A phylogenetic analysis of the grape genus (Vitis L.) reveals broad reticulation and concurrent diversification during neogene and quaternary climate change
- Yizhen Wan†1,
- Heidi R Schwaninger†2Email author,
- Angela M Baldo2, 3,
- Joanne A Labate2,
- Gan-Yuan Zhong2, 3 and
- Charles J Simon2
© Wan et al.; licensee BioMed Central Ltd. 2013
Received: 20 November 2012
Accepted: 28 May 2013
Published: 5 July 2013
Grapes are one of the most economically important fruit crops. There are about 60 species in the genus Vitis. The phylogenetic relationships among these species are of keen interest for the conservation and use of this germplasm. We selected 309 accessions from 48 Vitis species,varieties, and outgroups, examined ~11 kb (~3.4 Mb total) of aligned nuclear DNA sequences from 27 unlinked genes in a phylogenetic context, and estimated divergence times based on fossil calibrations.
Vitis formed a strongly supported clade. There was substantial support for species and less for the higher-level groupings (series). As estimated from extant taxa, the crown age of Vitis was 28 Ma and the divergence of subgenera (Vitis and Muscadinia) occurred at ~18 Ma. Higher clades in subgenus Vitis diverged 16 – 5 Ma with overlapping confidence intervals, and ongoing divergence formed extant species at 12 – 1.3 Ma. Several species had species-specific SNPs. NeighborNet analysis showed extensive reticulation at the core of subgenus Vitis representing the deeper nodes, with extensive reticulation radiating outward. Fitch Parsimony identified North America as the origin of the most recent common ancestor of extant Vitis species.
Phylogenetic patterns suggested origination of the genus in North America, fragmentation of an ancestral range during the Miocene, formation of extant species in the late Miocene-Pleistocene, and differentiation of species in the context of Pliocene-Quaternary tectonic and climatic change. Nuclear SNPs effectively resolved relationships at and below the species level in grapes and rectified several misclassifications of accessions in the repositories. Our results challenge current higher-level classifications, reveal the abundance of genetic diversity in the genus that is potentially available for crop improvement, and provide a valuable resource for species delineation, germplasm conservation and use.
KeywordsBiogeography Divergence time estimate Grapevine Molecular phylogeny Network Northern hemisphere Plant disjunction Reticulation SNP Vitis
Grapes (Vitis spp.) are one of the world’s most economically valuable fruit crops . They are widely used for wine, table grapes, raisins, juice, and spirits; recent trends have also focused on antioxidants and healthful products derived from grapes. Vitis vinifera L. subsp. vinifera (referred to as V. vinifera hereafter) is the most widely cultivated grape species but its productivity was historically limited due to its susceptibility to pests, diseases, and abiotic stress such as cold . Genes from wild grape germplasm have been used to improve biotic and abiotic tolerance and resistance in cultivated grapes.
Most previous molecular studies on the evolutionary history of Vitis were limited in taxonomic scope or marker choice [15–32]. Studies most similar in goals and pertinent to the present study were conducted by Aradhya et al. , Nie et al. , Liu et al. , and Zecca et al. . Aradhya et al.  obtained a taxon sample similar to the present study and used SSR and AFLP markers to study genetic diversity within Vitis. These markers have limited value for phylogeny reconstruction  and dating divergences was not attempted. Nie et al.  and Liu et al.  provided well-reasoned paleontological dates to estimate divergences in Vitaceae. These calibration points were applied in the present study. Zecca et al.’s  chronogram is a tantalizing expansion of the Vitis component of Nie et al.’s  chronogram. Their inferences were limited by the small number of markers and the limited variability available in those markers, which did not fully resolve the tip clades. Further, limited intra-specific replication (sampling) limited the ability to make species-level inferences.
Adding more data can be useful for resolving difficult phylogenies that were based on a few genes . The present study attempted to improve on three aspects of previous phylogeographical studies of Vitis by more extensive sampling of the nuclear genome, the species, and intraspecific variation. This study developed and used 27 nuclear gene markers and sequenced 309 accessions of 48 Vitis species, varieties, and four out-groups to: 1) reconstruct a phylogenetic hypothesis of the genus Vitis, 2) date important time points in the evolution of Vitis, 3) elucidate the biogeographic history of the genus, and 4) evaluate systematics of Vitis within the framework of phylogeny.
Molecular characteristics of the nuclear sequences of Vitis
Most Vitis accessions had complete sequence or had minimal missing data (Additional file 1). Indel sequences in the 27 gene markers were unambiguous and easy to align. The starting alignment matrix for the 27 gene markers and all 309 accessions was 11,437 bp long. Gap coding for Maximum Parsimony added 304 characters. Amino acid coding sequence accounted for 5,690 nt, 3’ or 5’ untranslated regions for 4,074 nt, and introns for 1,036 nt (Additional file 2). Because Trees by New Technology (TNT) does not output the number of parsimony-informative characters, we report the number of unique site patterns from Bayesian Evolutionary Analysis by Sampling Trees (BEAST) and Randomized Accelerated Maximum Likelihood (RAxML). The 52-OTU matrix had 1,855 unique site patterns; the 273-OTU matrix had 2,510 unique site patterns. Their distributions among the gene markers are listed in Additional file 2. Under the uncorrelated log-normal relaxed molecular clock, estimated from three combined runs in BEAST and calibrated with three fossil dates, the mean rate of substitution (meanRate) in the data set was 8.249×10-4 per million years (Effective Sample Size (ESS) = 959) and a coefficient of variation (CV) of 0.896 (ESS = 1284). Dividing this rate by the size of the data set (11,437 nt), the average rate of substitution in this data set was 7.2 × 10-8 substitutions per site per million years.
Extent of reticulation and network in Vitis
Ancestral polymorphism with subsequent lineage sorting is difficult to distinguish from reticulation based on a phylogenetic pattern [38–42]. In our study, six (22%) gene fragments showed significant tests for recombination based on the Phi Test : fragments 1313 (P = 0.033), 1973 (P = 0.033), 2415 (P = 0.006), 5069 (P = 0.033), 7022 (P = 0.047), 7029 (P = 0.019). When these fragments were excluded, the concatenated matrix still showed recombination, as expected (P = 0.000), representing independent assortment of markers. Lanier and Knowles  found that, in species-tree estimation, the gain of accuracy from sampling additional loci and/or individuals always exceeded inaccuracies related to recombination. Thus, in the present work no genes were excluded from subsequent analyses based on evidence of recombination.
