利用近红外技术预测向日葵籽仁品质性状

Prediction of qualitative characteristics of sunflower husked seed by near infrared spectroscopy

为实现向日葵育种材料品质性状的快速预测,选取154份向日葵籽仁样品,结合化学测定值和近红外光谱,利用化学计量学手段建立向日葵籽仁品质指标的近红外模型,评估其在籽仁粗蛋白、粗脂肪、油酸、亚油酸等重要品质性状测定中的可行性。结果表明,改进偏最小二乘法建模效果最佳,其粗脂肪、粗蛋白、油酸、亚油酸、饱和脂肪酸及不饱和脂肪酸的定标相关系数分别为0.975、0.950、0.973、0.951和0.913,交叉验证相关系数分别为0.969、0.939、0.915、0.927和0.711。用检验集对模型进行验证,粗脂肪、蛋白质、油酸、亚油酸、饱和及不饱和脂肪酸的外部检验决定系数(R^2)分别为0.959、0.950、0.937、0.906和0.930。本研究建立的模型质量较高,能够满足向日葵籽仁品质成分的快速测定,可为向日葵品质育种前期大量、快速的筛选提供技术支持。

For rapid and high throughput prediction of the quality of sunflower seeds,calibration equations were established on 154 materials of sunflower husked seeds.Combination methods included near-infrared spec troscopy(NIRS)scanning and chemical detecting with chemometrics method to determine the optimal calibration models.Results showed that the model of modified partial least-squares method was appropriate.The determination coefficient of NIRS model for crude fat,crude protein,oleic acid,linoleic acid,saturated fatty acid and unsaturated fatty acid were 0.975,0.950,0.973,0.951 and 0.913 respectively.Their cross-validation correlation coefficients were 0.969,0.939,0.915,0.927 and 0.711 correspondingly.Validation testing results showed that the external vali dation correlation coefficient for crude fat,crude protein,oleic acid,linoleic acid,saturated fatty acid and unsaturat ed fatty acid were 0.959,0.950,0.937,0.906 and 0.930 respectively.The results indicated that the established NIRS model could be used as a tool for rapid prediction of qualitative characteristics in sunflower husked seed for large-scale screening of sunflower quality breeding.