大豆主要脂肪酸含量近红外模型的建立

为提高大豆脂肪酸品种的育种进程,使用气相色谱法测量289份大豆籽粒样品的脂肪酸含量,并用近红外谷物品质分析仪NIRFIEX N-500采集大豆样品近红外光谱值,采用主成分回归法将化学试验测定的数值与采集的光谱曲线对应拟合整理分析建立定标模型。每种脂肪酸建立11个定标标准,选取决定系数定标Q值中最高的数值为定标模型。结果显示5种脂肪酸定标模型的适宜Q值分别为:棕榈酸0.823 5、硬脂酸0.854 1、油酸0.819 6、亚油酸0.829 1、亚麻酸0.836 3。最后验证结果表明,近红外法测量值与化学试验测定值误差均在1.5%误差范围之内,证明所建立模型质量较好,具有使用价值,可用于大豆育种材料脂肪酸含量的快速测量。

Establishment of NIRS Model for the Main Fatty Acids in Soybean

In order to improve the efficiency of soybean fatty acid breeding and to speed up the breeding process,the fatty acid content of 289 soybean grain samples measured by GC and NIRFIEX N-500 was used to collect the NIR spectra of soybean samples.Using the principal component regression (principal component regession,referred to as PCR),the calibration model was established by fitting and analyzing the value of the chemical experiment and the corresponding spectral curve.Eleven calibration criteria were established for each fatty acid,and the highest value of calibration coefficient Q was selected as the calibration model.The results showed the fit calibration Q value for the calibration model.The fit Q value of palmitic acid was 0.8235,stearic acid was 0.8541,oleic acid was 0.8196,linoleic acid was 0.8291,and linolenic acid was 0.8363,respectively.Validation results demonstrated that the errors between the near infrared detection and chemical test results were within 1.5％,which well explained the model and could be used to measure the fatty acid contents of soybean in breeding.