Estimation of divergence times in Vitis
Phylogenetic analyses in Vitis
ML analyses of 26 single genes analyzed independently without missing data yielded 26 very poorly resolved trees. No significant conflict was observed under ML, thus the sequences were concatenated.
The two runs timed out at 48 mil and 50 mil generations with good effective sample size (ESS; > > 200), but they did not converge on the exact same phylogenetic hypothesis. Both trees were highly concordant for the species-level clades but differed on the specific placement of some clades (Additional files 8 and 9). Because the 48 mil run (Additional file 8) had a higher mean log likelihood of the cold chain (LnL) after burnin, the posterior probabilities from this run were used to summarize supports (Figures 4 and 5). Overall comparisons of relationships above the species level were facilitated using the cartoon of this tree (Additional file 10).
Ancestral area analysis
Fitch parsimony identified Eastern/Southeastern North America (that also included Mexico and the Caribbean) as the origin of the most recent common ancestor of Vitis based on the strict consensus tree (Additional file 14).
Systematics of Vitis spp.
It was evident that the patterns were similar among the cartoons of the strict MP consensus (Figure 6), the highest ML value cladogram (Additional file 6) and the Bayesian cladogram (Additional file 10). The clades of Eurasian species were nested in North America as a monophyletic clade. The series Precoces Munson (containing V. riparia Michaux, V. acerifolia Raf., V. rupestris Scheele) together with V. arizonica Engelm. (belonging to the series Occidentales Munson, Additional file 15), V. blancoii Munson, V. bloodworthiana Comeaux, V. Xtreleasii Munson ex L. H. Bailey and V. girdiana Munson formed the sister clade. V. labrusca and V. aestivalis Michx. were grouped together as were V. cinerea (Engelm.) Engelm. ex Millardet, V. palmata Vahl, V. shuttleworthii House and V. mustangensis Buckley. All three analyses did not group V. monticola Buckley or V. californica Benth. with other species.
The position of several clades differed among the searches. This was a notable characteristic of this data set. Among the different BA, ML and MP trees shown, the OTU composition of species clades in general was quite consistent and well supported, but a few clade or species positions were inconsistent. Correspondingly, clades above the species level were often poorly supported. For example, V. yenshanensis J. X. Chen was grouped with V. amurensis Rupr. and V. coignetiae Pulliat ex Planch. in MP and BA analysis and in the network but not in ML. Similarly, the amurensis/coignetiae clade was basal to Eurasia in MP and BA, but placed within the Asian clade in ML and in the network. V. nesbittiana Comeaux were grouped with V. mustangensis and V. shuttleworthii in BA and ML but grouped with V. bloodworthiana and V. blancoii in MP and in the network.
To better understand clade support, we further investigated the synapomorphies defining clades of interest. The characters supporting nodes of interest and their level of homoplasy and gene source were summarized based on MP (Additional file 11). Summary statistics for specific nodes investigated are shown below the branches in Figures 4 and 5. Many species showed good support with high bootstrap values and posterior probabilities, presence of node-specific characters and support from multiple genes (letters in parentheses refer to marked nodes in Figures 4 and 5): V. shuttleworthii (D), V. palmata (E), V.cinerea (all varieties and including V. biformis Rose) (F), V. biformis (G), V. labrusca (H), V. californica (I), V. nesbittiana (J), V. girdiana (K), V. amurensis (L), V. qinlingensis (M), V. davidii (Rom. Caill.) Foëx(N), V. quinquangularis Rehd. (O), V. bashanica P.C. He (P), V. hancockii Hance (Q), V. davidii var. cyanocarpa (Rom. Caill.) Foëx (R; and not grouping with V. davidii), V. vinifera ssp. sylvestris (C. C. Gmel.) Hegi. (referred to as V. sylvestris hereafter) (S). The species V. vulpina L. (T) and V. monticola (U) each formed reliable species-level clades even in the absence of a species-specific SNP. Many higher level relationships were supported by a few characters and were of poor quality leading to the labile topology among major clades. Examples were the nodes defining the split between North America and Asia (B), Europe (with V. jacquemontii R. Parker) and China (C), and China without V. amurensis (V). These nodes had five to seven supporting characters that were frequently highly homoplasious (Figure 4; Additional file 11). The notable exception was the branch separating the Muscadinia from subg. Vitis (Figure 4 node A). It was supported by 93 characters of which 57 (61%) showed no homoplasy and represented 20 of the 27 (74%) gene markers. Other supported higher clades were the V. cinerea-V. palmata clade (X in Figure 4) and the (V. mustangensis, V. shuttleworthii, V. palmata, V. cinerea with V. biformis) clade (Y). Both nodes had a node-specific SNP, although bootstrap support and posterior probabilities were more consistent with other poorly supported higher-level clades.
The markers were informative in characterizing intra-specific variation in some species. The MP branch lengths (Additional file 16), MP and ML bootstrap supports and BA posterior probabilities (Additional files 7,8,13) supported intraspecific groupings well with non-zero branch lengths present in many species such as V. shuttleworthii, V. monticola, V. californica, V. palmata, V. labrusca, V. cinerea, V. aestivalis, V. sylvestris, V. adstricta, and V. davidii. Due to space constraints these supports were not summarized in (Figures 4 and 5).
Problems of phylogenetic study in Vitis
Based on MP, ML, and BA phylogenetic reconstruction methods, the nuclear DNA dataset in this study had extensive variation to address genus-wide relationships in Vitis. The markers characterized intraspecific variation, defined most species, and strongly supported subg. Vitis. However, many of the relationships and deeper nodes (above species within subg. Vitis) were characterized by low bootstrap values and were often supported by few characters with high homoplasy. Low clade support and high homoplasy may be caused by insufficient data, parallel changes, reversals and convergences, as well as different histories of genes caused by lineage sorting and reticulation [40, 46], and different clade sizes . The ascertainment bias (the systematic distortion in measuring the true frequency of SNPs due to sampling) introduced in the marker development phase may have selected markers with insufficient variation outside V. vinifera or phylogenetic depth, and differences may be distorted due to domestication. However, we re-sequenced whole fragments, rather than genotyping a priori identified SNPs, thus included additional markers not restricted by our selection criteria of intermediate variability (see Methods). Further, the deepest node was very well supported both in number and quality of characters, illustrating that there was sufficient phylogenetic signal available to define subg. Vitis. Similarly, several species (e.g. V. shuttleworthii, V. labrusca, V. palmata, V. hancockii, V. quinquangularis) were well supported by multiple non-homoplasious characters. Thus, simple lack of data was not the definitive reason for poor support of deep nodes. Additional data, using the Vitis9KSNP array  or others in development, markers developed by Lijavetzki et al.  and Vezzulli et al.  or next generation sequencing (NGS) may resolve this phylogenetic problem. However, these data will certainly add more noise (homoplasious characters). In addition, ascertainment of homology in the data set created by NGS is very intractable because of complex paleopolyploidization and gene duplication in the grapevine genome . It is possible that additional data from non-recombining chloroplast or mitochondrial DNA might add stable characters deeper within the tree. However, the literature [19, 23, 25, 29, 36, 48] suggests that nucleotide substitution rates in these datasets may be too slow to add much intra-generic information. Species reticulation and incomplete lineage sorting would still present a challenge. Parallel changes, reversals, and convergence are likely minor contributing factors to the observed homoplasy due to the shallow phylogenetic depth of this study involving moderate levels of evolutionary time; these factors were further minimized by locus selection criteria.
