论文部分内容阅读
本研究的目的是测定以下稳定估测因子的重复性：回归系数（ｂｉ）、去平均回归的回归系数（｜ｂｉ－１｜）、离回均均方（Ｓｄｉ２）、Ｓｈｕｋｌａ稳定变异（σｉ２）、基因型平均变异（Ｓｉ２）、基因型变异系数（ＣＶｉ）以及决定系数（ｒｉ２）和平均产量（ｘｉ）。这些统计数据来自于Ｌｏｕｉｓｉａｎａ州农业试验站３６个环境下种植的３套冬小麦产量表现性状的资料。重复性用Ｓｐｅａｒｍａｎ的序相关系数和Ｋｅｎｄａｌｌ的一致系数来估算。结果表明，（｜ｂｉ－１｜）、σｉ２和Ｓｄｉ２在任何两个亚环境间是不可重复的，Ｓｉ２和ｒｉ２的重复性很低。所有稳定性估算因子中，只有ｂＩＣＶｉ在各种亚环境单位间均是可重复的。ＣＶｉ由于受ｘｉ排序的影响对描述基因型稳定性来说是一个不可靠的统计参数。平均产量是最可重复的基因型特性。选择产量稳定性中的增益可通过ｂｉ和ｘｉ综合运用得以估计出来。若资料不适于线性回归模式，那么较低的Ｓｉ２（１型）或４型变异值（所有可预知环境中平均的非预知环境基因型平均变异值）可在产量稳定性选择中作为一个选择标准。
The purpose of this study was to determine the reproducibility of the following stable estimators: regression coefficient (bi), regression coefficient (bi-1 |), mean square deviation (Sdi2), and stable Shukla mutation (σi2) , Average genotype variation (Si2), coefficient of variation (CVi), and coefficient of determination (ri2) and average yield (xi). These statistics are based on data from three sets of winter wheat yield traits planted in 36 environments at the Louisiana State Agricultural Experiment Station. Repeatability is estimated using Spearman’s ordinal correlation coefficient and Kendall’s consensus coefficient. The results show that (| bi-1 |), σi2 and Sdi2 are non-repeatable between any two sub-environments, and the repeatability of Si2 and ri2 is very low. Among all the stability estimation factors, only bICVi is reproducible across various sub-environmental units. CVi is an unreliable statistical parameter for characterizing genotype stability due to the influence of xi ordering. The average yield is the most reproducible genotype. The gain in the choice of yield stability can be estimated by combining bi and xi. If data are unsuitable for linear regression models, lower values of Si2 (type 1) or type 4 variability (average of all unpredictable environmental genotypes in predictable environments) can be used as a selection criterion in yield stability selection .