Turjak Martin, Trontelj Peter
Department of Biology, Biotechnical faculty, University of Ljubljana, Večna Pot 111, SI-1000 Ljubljana, Slovenia.
Cladistics. 2012 Dec;28(6):627-638. doi: 10.1111/j.1096-0031.2012.00403.x. Epub 2012 May 22.
Synapomorphies are fundamental to phylogenetic systematics as they offer empirical evidence of monophyletic groups. However, no method exists to directly measure synapomorphy. Here, we propose a method that quantifies synapomorphy using the pattern of character state distribution over a cladogram separately for each character and for each clade. We define a fully synapomorphic character state as one shared by all of a clade's terminal taxa and at the same time completely absent from all terminal taxa outside that clade. The extent to which this condition is met corresponds to the support for the character state being synapomorphic or, in short, support for synapomorphy. It is calculated as the probability of randomly selecting, by multi-stage sampling following the topology of the tree, two terminals from inside a clade sharing the same character state and one terminal from outside the clade bearing a different character state. The method is independent of tree inference and free of transformational assumptions, and so can be applied to any tree and used for any type of discrete character. By measuring synapomorphy, the method offers a potential tool for determining diagnostic character states for taxa on different hierarchical levels, for evaluating alternative systems of character coding, and for evaluating clade support. We show how the method differs from ancestral character state reconstruction methods and goodness-of-fit indices. We demonstrate the behaviour of our method with several hypothetical scenarios and its potential use with two real-life examples.
共衍征对于系统发育系统学至关重要，因为它们提供了单系类群的实证证据。然而，目前还没有直接测量共衍征的方法。在此，我们提出一种方法，该方法通过分别针对每个性状和每个分支，利用分支图上性状状态分布模式来量化共衍征。我们将完全共衍征性状状态定义为一个分支的所有终端分类单元所共有的，同时在该分支之外的所有终端分类单元中完全不存在的性状状态。满足这一条件的程度对应于对该性状状态为共衍征的支持度，简而言之，就是对共衍征的支持度。它是通过按照树的拓扑结构进行多阶段抽样，随机从一个分支内部选择两个具有相同性状状态的终端，以及从该分支外部选择一个具有不同性状状态的终端的概率来计算的。该方法独立于树的推断，且没有转换假设，因此可应用于任何树，并用于任何类型的离散性状。通过测量共衍征，该方法为确定不同层次水平上分类单元的诊断性状状态、评估性状编码的替代系统以及评估分支支持度提供了一种潜在工具。我们展示了该方法与祖先性状状态重建方法和拟合优度指数的不同之处。我们用几个假设情景展示了我们方法的行为，并通过两个实际例子展示了其潜在用途。