Ancestral polymorphism, reticulation and incomplete lineage sorting
Homoplasy due to incomplete sorting of ancestral alleles is more likely when the time between lineage splitting is short (short branch especially when deep in the tree ) and the effective population size is large . The present estimates of divergence times showed that splitting events between the deeper clades occurred almost simultaneously within subg. Vitis. Myles et al.  found significant degrees of shared polymorphisms between North American wild grapevine species and European cultivated species, suggesting that grapevine species maintained large effective population sizes since their geographic isolation millions of years ago. Further, the linkage disequilibrium in V. vinifera is very low and haplotype blocks are very short [20–22, 51], indicating significant historical recombination within the species . There was significant recombination in several genes and in the concatenated dataset. Thus, the conflict in the NeighborNet (Figure 2) can be interpreted as evidence of shared ancestral polymorphisms mixed with reticulation and lineage sorting. The shared ancestral polymorphisms may be the cause of the central knot of conflict (an ancestral ocean of polymorphisms and reticulation) represented by the tight central mass of splits that represent incompatible and ambiguous signals in the data set , with the radiating splits representing progressive lineage sorting and reticulation within the lineages. Reticulating events include hybridization, recombination and horizontal gene transfer . The first two were likely major factors in the evolution of Vitis, while horizonatal gene transfer was an unlikely mechanism. We conclude that extensive reticulation deep in the tree and incomplete lineage sorting are the likely reason for the lack of support at higher level nodes.
Time frame of Vitisdiversification
Comparison of divergence estimates in Vitis among five studies that analyzed SNP data and used zero, two or three fossil calibration points
Xiang et al. 20002
Nie et al. 20103
Zecca et al. 20124,5
Liu et al. 20136
95% HPD (Ma)
subg. Vitis crown
95% HPD (Ma)
Europe/Near East - Asia
95% HPD (Ma)
4.98 (5.98- < 0.28)
95% HPD (Ma)
Higher level intra North America
95% HPD (Ma)
Higher level intra Asia
95% HPD (Ma)
95% HPD (Ma)
Fossil calibrations (Ma), Priors used
Vitaceae - Leea
85 ± 4.09
85 ± 4.0
90.7 ± 1.010
90.7 ± 1.0
58.5 ± 5.011
58.5 ± 5.0
58.5 ± 5.0
Old World - New world Parthenocissus
21.6 ± 6.412
5.75 ± 0.514
5.75 ± 0.5
Continental origin, dispersal and diversification of Vitis
The phylogenetic relationships and network of grapevines reflect the Northern hemisphere Cenozoic history. The extensive ancestral reticulations revealed by the network and analysis of individual genes suggested well connected ancestral populations and species throughout the distribution followed by increasing range-wide fragmentation, isolation, and differentiation. The ancestral area analysis and the recurring distributional trend of American paraphyly with Eurasia in this study suggested a progression from North America to Asia to Europe consistent with previous studies [17, 29, 33, 36]. However, Péros et al.  concluded that their analysis may support an Asian origin of Vitis. Fossils of Vitaceae have been found frequently in Western North American Eocene deposits (55.8 to 33.9 Ma) and have not yet been found in southeastern localities . Fossils of Vitis seed were found in deposits of the Rocky Mountains and Great Plains of North America [34, 53] and in central Europe [34, 54]. These findings assigned the oldest age of Vitis to the Paleocene (65.5-58.8 Ma). At that time the supercontinent Laurasia had only begun dividing into North America and Eurasia  and the climate was considerably warmer in the northern latitudes . These factors facilitated dispersal of warm-temperate terrestrial organisms in the northern hemisphere. Most East Asia–North American disjuncts from diverse families have had longer histories in North America than in Asia: of nine woody East Asian–East North American disjunct genera  all appeared earlier in the fossil record of North America than in that of Asia . Wen et al.  found many more lineages with North American origins and migration to Asia than vice versa. Nie et al.  argue for a North American origin of Ampelopsis (Vitaceae). Molecular phylogenetic analyses of several disjunct genera suggested a progression from East Asia to Eastern and Western North America [65, 66]. Thus, the balance of grape-specific information tends to support our findings of a North American origin for the most recent common ancestor of Vitis.
After fragmentation of a Paleo/Neogene range, our phylogenetic trees suggested isolation of some North American and Asian species during the Plio- and Pleistocene cooling cycles, post glacial range expansions, and ecological adaptation. Much of the current Eastern North American range of V. riparia, V. labrusca, V. aestivalis, and V. cinerea was unsuitable for Vitis during the Wisconsinan glaciations due to coverage by the polar ice sheet and harsh conditions along its southern edge (Figure 1). These species must have expanded to their large present ranges after the glacial period. Large range expansions with post glacial warming were also promoted by the physiographic homogeneity of Eastern North America . Fragmentation and local adaptations were evident in physiographically heterogeneous western North America and temperate eastern Asia. The North American species V. shuttleworthii, V. nesbittiana, V. girdiana, V. palmata, V. bloodworthiana and V. blancoii have smaller ranges and multiple species-specific SNP character changes. Similarly, physiographically diverse eastern Asia  had three species with multiple species-specific SNPs: V. bashanica, V. hancockii, V. quinquangularis. Local adaptations in heterogeneous environments likely lead to smaller population sizes and thus more rapid loss or fixation of novel characters .
