Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA, USA.
Institute for Quantitative Social Science, Harvard University, Cambridge, MA, USA.
Mol Ecol. 2018 Mar;27(5):1296-1308. doi: 10.1111/mec.14520. Epub 2018 Feb 27.
Ecologists are increasingly making use of molecular phylogenies, especially in the fields of community ecology and conservation. However, these phylogenies are often used without full appreciation of their underlying assumptions and uncertainties. A frequent practice in ecological studies is inferring a phylogeny with molecular data from taxa only within the community of interest. These "inferred community phylogenies" are inherently biased in their taxon sampling. Despite the importance of comprehensive sampling in constructing phylogenies, the implications of using inferred community phylogenies in ecological studies have not been examined. Here, we evaluate how taxon sampling affects the quantification and comparison of community phylogenetic diversity using both simulated and empirical data sets. We demonstrate that inferred community trees greatly underestimate phylogenetic diversity and that the probability of incorrectly ranking community diversity can reach up to 25%, depending on the dating methods employed. We argue that to reach reliable conclusions, ecological studies must improve their taxon sampling and generate the best phylogeny possible.
生态学家越来越多地利用分子系统发育，特别是在群落生态学和保护领域。然而，这些系统发育往往在不完全了解其潜在假设和不确定性的情况下使用。生态研究中的一个常见做法是仅从感兴趣的群落中的分类单元推断具有分子数据的系统发育。这些“推断的群落系统发育”在其分类单元采样中固有偏见。尽管在构建系统发育时全面采样很重要，但在生态研究中使用推断的群落系统发育的影响尚未得到检验。在这里，我们使用模拟和实证数据集评估分类单元采样如何影响群落系统发育多样性的量化和比较。我们证明，推断的群落树大大低估了系统发育多样性，并且根据所使用的定年方法，错误地对群落多样性进行排名的概率可能高达 25%。我们认为，为了得出可靠的结论，生态研究必须改进其分类单元采样并生成最佳的系统发育。
