Koizumi Itsuro, Yamamoto Shoichiro, Maekawa Koji
Department of Biological and Environmental Sciences, PO Box 65 (Viikinkaari 1), FIN-00014, University of Helsinki, Finland. itsuro@
Mol Ecol. 2006 Oct;15(11):3175-89. doi: 10.1111/j.1365-294X.2006.03019.x.
Isolation by distance is usually tested by the correlation of genetic and geographic distances separating all pairwise populations' combinations. However, this method can be significantly biased by only a few highly diverged populations and lose the information of individual population. To detect outlier populations and investigate the relative strengths of gene flow and genetic drift for each population, we propose a decomposed pairwise regression analysis. This analysis was applied to the well-described one-dimensional stepping-stone system of stream-dwelling Dolly Varden charr (Salvelinus malma). When genetic and geographic distances were plotted for all pairs of 17 tributary populations, the correlation was significant but weak (r(2) = 0.184). Seven outlier populations were determined based on the systematic bias of the regression residuals, followed by Akaike's information criteria. The best model, 10 populations included, showed a strong pattern of isolation by distance (r(2) = 0.758), suggesting equilibrium between gene flow and genetic drift in these populations. Each outlier population was also analysed by plotting pairwise genetic and geographic distances against the 10 nonoutlier populations, and categorized into one of the three patterns: strong genetic drift, genetic drift with a limited gene flow and a high level of gene flow. These classifications were generally consistent with a priori predictions for each population (physical barrier, population size, anthropogenic impacts). Combined the genetic analysis with field observations, Dolly Varden in this river appeared to form a mainland-island or source-sink metapopulation structure. The generality of the method will merit many types of spatial genetic analyses.
距离隔离通常通过对所有成对种群组合的遗传距离和地理距离的相关性进行检验。然而，这种方法可能会因少数高度分化的种群而产生显著偏差，并丢失单个种群的信息。为了检测异常种群并研究每个种群基因流和遗传漂变的相对强度，我们提出了一种分解成对回归分析方法。该分析方法应用于对溪流栖息的花羔红点鲑（Salvelinus malma）描述详尽的一维 stepping-stone 系统。当绘制 17 个支流种群所有成对种群的遗传距离和地理距离时，相关性显著但较弱（r(2) = 0.184）。基于回归残差的系统偏差并遵循赤池信息准则，确定了 7 个异常种群。包含 10 个种群的最佳模型显示出强烈的距离隔离模式（r(2) = 0.758），表明这些种群中基因流和遗传漂变之间达到了平衡。还通过将成对遗传距离和地理距离与 10 个非异常种群进行绘图分析每个异常种群，并将其归类为三种模式之一：强遗传漂变、有限基因流的遗传漂变和高水平基因流。这些分类通常与每个种群的先验预测（物理屏障、种群大小、人为影响）一致。将遗传分析与实地观察相结合，这条河流中的花羔红点鲑似乎形成了大陆 - 岛屿或源 - 汇集合种群结构。该方法的通用性将有利于多种类型的空间遗传分析。