The underlying evolutionary scenario for Vitis is consistent with origin in the Eocene, a time of maximum development of temperate Paleo/Neogene forests. This was followed by diversification in the mid-Oligocene, the rise of subg. Vitis in the early Miocene, the North American and Asian disjunction in the late Miocene, range restriction and fragmentation and speciation during the Pliocene and Pleistocene cooling cycles. These caused the primary divisions within Vitis as well as species-level and some intra-specific divisions . The North Atlantic land bridge was present in the early Paleogene [69, 70] and may have no longer existed when Vitis arose, leaving Beringia as the major route for potential gene flow. The area of the Bering and Chukchi seas lay above sea level for most of the last 50 to 60 M years  and was suitable for exchanges of temperate plants  until the establishment of the Bering Seaway 3.5-5 Ma , permitting genetic exchange at least until late Miocene to which the disjunction was timed. The Pleiocene/Pleistocene cooling cycles are well known to have caused range restrictions, survival in refugia, and diversifications in many groups of organisms , both on land and in the sea. This study shows clearly that Vitis was also a part of this great biogeographic phenomenon.
The systematics of Vitis is a challenging area of taxonomy. Our findings confirmed the tenuous nature of many grapevine species and especially higher groupings such as series. The apparent species-specific SNPs are good candidates to apply in species delineation investigations of grapes.
The present study found very low support for all series that included more than one species except for the Munson/Moore series Precoces/Ripariae (Figure 4, node W). Other well supported higher-level groupings were subg. Vitis (Figure 4, node A) and genus Vitis, supporting the division of the genus Vitis into two sections [4, 74, 75] or subgenera [5, 76]. Additional file 15 lists a synopsis of the major Vitis classifications. Only Galet  assigned Asian species to series. The most comprehensive treatment of Chinese Vitis did not apply a series-level classification. Most Chinese species could be assigned to one series if series were to be used (Figure 5). This may not include V. amurensis, V. coignetiae and V. yenshanensis as these species in some analyses grouped firmly within Asia (as opposed to Figure 5 where they are basal to Eurasia). It appears as if V. jacquemontii should be assigned to the series Viniferae. However, our accessions had perfect flowers, suggesting past hybridization with V. vinifera. The phylogenetic position intermediate between the Asian and Eurasian species and the well-defined split revealed in the network (Figure 2) supported this conclusion.
The derived position of V. sylvestris was unexpected. V. sylvestris is the suggested progenitor of V. vinifera while the phylogenetic position suggests that V. sylvestris was derived from V. vinifera (Figure 5). This may be an artifact of the tenuous nature of most higher-level relationships revealed in this study. It could also be a result of the nature of selection and clonal propagation that all V. vinifera cultivars included in the present study have been subjected to, some of them potentially for thousands of years [77, 78]. Evolution is arrested by clonal propagation, leaving the naturally evolving wild species to appear more derived. Myles et al.  concluded that current commercial V. vinifera varieties are only one or two generations removed from the wild V. sylvestris.
Mullins et al.  hypothesized Asian/North American sister species pairs for V. coignetiae/V. labrusca and for V. jaquemontii/V. tiliifolia (V. lanata and V. caribaea in ). Our results did not support sister pair relationships for V. coignetiae /V. labrusca as these species placed solidly into well separated Asian and North American clades, respectively. Our results are inconclusive with respect to the V. jacquemontii/V. tiliifolia pair due to the possible hybrid nature of V. tiliifolia accessions in general and the dispersed positions of V. tiliifolia samples.
V. girdiana has been considered to be a variety of V. arizonica, a variety of V. californica and its own separate species . Our results preliminarily identified V. girdiana as a well supported independent species (using the general lineage concept  and diagnosability ) with five species-specific SNPs. More samples need to be investigated to assess the discriminatory power of these SNPs. Wada  also identified a monophyletic V. girdiana cluster, although it had poor bootstrap support.
Samples 080-084 came to us as V. cinerea (Engelm.) Engelm. ex Millardet var. floridana (Munson) but placed solidly into the V. aestivalis clade. This highlighted confusion in the past related to the synonym Vitis simpsonii that has been claimed for two different species as described in Comeaux , one belonging to Aestivales and the other to Cinerescentes. The synonym V. rufotomentosa has the same problem. Our study showed conclusively that these accessions belong to V. aestivalis. Several additional accessions were identified as misnamed and others were recognized as hybrids (Additional file 1). Finally, two accessions, 111V. flexuosa DVIT1385 and 304V. wilsoniae Wangmaiputao, were of Asian origin yet grouped with North American accessions and remain anomalies that could not be resolved.
This is the first study to apply sequences of a large number of nuclear loci combined with extensive species and intraspecific sampling to the phylogeny and biogeographic history of Vitis and the problem of Vitis systematics. The genome-wide sampling of SNPs provided insight into the evolutionary history of the grape genus and supported previous notions of Paleogene origins, range fragmentation, and recent nature of the species, joining Vitis with the large group of organisms whose extant species differentiated in response to Pliocene and Quarternary climate change . We found that the most recent common ancestor of Vitis was North American. The major clades formed throughout the native distribution at 23-8 Ma (broad range due to large HPDs), suggesting that vicariance (the fragmentation of a large Paleo/Neogene Northern hemisphere distribution) in conjunction with local adaptation, was a dominant force in structuring genetic diversity of extant Vitis spp. We demonstrated that genome-wide nuclear SNPs were a productive approach to address questions at and below the species level in grapes. Many species were well supported, and the markers with low homoplasy defining those lineages will likely be useful in species delineation and assessing the reliability of different morphological taxonomic characters. Most higher-level relationships within the genus suffered from weak support. The genus itself was extremely well supported. This suggested that the phylogenetic signal was too weak to overcome the level of noise created by evolutionary forces acting within the Vitis gene pool. Two of the most important forces, probably acting concurrently or alternating, are incomplete lineage sorting of ancestral polymorphism and reticulation. Broad reticulation across many species probably prevented the ancestral gene pool from diverging during the Neogene forest stage, maintained reproductive compatibility, and is still acting today as evidenced by the prevalence of hybrids found in the wild and in repository collections. However, climatic oscillations during the Pliocene and Quaternary, coupled with physiographic heterogeneity, provided enough recent barriers to gene flow to facilitate evolutionary divergence. In light of the recency of divergence and diffuse genetic boundaries, higher-level taxonomic groupings, such as series, may be misleading.
