Department of Biology, University of Oregon, Eugene, OR, USA.
Institute of Ecology and Evolution, University of Oregon, Eugene, OR, USA.
Evolution. 2022 Jul;76(7):1625-1637. doi: 10.1111/evo.14515. Epub 2022 May 24.
Numerous phylogenetic studies reported the existence of a pervasive scaling relationship between the ages of extant eukaryotic clades and their estimated diversification rates. The causes of this age-rate-scaling (ARS), whether biological and/or artifactual, remain unresolved. Here we fit diversification models to thousands of eukaryotic time-calibrated phylogenies to explore multiple potential causes of the ARS including parameter non-identifiability, model inadequacy, biases in taxonomic practice, and an important and ubiquitous form of sampling bias-preferentially analyzing larger extant clades. We distinguish between two mechanism by which such sampling biases can cause an ARS: First, by favoring clades that happen to be unusually large merely by chance (i.e., due to the stochastic nature of the cladogenic process), thus leading to rate overestimation, and second, by favoring clades that have truly higher diversification rates. We find that, of the proposed explanations, only sampling biases are likely to contribute to the observed ARS. We develop methods for fully correcting for sampling bias mechanism 1, and find that despite these corrections a substantial ARS remains. We then confirm using simulations that preferring trees with truly higher rates (mechanism 2) likely explains this residual ARS. Since we do not have a completely unbiased sample of clades, including extinct ones, for phylogenetic analyses, it is difficult to demonstrate unambiguously that sampling biases are the sole cause of the ARS. Sampling biases are, however, a parsimonious and plausible explanation for this widely observed macroevolutionary pattern, and this has implications for how we interpret the distribution of diversification rate estimates in extant clades.
大量系统发育研究报告称，现存真核生物类群的年龄与其估计的多样化率之间存在普遍的缩放关系。这种年龄-速率缩放（ARS）的原因，无论是生物的还是人为的，仍然没有解决。在这里，我们拟合了数千个真核生物时间校准系统发育树的多样化模型，以探索 ARS 的多种潜在原因，包括参数不可识别性、模型不足、分类实践中的偏差，以及一种重要且普遍存在的采样偏差形式——优先分析更大的现存类群。我们区分了两种机制，即采样偏差如何导致 ARS：首先，通过偏爱碰巧由于随机性质的有丝分裂过程而异常大的类群（即，由于随机性质的有丝分裂过程），从而导致高估率，其次，通过偏爱真正具有更高多样化率的类群。我们发现，在所提出的解释中，只有采样偏差可能导致观察到的 ARS。我们开发了用于完全纠正采样偏差机制 1 的方法，并发现尽管进行了这些校正，仍然存在大量的 ARS。然后，我们使用模拟确认，偏爱真正具有更高速率的树（机制 2）可能解释了这种剩余的 ARS。由于我们没有进行系统发育分析的类群的完全无偏样本，包括已灭绝的类群，因此很难明确证明采样偏差是 ARS 的唯一原因。然而，采样偏差是对这种广泛观察到的宏观进化模式的一种简单而合理的解释，这对我们如何解释现存类群中多样化率估计的分布具有影响。
