State Key Laboratory of Genetic Resources and Evolution, Kunming Institute of Zoology, Chinese Academy of Sciences, 32 Jiaochang Donglu, Kunming, Yunnan, 650223, China.
Graduate School of Human and Environmental Studies, Kyoto University, Yoshida Nihonmatsu, Kakyo-ku, Kyoto, 606-8501, Japan.
Sci Rep. 2017 Nov 23;7(1):16124. doi: 10.1038/s41598-017-16309-1.
Few empirical studies have compared coalescent-based methods to distance-based methods for delimitation of less divergent species. In this study, we used two coalescent-based (BFD and BPP) and two distance-based barcoding (ABGD and jMOTU) methods to delimit closely related species in the Kurixalus odontotarsus species group. Phylogenetic analyses revealed that the K. odontotarsus species group comprises 11 distinct maternal clades with strong support values. Based on the genetic and morphological evidences, we consider that species diversity in the K. odontotarsus species group was underestimated and the 11 clades represent 11 species, of which six are unnamed. The coalescent-based delimitations decisively supported the scenario of 11-species corresponding to the 11 clades. However, the distance-based ABGD only obtained 3-6 candidate species, which is not consistent with morphological evidence. These results indicate that BFD and BPP are more conservative than ABGD to false negatives (lumping). Method of fixed threshold (jMOTU) may obtain a resolution similar to that inferred by BFD and BPP, but it severely relies on subjective choice of the threshold and lacks statistical support. We consider that coalescent-based BFD and BPP approaches outperform distance-based methods for delineation of less divergent species.
很少有实证研究比较了基于合并的方法和基于距离的方法在界定分歧较小的物种方面的差异。在本研究中,我们使用了两种基于合并的(BFD 和 BPP)和两种基于距离的条形码方法(ABGD 和 jMOTU)来界定近似相关的物种在 Odontotarsus kurixalus 物种组中。系统发育分析表明,Odontotarsus kurixalus 物种组包含 11 个具有强烈支持值的独特母系分支。基于遗传和形态学证据,我们认为 Odontotarsus kurixalus 物种组的物种多样性被低估了,这 11 个分支代表 11 个物种,其中 6 个尚未命名。基于合并的界定方法果断地支持了 11 个物种对应 11 个分支的情景。然而,基于距离的 ABGD 仅获得了 3-6 个候选物种,这与形态学证据不一致。这些结果表明,BFD 和 BPP 比 ABGD 更保守,不易出现假阴性(合并)。固定阈值方法(jMOTU)可能获得与 BFD 和 BPP 推断相似的分辨率,但它严重依赖于阈值的主观选择,缺乏统计支持。我们认为,基于合并的 BFD 和 BPP 方法在界定分歧较小的物种方面优于基于距离的方法。