水稻籽粒直链淀粉含量非破坏性活体测定方法研究

Nondestructive Technique for Determining Amylose Content of Intact Rice Grains

对533份杂交稻和常规稻材料的糙米和精米直链淀粉含量进行化学分析,并结合近红外透射光谱数据,采用改进偏最小二乘法(MPLS)分别建立糙米、精米的直链淀粉含量(AC)预测定标模型,并对其进行内部和外部验证.结果表明,杂交稻糙米及精米模型的定标相关系数(RSQ)分别为0.873和0.922,常规稻糙米及精米模型的RSQ则分别为0.924和0.939,定标标准偏差(SEC)分别为1.100,0.956,1.537,1.547;内部交叉验证预测值和真实值之间的RSQ分别为0.866,0.901,0.892和0.921,外部验证的RSQ分别为0.9506,0.9352,0.9116,0.9180,所建模型的相关性较高,预测值与真实值之间的误差小.常规稻模型可应用于大量育种材料快速、无损的早代筛选,杂交稻模型可用于新组合直链淀粉含量的快速鉴定,促进稻米品质改良,提高育种效率.

The ventional rice, amylose conte was analyzed the regression method of mo nt （AC） of 53 with chemical dified partial 1 3 brown and milled rice samples, including hyb methods. Four calibration models of AC were east square （MPLS） used in combination with rid rice and con- established with the NITS （near- infrared transmission spectroscopy） data. Cross-validation and external validation were made to the predication capacity of the calibration equations. The results indicated that the regression （RSO） of the calibration in hybrid brown and milled rice were 0. 873 and 0. 922 and the standard e evaluate squared rrors for calibration （SEC） were 1. 100 and 0. 956, respectively, while for the conventional rice genotypes, their RSO of were 0. 924 and 0. 939 and the SEC were 1. 537 and 1. 547, respectively. The RSO of internal cross-validation were 0. 866, 0. 901, 0. 892 and 0. 921, and the RSO of external validation were 0. 950 6, 0. 935 2, 0. 911 6 and 0. 918 0, respectively. These results the models established in this study have significant correlation and applied for fast selec relatively low error between near-infrared value and true value and, therefore, can be tion for amylase content in early generations, so as to promote rice quality improve- ment and enhance breeding efficiency.