论文部分内容阅读
籽粒直链淀粉质量分数的快速、无损伤测定是高直链淀粉玉米育种的关键环节。以196份高直链单籽粒和360份高直链单穗玉米籽粒为样本,分别利用碘染色法和一阶导数+标准正态变量变换(SNV)的光谱预处理法,构建单籽粒和单穗籽粒直链淀粉质量分数的NIRS分析模型,并通过分割建模样品化学值变异范围的方法建立2个单穗籽粒的NIRS子模型,以期提高对单穗籽粒样品的预测准确度。结果表明,所建立的4个模型的交叉验证标准偏差(RMSECV)分别为1.805、3.370、2.394、2.408,预测标准偏差(RMSEP)分别为2.017、3.205、2.369、2.596,各项决定系数(R2cal、R2cv、R2val)为0.626 1~0.897 0。表明,所建玉米单籽粒NIRS模型的预测准确度较高,可用于早代玉米单籽粒直链淀粉质量分数的鉴定;单穗NIRS子模型能够在一定程度上弥补单穗NIRS模型在预测准确度上的不足,将总模型与子模型配合使用能提高预测准确度。
Rapid, non-destructive determination of amylose content in grain is a key step in high-amylose corn breeding. A total of 196 high-linear single grain and 360 high-linear single-ear maize kernels were used as samples. Iodine staining and first-order derivative + standard normal-variant transformation (SNV) NIRS model of starch mass fraction was established. Two NIRS sub-models of single-spike grain were established by dividing the range of chemical value of the model samples, in order to improve the prediction accuracy of single-spike grain samples. The results showed that the RMSECV of the four models were 1.805,3.370,2.394,2.408 respectively, RMSEP were 2.017,3.205,2.369,2.596, R2cal, R2cv, R2val) is 0.626 1 to 0.897 0. The results showed that the NIRS model of maize single grain had higher prediction accuracy and could be used to identify the amylose content of single grain early maize. The single-ear NIRS sub-model could make up for the single-ear NIRS model to predict the accuracy On the deficiencies, the use of the total model with the sub-model can improve the prediction accuracy.