Phylogenomics, biogeography, and evolution of the blue‐ or white‐fruited dogwoods (Cornus)—Insights into morphological and ecological niche divergence following intercontinental geographic isolation

Abstract

The eastern Asian (EA)–eastern North American (ENA) floristic disjunction represents a major pattern of phytogeography of the Northern Hemisphere. Despite 20 years of studies dedicated to identification of taxa that display this disjunct pattern, its origin and evolution remain an open question, especially regarding post‐isolation evolution. The blue‐ or white‐fruited dogwoods (BW) are the most species‐rich among the four major clades of Cornus L., consisting of ~35 species divided into three subgenera (subg. Yinquania, subg. Mesomora, and subg. Kraniopsis). The BW group provides an excellent example of the EA–ENA floristic disjunction for biogeographic study due to its diversity distribution centered in eastern Asia and eastern North America, yet its species relationships and delineation have remained poorly understood. In this study, we combined genome‐wide markers from RAD‐seq, morphology, fossils, and climate data to understand species relationships, biogeographic history, and ecological niche and morphological evolution. Our phylogenomic analyses with RAxML and MrBayes recovered a strongly supported and well‐resolved phylogeny of the BW group with three intercontinental disjunct clades in EA and ENA or Eurasia and North America, of which two are newly identified within subg. Kraniopsis. These analyses also recovered a potential new species but failed to resolve relationships within the C. hemsleyi–C. schindleri complex. In an effort to develop an approach to reduce computation time, analysis of different nodal age settings in treePL suggests setting a node's minimum age constraint to the lower bound of a fossil's age range to obtain similar ages to that of BEAST. Divergence time analyses with BEAST and treePL dated the BW stem back to the very Late Cretaceous and the divergence of the three subgenera in the Paleogene. By integrating fossil ages and morphology, a total evidence‐based dating approach was used in conjunction with time‐slice probabilities of dispersal under a DEC model to resolve ancestral ranges of each disjunct in the Miocene: Eurasia and ENA (disjunct 1), EA and western North America (disjunct 2), and EA (disjunct 3). The dated biogeographic history supports dispersal via the North Atlantic Land Bridge in the late Paleogene in disjunct 1 and dispersal via the Bering Land Bridge in the Miocene for disjuncts 2 and 3. Character mapping with a stochastic model in phytools and comparison of ecological niche, morphospace, and rate of evolution indicated differential divergence patterns in morphology, ecological niche, and molecules between disjunct sisters. Although morphological stasis was observed in most of the characters, evolutionary changes in growth habit and some features of leaf, flower, and fruit morphology occurred in one or both sister clades. A significant differentiation of ecological habitats in temperature, precipitation, and elevation between disjunct sisters was observed, suggesting a role of niche divergence in morphological evolution post‐isolation. The patterns of evolutionary rate between morphology and molecules varied among disjunct clades and were not always congruent between morphology and molecules, suggesting cases of non‐neutral morphological evolution driven by ecological selection. Our phylogenetic evidence and comparisons of evolutionary rate among disjunct lineages lend new insights into the formation of the diversity anomaly between EA and ENA, with particular support of an early diversification in EA. These findings, in conjunction with previous studies, again suggest that the EA–ENA disjunct floras are an assembly of lineages descended from the Mesophytic Forests that evolved from the early Paleogene “boreotropical flora” through varied evolutionary pathways across lineages.

Floral morphogenesis of the Maddenia and Pygeum groups of Prunus (Rosaceae), with an emphasis on the perianth

“…Morphological and anatomical characters are used increasingly in the molecular phylogenetic framework to test evolutionary and systematic hypotheses (Gerrath et al, 2017; Lindelof et al, 2020; Wang et al, 2020b; Zhou et al, 2020). Although the morphology of the mature raceme is similar within the racemose group of Prunus , their position and development show great diversity (Table 1).…”

Integrating coalescent‐based species delimitation with ecological niche modeling delimited two species within theStewartia sinensiscomplex (Theaceae)

“…inheritance, and more similar cases continue to be reported (e.g., Yang et al, 2019Yang et al, , 2020Zhang et al, 2020). In addition to morphological and genetic data, the importance of ecological information is soaring in systematic studies (Rissler & Apodaca, 2007;Ma et al, 2015;Shrestha & Zhang, 2015;Su et al, 2015;Gutiérrez-Tapia & Palma, 2016;Zheng et al, 2017;Wang et al, 2019;Collart et al, 2020;Li et al, 2020;Lindelof et al, 2020;Shao et al, 2020;Tran et al, 2020;Zhou et al, 2020).…”

“…Yang and Phoebe zhennan S. K. Lee & F. N. Wei based on lines of evidence from morphology and genome‐wide genetic inheritance, and more similar cases continue to be reported (e.g., Yang et al, 2019, 2020; Zhang et al, 2020). In addition to morphological and genetic data, the importance of ecological information is soaring in systematic studies (Rissler & Apodaca, 2007; Ma et al, 2015; Shrestha & Zhang, 2015; Su et al, 2015; Gutiérrez‐Tapia & Palma, 2016; Zheng et al, 2017; Wang et al, 2019; Collart et al, 2020; Li et al, 2020; Lindelof et al, 2020; Shao et al, 2020; Tran et al, 2020; Zhou et al, 2020). Environmental niche models can be used to describe the ecological preferences or tolerances of species towards certain habitats.…”

Climatic niche comparisons of eastern North American and eastern Asian disjunct plant genera

“…However, pairwise analyses among all pairs of congeneric species were used because many of these genera comprise only two species and phylogenies are only available for some of these genera. Genera, or subgenera in the cases of Catalpa subgenus Catalpa (Li, 2008; Olsen & Kirkbride, 2017) and Cornus subgenus Mesomora (Lindelof et al., 2020), selected for study were also restricted to those for which occurrence data could be acquired for at least one EA and one ENA species, so that pairwise comparisons within genera/subgenera could be made. Occurrence data for 124 species representing 36 genera (Supporting Information Table S2 lists all 124 species for which occurrence data were obtained; Supporting Information Table S3 lists all species used in analyses) that exhibit the disjunction were successfully downloaded from iDigBio (https://www.idigbio.org/) and the Global Biodiversity Information Facility (GBIF; https://www.gbif.org/) via the ‘spocc’ v.0.7.0 package in R (Chamberlain, 2017).…”