A total of 309 accessions of 48 species or varieties (~80% of the approximately 60 known species of the genus) and outgroups were sampled in this study: 21 species from Asia, both European species, and 25 species and varieties from North America (Figure 1; Additional file 1). These samples were obtained from: 1) the Grape Germplasm Collection at the Northwest A&F University (NAFU), Yangling, Shaanxi Province, China, (DNA), 2) USDA-ARS, Plant Genetic Resources Unit (PGRU), Geneva, NY, USA, and 3) USDA-ARS, National Clonal Germplasm Repository (NCGR), Davis, CA, USA. Four closely related genera based on chloroplast and nuclear markers [19, 28, 32, 34] were chosen as outgroups in the dating of divergences using BEAST: Parthenocissus spp. Planch, Ampelopsis glandulosa (Wall.) Momiy. var. brevipedunculata (Maxim.) Momiy, Leea coccinea Planch. ‘Rubra’, Cayratia japonica Thunb. Two of the outgroup genera (five species of Cissus and Cayratia japonica) were obtained from a research collection (Dr. P. Cousins, USDA/ARS, presently E & J Gallo Winery). The outgroup Leea coccinea “Rubra” was grown from seeds obtained from Carter Seeds (Vista, California).
No cultivars of Vitis spp. were included except for V. vinifera ssp. vinifera for which no wild accessions are known . To mitigate long-branch attraction, the 40 chromosome Vitis rotundifolia Michx. subg. Muscadinia was used as the outgroup in analyzing subg. Vitis. This was justified by all preliminary analyses on the complete data set that identified V. rotundifolia as the sister species to subgenus Vitis and it is consistent with other studies e.g. [23, 29, 33–36]. One to 27 accessions or genotypes were sampled per species. All available varieties (not cultivars) of a species were sampled. Similarly, widely distributed species were more extensively sampled to include potential geographic differentiation. A few sibling groups were also included to test the ability of the markers to place or distinguish those accessions.
Based on preliminary analyses, accessions that placed in unexpected positions or had very weak support on preliminary phylogenetic trees were submitted to a taxonomic expert (Dr. P. Cousins) for an independent assessment of species identity, but without indicating the nature of the conflict. Additional file 1 lists the results of all assessments. The labels in the figures and tables indicate the corrected names unless otherwise indicated. Exact geographic coordinates of origin were not available for most accessions. Accessions and pertinent details are listed in Additional file 1. The accessions located in the US repositories can be requested through the Genetic Resources Information Network (GRIN)  and plant materials (leaves, cuttings) can be requested from the clonally maintained vines at these sites.
DNA isolation and re-sequencing
DNA was isolated from fresh or frozen young leaves and apical meristems using a modified CTAB (cetyltrimethylammonium bromide; Sigma H6292) protocol [84, 85] with 2-5% PVP (Polyvinylpyrrolidone, mol. wt. 40,000; Sigma PVP40) in the extraction buffer to remove secondary compounds, two chloroform purifications to remove proteins and a NaCl and ethanol precipitation to remove polysaccharides.
Primer screening was performed in 25 μL PCR volumes. 50 μL PCR volumes were cleaned for sequencing, concentrated and used in 12 μL cycle sequencing reactions. Additional file 17 provides the detailed conditions.
The exploratory sequencing was performed in-house at PGRU on ABI-3100xl Genetic Analyzer. The high-throughput sequencing (30 gene fragments for 309 accessions) was performed by Genaissance Pharmaceuticals, Inc (New Haven CT, USA). Both strands were separately sequenced using the PCR forward or reverse primer.
SNP discovery and selection
Expressed Sequence Tags (ESTs)  of Vitis vinifera and grape mRNAs in NCBI in 2004 were sub-clustered and surveyed to predict SNPs using an in-house pipeline as described by Labate and Baldo . The 62 variably-sized EST libraries and additional grape mRNAs included 108,429 V. vinifera sequences which formed 3,792 clusters. Because EST data are often based on one sequencing pass and are not filtered for error, a predicted SNP may not be verifiable. Because this study was intended to survey broadly across the entire genus, gene markers that were predicted to be monomorphic among V. vinifera were discarded. Markers with extreme levels of polymorphism were also excluded to minimize possible selection of duplicated loci.
Pairs of PCR primers were designed using the program ‘Primer 3’  for 281 gene fragments of 400-600 base pairs (bp) containing moderate polymorphism. The amplifications were tested using three DNA samples, one each from Asia (V. romanetii Rom. Caill. ‘Jiangxi2’), Europe (V. vinifera ‘Rotberger’, DVIT2339) and North America (V. rotundifolia, DVIT1689). Robust, single bands were obtained for 201 of 281 primer pairs (71.7%). Then 96 primer pairs with robust single bands were chosen for re-sequencing to test sequence quality using eight species (V. cinerea (PI588575), V. labrusca L. (PI588194), V. amurensis (Zuoshan1), V. quinquangularis (Weinan3), V. romanetii (Pingli7), V. davidii (Xuefeng), V. hancockii (Lingye_F) and V. yenshanensis (Yanshan_F). Thirty of the most consistently amplifiable gene fragments, both within Vitis and outgroups, with suitable polymorphisms and only minor sequence length variation in the eight tested, were re-sequenced in a total of 309 accessions (Additional file 1). Predicted genes were identified in comparison with the NCBI non-redundant protein sequence database. When the V. vinifera genome sequence became available , the primer sequences and gene fragments were BLASTed  against this genome to determine their chromosome locations and confirm their homology and identity. When the final dataset was assembled, three gene markers were excluded because of unalignable indels (one marker), and suspected duplicate loci (two markers). The sequences of the 27 final primer pairs and supporting information are listed in Additional file 2.
Sequence alignment, data sets, coding of gaps
ProSeq  was used for editing sequence based on trace files, and Mutation Surveyor (Soft Genetics) was used for base calling. Heterozygotes were manually edited to use the IUPAC-IUB symbols for nucleotide nomenclature . The results from ProSeq and Mutation Surveyor were compared for accuracy, and nearly 100% agreement was found. Discrepancies were resolved by examining trace files manually. Sequences were aligned manually and also aligned using Clustal W  with default parameters.
