intact peanut seeds, and therefore are of value in peanut breeding for improved fatty acid composition, and in quality control in high-oleic peanut seed production and raw material for peanut oiL/food processors.