Laboratoire d'Ecologie Alpine, CNRS UMR 5553, Université Joseph Fourier, Grenoble, France.
Mol Ecol. 2011 Aug;20(15):3144-55. doi: 10.1111/j.1365-294X.2011.05164.x. Epub 2011 Jun 20.
A plant species' genetic population structure is the result of a complex combination of its life history, ecological preferences, position in the ecosystem and historical factors. As a result, many different statistical methods exist that measure different aspects of species' genetic structure. However, little is known about how these methods are interrelated and how they are related to a species' ecology and life history. In this study, we used the IntraBioDiv amplified fragment length polymorphisms data set from 27 high-alpine species to calculate eight genetic summary statistics that we jointly correlate to a set of six ecological and life-history traits. We found that there is a large amount of redundancy among the calculated summary statistics and that there is a significant association with the matrix of species traits. In a multivariate analysis, two main aspects of population structure were visible among the 27 species. The first aspect is related to the species' dispersal capacities and the second is most likely related to the species' postglacial recolonization of the Alps. Furthermore, we found that some summary statistics, most importantly Mantel's r and Jost's D, show different behaviour than expected based on theory. We therefore advise caution in drawing too strong conclusions from these statistics.
植物物种的遗传种群结构是其生活史、生态偏好、在生态系统中的位置和历史因素复杂组合的结果。因此，存在许多不同的统计方法可以衡量物种遗传结构的不同方面。然而，对于这些方法如何相互关联以及它们与物种的生态学和生活史有何关联，我们知之甚少。在这项研究中，我们使用了来自 27 种高山物种的 IntraBioDiv 扩增片段长度多态性数据集，计算了 8 个遗传综合统计量，这些统计量与一组 6 种生态和生活史特征进行了联合相关分析。我们发现，计算出的综合统计量之间存在大量冗余，并且与物种特征矩阵之间存在显著关联。在多元分析中，27 种物种中存在两个主要的种群结构方面。第一个方面与物种的扩散能力有关，第二个方面很可能与物种在阿尔卑斯山的后冰川再殖民化有关。此外，我们发现一些综合统计量，尤其是 Mantel 的 r 和 Jost 的 D，表现出与理论预期不同的行为。因此，我们建议在从这些统计数据中得出过于强硬的结论时要谨慎。
