Direct optimization of unaligned sequence characters provides a natural framework to explore the sensitivity of phylogenetic hypotheses to variation in analytical parameters. Phenotypic data, when combined into such analyses, are typically analyzed with static homology correspondences unlike the dynamic homology sequence data. Static homology characters may be expected to constrain the direct optimization and thus, potentially increase the similarity of phylogenetic hypotheses under different cost sets. However, whether a total‐evidence approach increases the phylogenetic stability or not remains empirically largely unexplored. Here, I studied the impact of static homology data on sensitivity using six empirical data sets composed of several molecular markers and phenotypic data. The inclusion of static homology phenotypic data increased the average stability of phylogenetic hypothesis in five out of the six data sets. To investigate if any static homology characters would have similar effect, the analyses were repeated with randomized phenotypic data, and with one of the molecular markers fixed as static homology characters. These analyses had, on average, almost no effect on the phylogenetic stability, although the randomized phenotypic data sometimes resulted in even higher stability than empirical phenotypic data. The impact was related to the strength of the phylogenetic signal in the phenotypic data: higher average jackknife support of the phenotypic tree correlated with stronger stabilizing effect in the total‐evidence analysis. Phenotypic data with a strong signal made the total‐evidence trees topologically more similar to the phenotypic trees, thus, they constrained the dynamic homology correspondences of the sequence data. Characters that increase phylogenetic stability are particularly valuable for phylogenetic inference. These results indicate an important role and additive value of phenotypic data in increasing the stability of phylogenetic hypotheses in total‐evidence analyses.

中文翻译：

---

静态同源性的表型特征在其他动态同源性特征的直接优化下增加系统发育稳定性

未对齐序列特征的直接优化为探索系统发育假设对分析参数变化的敏感性提供了一个自然框架。与动态同源序列数据不同，表型数据在结合到此类分析中时，通常使用静态同源对应进行分析。静态同源性特征可能会限制直接优化，因此可能会增加不同成本集下系统发育假设的相似性。然而，全证据方法是否会增加系统发育稳定性在很大程度上仍未得到经验探索。在这里，我使用由几个分子标记和表型数据组成的六个经验数据集研究了静态同源数据对灵敏度的影响。包含静态同源表型数据增加了六个数据集中五个数据集中系统发育假设的平均稳定性。为了研究任何静态同源性特征是否会产生类似的效果，使用随机表型数据重复分析，并将分子标记之一固定为静态同源性特征。平均而言，这些分析对系统发育稳定性几乎没有影响，尽管随机表型数据有时会导致比经验表型数据更高的稳定性。这种影响与表型数据中系统发育信号的强度有关：表型树的平均折刀支持度越高，与全证据分析中的稳定作用越强相关。具有强信号的表型数据使全证据树在拓扑上与表型树更加相似，从而限制了序列数据的动态同源对应关系。增加系统发育稳定性的特征对于系统发育推断特别有价值。这些结果表明表型数据在提高全证据分析中系统发育假设的稳定性方面具有重要作用和附加价值。