Extensive preliminary phylogenetic and PCA  (Additional file 18) analyses using all (Additional files 19, 20, with partitions listed in Additional file 21) and subsets of OTUs revealed known and new hybrids which were excluded in the final analysis because phylogenetic trees can be strongly influenced by hybrid taxa . This does not guarantee that the final analyses were devoid of hybrids as they are not always identifiable based on morphology. The final phylogenetic data set contained 273 OTU composed of subgenera Vitis and Muscadinia (Additional files 1, 19, 20). This dataset was modified for dating divergences using BEAST: 1) Four outgroup taxa were added to match calibration points in Nei et al. : 060_Ampelopsis brevipedunculata, 096_Leea coccinea ‘Rubra’, 129_Cayratia japonica, 247_Parthenocissus spp.. Leea and Cayratia had substantial amounts of missing data (Additional file 1). 2) With the presence of multiple individuals per species dating is a more complex issue and would preferably apply coalescence methods. Preliminary analyses indicated that the present data set was not sufficiently informative to allow a well-supported coalescent analysis. Thus, the number of ingroup OTUs was reduced to one accession per species and variety for efficient calculations  and to satisfy the Yule model of speciation. These modifications resulted in the 52-OTU dataset (Additional file 22). Additional file 1 summarizes the members of each analysis. Preliminary ML analyses were conducted on all single gene fragments, and partitioned and unpartitioned concatenated sets for a total evidence dataset .
Gaps were treated as characters using ‘simple indel coding' (SIC)  and implemented in SeqState 1.4.1 . Simple gap coding was chosen because it is a preferred coding method for empirical studies [99, 100]. Inclusion of gaps in phylogenetic analyses is limited by the optimality criterion used for phylogenetic inference. Gap information was used for parsimony analysis only. ML and BA treated gaps as missing data. Combinability of DNA partitions was ascertained using Wien’s  method: existence of corresponding but incongruent clades with bootstrap support greater than 70% are seen as support for not concatenating data sets.
Test for recombination, network analyses
The best single gene tree from each locus was combined into a file from which a consensus network was constructed in Splitstree4 . Thresholds used were 0.04 (all splits present in at least one tree, 1/26), 0.08 (splits present in at least two of the trees), 0.5 (splits present in half the trees), 0.9 (splits present in 90% of the trees).
Calculation of divergence time
Bayesian (BA) estimates in the BEAST V1.7.4 [95, 106] software were used to estimate divergence dates using Markov Chain Monte Carlo (MCMC) sampling. Trees were visualized in Figtree V1.3.1 . Many preliminary runs were conducted on the partitioned and unpartitioned data file to explore parameters. Operators were optimized automatically. The final .xml files (Additional files 22, 23) were run ten times using the maximum time available at the Computational Biology Service Unit (CBSU) BioHPC computer cluster at Cornell University. The conditions for the partitioned runs were: 27 unlinked partitions with individual substitution models, estimated frequencies of nucleotides, a random starting tree, an uncorrelated relaxed clock with log normally distributed uncorrelated rates between branches , Yule model of speciation (a pure birth process), default operators modified based on preliminary runs, auto-optimization turned on, and parameters were sampled every 10,000 steps. The conditions for the unpartitioned runs were identical except that there was only one partition and one substitution model. Marker-specific and whole-dataset-specific substitution models were determined in Findmodel  and are listed in Additional file 2. Findmodel uses Weighbor, PAML and methods in Modeltest  to determine substitution models. Following the reasoning of Nie et al.  a normal prior was used with a mean of 58.5 Ma (st. dev. = 5 Ma) for the stem age of Vitis. Liu et al.’s  reasoning was adopted to 1) place the second calibration point of V. labrusca and closely related North American relatives in the subg. Vitis at 5.75 Ma (st. dev. = 0.5 Ma), and 2) fix the stem age of Vitaceae with a normal prior distribution of 90.7 Ma (st. dev. = 1 Ma). Runs from the same .xml file were combined if they shared the same trace, met general quality requirements outlined in Additional file 24, and if the addition increased the ESS of key parameters. The three unpartitioned runs with the highest (almost identical) likelihoods were combined after 10% burnin was removed to produce the chronogram in Figure 3. Unpartitioned runs were used because the combined partitioned runs did not have sufficient support for important parameters. Five runs of the partitioned dataset were combined after removing a burnin of 10-75% to compute the evolutionary rates (gene.meanRate) of the individual genes. These gene.meanRates had large ESS. Means and 95% Highest Posterior Density (HPD) from these combined runs were computed using TreeAnnotator.
Phylogenetic and ancestral area analyses
To evaluate species we used a phylogenetically based general lineage concept, where species are defined as separately evolving segments of metapopulation lineages . Additional criteria of more stringent species definitions were considered, such as monophyly  and diagnosability [81, 114].
The 273-OTU matrix was analyzed under the ML criterion using RAxML  versions 7.0.4, 7.2.6 and HPC2 at the web portal of the Cyber Infrastructure for Phylogenetic Research (CIPRES) cluster. The data were partitioned by gene fragment (Additional file 21). All characters were included. Gap coding was removed. Indels were treated as missing data. Twenty replicate searches were run on this final data set using a rapid hill climbing algorithm and the GTRGAMMA (= GTR + Optimization of substitution rates + GAMMA model of rate heterogeneity, the alpha parameter was estimated) model of substitution as recommended by the program’s author, and the default of 25 rate categories. The rapid bootstrapping option  was chosen to generate 1,000 bootstrap replicates. The best-scoring ML tree was obtained in the same search and bootstrap values were annotated. Output was visualized in Dendroscope V2.2.2 .
To test for conflict among genes, a preliminary analysis was performed for each gene fragment using the same parameters as above but without a partitioned model. Each gene was run in 10 replicates. Five more replicates were added if the final maximum likelihood values varied extensively among replicates. To keep trees comparable no OTUs were excluded. One-thousand bootstrap replicates were collected for each gene fragment. Incongruent clades with bootstrap support of 70% or greater were considered as support for not combining data sets [42, 100]. Due to computational limitations, the single gene analyses were performed only in RAxML at CIPRES. Because of the sparse yield of information and low information content of most markers, this computational expense was not repeated with the 273-OTU matrix.
Bayesian analyses were performed on the 273-OTU matrix using Mr. Bayes  on the concatenated, non-partitioned data set using the K80 substitution model (Nst = 2, 4 by 4) plus Gamma, as determined in Findmodel. Multiple short runs were performed to determine the temperature and number of chains that would support chain swap. The two final and longest runs of 48 and 50 million generations (fitting just within the 168 hr time limit) were run with 8 chains, temperature of 0.10, and sample frequency every 5,000 generations. Tracer V1.5  was used to evaluate the MCMC runs, TreeAnnotator v1.6.1  and Figtree v1.3.1 were used to annotate and visualize maximum credibility trees listing all posterior probabilities. Burn in was 2.5 million (50 mil run) and 10 million (48 mil run).
