论文部分内容阅读
从玉米自交系种子的遗传距离、近红外光谱距离、品种鉴别模型性能三方面进行分析,探索三者间的关系。采用三组(共15对)遗传关系远近不同的玉米自交系种子作为实验材料,通过简单重复序列(simple sequence repeats,SSR)标记计算自交系种子间的遗传距离;种子的近红外光谱经预处理后降维到主成分分析(principal component analysis,PCA)空间中,计算各个自交系种子样本中心点之间的欧氏距离,作为对应的近红外光谱距离;使用仿生模式识别方法建立鉴别模型,用模型的鉴别正确率评价模型的性能。分析结果表明,自交系种子间遗传距离与近红外光谱距离相关性为0.986 8,与模型鉴别正确率的相关性为0.911 0,相关性显著。说明近红外光谱可以反映出玉米自交系种子之间的遗传关系,遗传关系的远近影响品种鉴别模型的性能,遗传距离越小,近红外光谱距离越小,模型鉴别能力也越差。实际应用中有望利用近红外光谱技术分析玉米自交系的遗传关系,对遗传育种、品种识别、纯度分选等具有重要意义;且建立品种鉴别模型时,应充分考虑遗传关系较近的玉米自交系对模型性能的影响。
From three aspects of the genetic distance, near-infrared spectral distance of the seeds of the inbreds of maize, and the performance of the variety identification model, the relationship among the three was explored. Three groups (15 pairs) of maize inbred lines with different genetic relationships were used as experimental materials to calculate the genetic distance between inbred lines by simple sequence repeats (SSR) markers. The near-infrared spectra of seeds The pre-treatment dimension was reduced to the principal component analysis (PCA) space to calculate the Euclidean distance between the center points of the seed samples of each inbred line as the corresponding near-infrared spectral distance. Bionic pattern recognition was used to establish the identification Model, the correctness of the model is used to evaluate the performance of the model. The results showed that there was a significant correlation between genetic distance and NIR spectra distance of inbred lines of 0.986 8, and the correlation with the accuracy of model identification was 0.911 0. It is concluded that near-infrared spectroscopy can reflect the genetic relationship between seeds of maize inbred lines. The genetic relationship between the near and far influences the performance of the variety identification model. The smaller the genetic distance and the smaller the near-infrared spectral distance, the worse the model discrimination ability. In practical applications, it is expected to analyze the genetic relationship of maize inbred lines by near-infrared spectroscopy, which is of great significance for genetic breeding, variety identification and purity sorting. When establishing the identification model of varieties, the genetic relationship between maize The Impact of Crossing on the Performance of the Model.