Data from: Estimating diversification rates for higher taxa: BAMM can give problematic estimates of rates and rate shifts

Estimates of diversification rates are invaluable for many macroevolutionary studies. Recently, an approach called BAMM (Bayesian Analysis of Macro-evolutionary Mixtures) has become widely used for estimating diversification rates and rate shifts. At the same time, several papers have concluded that estimates of net diversification rates from the method-of-moments (MS) estimators are inaccurate. Yet, no studies have compared the ability of these two methods to accurately estimate clade diversification rates. Here, we use simulations to compare their performance. We found that BAMM yielded relatively weak relationships between true and estimated diversification rates. This occurred because BAMM underestimated the number of rates shifts across each tree, and assigned high rates to small clades with low rates. Errors in both speciation and extinction rates contributed to these errors, showing that using BAMM to estimate only speciation rates is also problematic. In contrast, the MS estimators (particularly using stem group ages), yielded stronger relationships between true and estimated diversification rates, by roughly twofold. Furthermore, the MS approach remained relatively accurate when diversification rates were heterogeneous within clades, despite the widespread assumption that it requires constant rates within clades. Overall, we caution that BAMM may be problematic for estimating diversification rates and rate shifts.分化速率（diversification rates）的估计结果对诸多宏观进化研究而言至关重要。近年来，一种名为BAMM（贝叶斯宏观进化混合分析，Bayesian Analysis of Macro-evolutionary Mixtures）的方法被广泛用于分化速率与速率转变的估计。与此同时，已有多项研究指出，基于矩量法（method-of-moments，MS）估计量得到的净分化速率估计结果并不准确。然而，目前尚无研究对比这两种方法在准确估计支系（clade）分化速率方面的性能差异。本研究通过模拟实验对比二者的表现。研究发现，BAMM得到的真实分化速率与估计分化速率之间的相关性相对较弱。这一现象的成因在于，BAMM低估了每棵系统发育树中的速率转变次数，且将高速率赋值给了实际速率偏低的小型支系。物种形成速率与灭绝速率的误差共同引发了上述问题，这表明仅通过BAMM估计物种形成速率同样存在缺陷。与之形成鲜明对比的是，矩量法估计量（尤其是采用茎群年龄（stem group ages）的方案）得到的真实与估计分化速率间的相关性更强，提升幅度约为两倍。此外，尽管学界普遍认为矩量法要求支系内部的分化速率保持恒定，但当支系内部存在分化速率异质性时，该方法仍能保持相对较高的准确性。综上，本研究提请学界注意，使用BAMM进行分化速率与速率转变的估计可能存在诸多问题。