Maximum Parsimony (MP)
The software package TNT, Hennig Society version  was used to analyze the unpartitioned 273-OTU matrix under parsimony, assuming unordered character state transformation and equal weights (Fitch parsimony) . Uninformative characters were excluded and gaps were coded. The efficient option of “driven search” in TNT was used for the search. This option searched until a minimum tree length was found a certain number of times and then a consensus was estimated. After a second round of searching the new consensus was compared to the previous one, and so on until the consensus stabilized . The driven searches included the ratchet , tree drifting, tree fusion and sectorial search . The default settings were used except that the consensus was stabilized four times instead of twice. The search was repeated four times using a different random starting seed and without specifying a target score. The strict consensus tree was constructed using all most parsimonious trees from all four searches. Bootstrap support was based on 1,000 replicates of driven searches using the same search components and default parameters. Synapomorphies were optimized and listed, and those of interest were reconstructed. The character values, indicating the level of homoplasy, were obtained in TNT to study the support at nodes of interest. Trees for illustrations were exported in a NEXUS format, manually converted to Newick trees, visualized and annotated using Dendroscope V 2.2.2.
Ancestral area analysis
The geographic distribution was partitioned into four continental area units that correspond to broad distributional trends in Vitis: Asia, Europe/near East, E- and SE-North America (including Mexico and the Caribbean) and Western North America (Figure 1). An area code was added to each accession in the 273-OTU data matrix (Additional file 20). Fitch optimization (reversible parsimony)  was performed in TNT to optimize the area on the strict consensus tree .
Biogeography: the Eastern Asia-Eastern North American disjunction
The genus Vitis contributes to one of the great distributional phenomena in plant biogeography, the Eastern Asia-Eastern North American disjunction among the temperate to warm temperate northern hemisphere taxa [69, 126–129]. Up to 30 species are native to a vast area in eastern Asia, China, Japan and Java, two species across middle Asia and Europe, and up to 28 species across the eastern and southwestern US and Mexico  (Figure 1). A small number of species extend into the Tropics both in Asia and in North America [6, 130–132]. There is widespread agreement that these disjunct floras are relicts of plant communities that were distributed throughout a large part of the Northern Hemisphere during much of the Paleogene and early Neogene (formerly the Tertiary) Periods (65-15 Ma) [69, 70, 126, 129, 133, 134]. Communities on different continents were linked by migration across the Bering Land Bridge, linking North America and Asia beginning in the Miocene , and across the North Atlantic Land Bridge, linking North America and Europe particularly in the early Eocene [65, 129]. Wild grapes are a savored food of birds and some small mammals, providing dispersal for these species. Intra-continental migration was impeded between Europe and Asia by an epicontinental seaway (Cretaceous-Eocene) as was migration between east and west North America (upper Cretaceous), followed by regions of dry continental climates [65, 129, 134]. Climatic cooling at the start of the Oligocene (33.9 Ma) gave rise to the Mixed Mesophytic Forest of deciduous and evergreen trees and associated taxa that comprise the modern Paleogene/Neogene relict floras , among them the early grapevines. The flora retreated into refugial regions in response to Pliocene cooling (5.3-2.5 Ma) and Quaternary glaciations (2.5-0 Ma) . Tectonic uplifting of mountain ranges and plateaus during the Pliocene into the Holocene, and concurrent reduction in precipitation caused further partitioning of the East Asian habitats [63, 133]. Fossil distributions suggest that, by the end of the Neogene, the genusVitis was widely distributed in the Northern Hemisphere . As detailed in Nei et al. , the oldest reliable Vitis seeds are from the Paleocene (65.5-55.8 Ma) [53, 54] and were not detected in the preceding Cretaceous period. Important estimated time points in the Vitaceae diversifications were: 1) the divergence of Vitaceae and Leeaceae (stem age of Vitaceae), estimated by Magallón and Castillo  at 90.82 – 90.65 Ma, this estimate was based on a five gene data set (chloroplast rbcL, atpB,matK,and nuclear 18S and 26S nrDNA) obtained from GenBank, and conversion to absolute time using three fossil reference time points, 2) the divergence of the Ampelocissus-Vitis clade in the Tiffian stage of the Paleocene (62.0-56.8 Ma) based on fossil evidence synthesized in Nei at al. , and 3) the presence of well preserved Vitis seed at the late Neogene Gray Fossil site in Tennessee (7-4.5 Ma) .
The data sets supporting the results of this article are included within the article and its additional files. NCBI accession numbers: [Genbank: JX952227-JX960379, EMBL: HF544510-HF544512]. Additional file 1 lists the sequence accession number for all OTUs; Additional file 2 lists accession numbers by marker. Alignment, phylogenies, trees and BEAST .xml files were deposited at http://dx.doi.org/10.5061/dryad.s1s75.
We would like to thank Ashley Egan for supplying crucial sparks and expertise at several analytical key points of this project. Warren Lamboy worked extensively on data accuracy, management, and early analysis of datasets. Pablo Goloboff answered many basic and operational questions about TNT. Mark Miller and Wayne Pfeiffer among others at the San Diego Supercomputer Center gave very effective support to RAxML and endeavored to bring BEAST online at the CIPRES portal. Lauren Chan provided answers to specific questions in BEAST and Mr. Bayes. Peter Cousins supplied outgroup material and offered excellent advice on grape species identification and anything else related to grapes. Bob Nearpass provided local IT support, NCGR in Davis, CA supplied tissue samples. Anonymous reviewers provided suggestions that improved this work. Part of this work was carried out using the resources of the Computational Biology Service Unit from Cornell University which is partially funded by Microsoft Corporation. USDA is an equal opportunity provider and employer.
This project was funded by the United States Department of Agriculture, Agricultural Research Service, CRIS Project Number 1910-21000-020-00D. YW was supported by the China Scholarship Council (22861057), the Shaanxi Natural Science Foundation (No. 2004C103), and the Young Scientist Foundation of NWAFU (QN2009-013).
