Weir B S
Genetics. 1976 Nov;84(3):639-59. doi: 10.1093/genetics/84.3.639.
The statistical assessment of gene-frequency data on protein polymorphisms in natural populations remains a contentious issue. Here we formulate a test of whether polymorphisms detected by electrophoresis are in accordance with the stepwise, or charge-state, model of mutation in finite populations in the absence of selection. First, estimates of the model parameters are derived by minimizing chi-square deviations of the observed frequencies of genotypes with alleles (0,1,2...) units apart from their theoretical expected values. Then the remaining deviation is tested under the null hypothesis of neutrality. The procedure was found to be conservative for false rejections in simulation data. We applied the test to Ayala and Tracey 's data on 27 allozymic loci in six populations of Drosophila willistoni . About one-quarter of polymorphic loci showed significant departure from the neutral theory predictions in virtually all populations. A further quarter showed significant departure in some populations. The remaining data showed an acceptable fit to the charge state model. A predominating mode of selection was selection against alleles associated with extreme electrophoretic mobilities. The advantageous properties and the difficulties of the procedure are discussed.
对自然种群中蛋白质多态性的基因频率数据进行统计评估仍然是一个有争议的问题。在此，我们设计了一个检验方法，用于判断通过电泳检测到的多态性是否符合有限种群在无选择情况下的逐步突变模型或电荷状态突变模型。首先，通过最小化观察到的基因型频率与理论预期值之间的卡方偏差来推导模型参数估计值，这些基因型的等位基因相差（0、1、2……）个单位。然后，在中性假设下检验剩余偏差。结果发现，该程序对于模拟数据中的错误拒绝情况较为保守。我们将该检验应用于阿亚拉（Ayala）和特雷西（Tracey）关于威氏果蝇（Drosophila willistoni）六个种群中27个等位酶位点的数据。几乎在所有种群中，约四分之一的多态位点显示出与中性理论预测有显著偏差。另外四分之一在某些种群中显示出显著偏差。其余数据与电荷状态模型拟合良好。主要的选择模式是针对与极端电泳迁移率相关的等位基因进行选择。本文还讨论了该程序的有利特性和困难之处。
