单粒花生蔗糖含量近红外预测模型的建立

Establishment of near-infrared reflectance spectroscopy model for predicting sucrose content of single seed in peanut

蔗糖含量是影响花生口感和风味的重要因素,培育高蔗糖甜味品种已成为食用型花生遗传改良的重要目标。因此,建立单粒花生蔗糖含量的近红外预测模型有助于加快甜花生品种选育进程。本研究选择128份遗传多样性丰富的代表性材料,采集了近红外光谱,利用高效液相色谱-折光指数检测器(HPLC-RID)测得蔗糖含量化学值,并利用偏最小二乘法(PLS)建立了单粒花生蔗糖含量的数学预测模型,其决定系数(R^(2))为0.913,交叉验证根均方差(RMSECV)为0.750。另选用50粒花生种子对预测模型进行外部验证,预测值和化学值的相关系数达0.92,表明本研究建立的模型预测值准确可靠。本研究建立的单粒花生蔗糖含量预测模型可以应用于杂交早期世代育种材料蔗糖含量的选择,也可以应用于高蔗糖材料纯度的筛选和鉴定,为食用型花生品种选育和产业化应用提供技术支撑。

Sucrose content is an important factor to determine the taste and flavor of peanuts.Breeding sweet peanut varieties with high sucrose content has been a main objective for the edible peanut genetic improvement.Developing a near-infrared reflectance spectroscopy model for predicting the sucrose content in single peanut kernel will accelerate the process of sweet peanut breeding.In this study,128 representative materials with abundant genetic diversity were selected,the near-infrared spectral data were collected,and the sucrose content of each seed was determined by high performance liquid chromatography(HPLC)with RID detector.Based on spectral data and chemical value of sucrose content,a calibration model for predicting the sucrose content in single seed with a coefficient of determination(R^(2))of 0.913 and a root mean square error of cross validation(RMSECV)of 0.750,was built up by partial least squares(PLS)method.50 peanut seeds were analyzed by both NIR and HPLC for external validation and the correlation coefficient between the prediction value and the chemical value reached 0.92,indicating that this model could predict sucrose content with adequate accuracy and reliability.This calibration model for sucrose content in single seed could be applied to the selection of high sucrose lines in the early generations of hybrid progenies and in the purity monitoring of seed and raw material,which will be a supportive technique for the edible peanut variety breeding and industrial application.