Alternative treatments of genetic distances for species delimitation in Callosciurinae and Sciurinae (Rodentia: Sciuridae).
Alternative treatments of genetic distances for species delimitation in Callosciurinae and Sciurinae (Rodentia: Sciuridae).
复制标题
DOI：
10.1016/j.gene.2019.03.048
复制
发表时间：
2019-06
期刊：
影响因子：
3.5
通讯作者：
Wei Zhang;Fang Li;Nathan Whiting-Wagner;Yuchun Li
中科院分区：
文献类型：
--
作者：
Wei Zhang;Fang Li;Nathan Whiting-Wagner;Yuchun Li
有许多可用的 DNA 分类方法，但单基因座数据与基于距离的分析相结合仍然在实践中主导物种鉴定。遗传距离阈值通常用于将遗传多样性分配给与物种相对应的实体，但因其非普遍性或任意性而受到批评。本研究研究了一种确定遗传阈值的替代方法，该方法使用迭代姐妹进化枝聚类在系统发育框架内进行姐妹物种边界检测。该方法使用 cytbgene 分别应用于啮齿类动物 Callosciurinae 和 Sciurinae 的两个近亲亚科。我们的结果表明，Callosciurinae 和 Sciurinae 的遗传阈值分别为 4.0% 和 6.0%，表明优化的阈值可能是谱系特异性的。使用这些分类单元划分阈值得到了与流行的 ABGD 方法非常相似的推定物种数，并且 Callosciurinae 和 Sciurinae 的物种多样性分别增加了 74.2% 和 20.7%。这表明存在其他需要进一步调查的神秘物种。我们通过模拟树构建和采样限制的影响进一步测试了我们方法的性能限制。结果表明，它对于不同的树木表现同样良好，但当种间或种内采样不足时则无法发挥作用。这些发现支持当大型数据集仅提供单位点信息时，该方法作为物种描述的替代工具的可行性。
There are many DNA taxonomy methods available, but single-locus data coupled with distance-based analyses still dominate species identification in practice. Genetic distance thresholds are often used for assigning genetic diversity into entities corresponding to species but are criticized for non-universality or arbitrary. This study investigated an alternative approach for determining genetic thresholds that used iteratively sister-clade clustering for sister species boundary detection within a phylogenetic framework. This method was separately applied to two close subfamilies of Callosciurinae and Sciurinae in Rodentia using the cytbgene. Our results showed that genetic thresholds for Callosciurinae and Sciurinae were 4.0% and 6.0%, respectively, indicating that the optimized thresholds could be lineage-specific. The use of these thresholds for taxa partitioning yielded a very similar putative species number as the prevailing ABGD method, and increased species diversity by 74.2% and 20.7% in Callosciurinae and Sciurinae, respectively. This suggested that additional cryptic species were present that warrant further investigation. We further tested the performance limitation of our method by simulating the impacts of tree construction and sampling limitation. The results showed that it preformed equally well for different trees but failed to work when inter- or intraspecific sampling is insufficient. These findings support the feasibility of this approach as an alternative tool for species delineation when only single-locus information is available for large datasets.