Phylogenetic analysis of Primula section Primula reveals rampant non-monophyly among morphologically distinct species
详细信息 查看全文
- 作者：Alex ; er N. Schmidt-Lebuhn ; Jurriaan M. de Vos ; Barbara Keller ; Elena Conti
- 关键词：Ancestral polymorphism ; Bayesian phylogenetics ; Hybridization ; Incomplete lineage sorting ; Phylogeny ; Primula
- 刊名：Molecular Phylogenetics and Evolution
- 出版年：2012
- 出版时间：October, 2012
- 年：2012
- 卷：65
- 期：1
- 页码：23-34
- 全文大小：1693 K
文摘
The type section of Primula (Primulaceae), here considered to include seven species, is phylogenetically quite isolated in its genus. Although its species are popular ornamentals, traditional medicinal plants and model organisms for the study of heterostyly, the section has not yet been studied from a phylogenetic or evolutionary perspective. Using phylogenetic analysis of nuclear ITS and plastid data from all species and subspecies, we find that widespread Primula elatior is genetically heterogeneous and non-monophyletic to most if not all of the other ingroup taxa. The Genealogical Sorting Index (GSI) indicates that the assumption of all currently accepted species being independent lineages is consistent with the data. It is possible that P. elatior in its current circumscription may represents the disjointed remnant of an ancestral species from which the other recognized species diverged. However, based on available data, the alternative possibility of introgression explaining the non-monophyly of this species cannot be excluded. Species trees show P. elatior and P. veris as sister species. Primula vulgaris and P. juliae are closely related, while, in contrast to previous assumptions, P. renifolia does not appear to be a close relative of P. megaseifolia. With the section¡¯s isolation from the rest of the genus and very short internal branches, our dataset also presents a case study of the confounding effects of different branch length priors on the Bayesian estimation of resulting branch length estimates. Experimental runs using different priors confirm the problem of resulting estimates varying by orders of magnitude, while topology and relative branch lengths seem unaffected.