- Food and Agriculture Organization of the United Nations: FAO. [http://faostat.fao.org]
- Reisch BI, Pratt C: Grapes. Fruit breeding. Edited by: Janick J, Moore JN. 1996, New York: Wiley, 297-369. 2Google Scholar
- Comeaux BL, Nesbitt WB, Fantz PR: Taxonomy of the native grapes of North Carolina. Castanea. 1987, 52: 197-215.Google Scholar
- Galet P: Cepages et vignobles de France. Tome 1 Les vignes Americaines. 1988, Dehan: MontpellierGoogle Scholar
- Moore MO: Classification and systematics of Eastern North American Vitis L. (Vitaceae) north of Mexico. Sida. 1991, 14: 339-367.Google Scholar
- Chen Z, Hui R, Wen J: Vitaceae. Flora of China. 2007, 12: 173-222. PDF at [http://flora.huh.harvard.edu/china/mss/volume12/Vitaceae.pdf]Google Scholar
- Alleweldt G, Spiegel-Roy P, Reisch B: Grapes (Vitis). Genetic resources of temperate fruit and nut crops. Edited by: Moore JN, Ballington JR. 1990, Wageningen: Acta Hort 290, 291-337.Google Scholar
- Patel GI, Olmo HP: Cytogenetics of Vitis: I. The hybrid V. vinifera x V. rotundifolia. Am J Bot. 1955, 42: 141-159. 10.2307/2438464.Google Scholar
- Jaillon O, Aury J-M, Noel B, Policriti A, Clepet C, Casagrande A, Choisne N, Aubourg S, Vitulo N, Jubin C, Vezzi A, Legeai F, Hugueney P, Dasilva C, Horner D, Mica E, Jublot D, Poulain J, Bruyère C, Billault A, Segurens B, Gouyvenoux M, Ugarte E, Cattonaro F, Anthouard V, Vico V, Del Fabbro C, Alaux M, Di Gaspero G, Dumas V, et al: The grapevine genome sequence suggests ancestral hexaploidization in major angiosperm phyla. Nature. 2007, 449: 463-467. 10.1038/nature06148.PubMedGoogle Scholar
- Mullins MG, Bouquet A, Williams LE: The grapevine and its relatives. Biology of the grapevine. Edited by: Mullins MG, Bouquet A, Williams LE. 1992, Cambridge: Cambridge University Press, 17-36.Google Scholar
- Comeaux BL: Taxonomic studies on certain native grapes of the Eastern United States. PhD thesis. 1984, North Carolina State University: Department of Horticultural ScienceGoogle Scholar
- Rogers DJ, Rogers CF: Systematics of North American grape species. Am J Enol Vitic. 1978, 29: 73-78.Google Scholar
- Barrett HC, Cramer SG, Rhodes AM: A taximetric study of interspecific variation in Vitis. Vitis. 1969, 8: 177-187.Google Scholar
- Reinig WF: Die Holarktis. 1937, Jena: Gustav Fischer VerlagGoogle Scholar
- Arnold C, Rosetto M, McNally J, Henry RJ: The application of SSRs characterized for grape (Vitis vinifera) to conservation studies in Vitaceae. Am J Bot. 2002, 89: 22-28. 10.3732/ajb.89.1.22.PubMedGoogle Scholar
- Aradhya MK, Dangle G, Prins BH, Boursiquot J- M, Walker MA, Meredith CP, Simon CJ: Genetic structure and differentiation in cultivated grape, Vitis vinifera L. Genet Res, Camb. 2003, 81: 179-192. 10.1017/S0016672303006177.Google Scholar
- Aradhya MK, Koehmstedt A, Prins BH, Dangle GS, Stover E: Genetic structure, differentiation, and phylogeny of the genus Vitis: implications for genetic conservation. Acta Hort. 2008, 799: 43-49.Google Scholar
- Grassi F, Labra M, Imazio S, Ocete-Rubio R, Failla O, Scienza A, Sala F: Phylogeographical structure and conservation genetics of wild grapevine. Conserv Genet. 2006, 7: 837-845. 10.1007/s10592-006-9118-9.Google Scholar
- Ingrouille MJ, Chase MW, Fay MF, Bowman D, Van der Bank M, Bruijn ADE: Systematics of Vitaceae from the viewpoint of plastid rbcl DNA sequence data. Biol J Linn Soc. 2002, 138: 421-432. 10.1046/j.1095-8339.2002.00028.x.Google Scholar
- Lijavetzky D, Cabezas JA, Ibanez A, Rodriguez V, Martinez-Zapater JM: High throughput SNPdiscovery and genotyping in grapevine (Vitis vinifera L.) by combining a re-sequencing approach and SNPlex technology. BMC Genomics. 2007, 8: 424-10.1186/1471-2164-8-424.PubMed CentralPubMedGoogle Scholar
- Myles S, Boyko AR, Owens C, Brown PJ, Grassi F, Aradhya MK, Prins B, Reynolds A, Chia J-M, Ware D, Bustamante CD, Buckler ES: Genetic structure and domestication history of the grape. Proc Natl Acad Sci. 2011, 108: 3530-3535. 10.1073/pnas.1009363108.PubMed CentralPubMedGoogle Scholar
- Myles S, Chia J-M, Hurwitz B, Simon C, Zhong GY, Buckler E, Ware D: Rapid genomic characterization of the genus Vitis. PLoS One. 2010, 5 (e8219): 1-9.Google Scholar
- Péros J-P, Berger G, Portemont A, Boursiquot J-M, Lacomb T: Genetic variation and biogeography of the disjunct Vitis subg. Vitis (Vitaceae). J Biogeogr. 2011, 38: 471-486. 10.1111/j.1365-2699.2010.02410.x.Google Scholar
- Arroyo-Garcia R, Ruiz-Garcia L, Bolling L, Ocete R, Lopez MA, Arnold C, Ergul A, Soylemezoglu G, Uzun HI, Cabello F, Ibanez J, Aradhya MK, Atanassov A, Atanassov I, Balint S, Cenis JL, Costantini L, Goris-Lavets S, Grando MS, Klein B, McGovern PE, Merdinoglu D, Pejic I, Pelsi F, Primikirios N, Risovannaya V, Roubelakis-Angelakis KA, Snoussi H, Sotiri P, Tamhankar S, et al: Multiple origins of cultivated grapevine (Vitis vinifera L. ssp. sativa) based on chloroplast DNA polymorphisms. Mol Ecol. 2006, 15: 3707-3714. 10.1111/j.1365-294X.2006.03049.x.PubMedGoogle Scholar
- Rossetto M, Crayn DM, Jackes BR, Porter C: An updated estimate of intergeneric phylogenetic relationships in the Australian Vitaceae. Can J Bot. 2007, 85: 722-730. 10.1139/B07-022.Google Scholar
- Rossetto M, McNally J, Henry RJ: Evaluating the potential of SSR flanking regions for examining taxonomic relationships in the Vitaceae. Theor Appl Genet. 2002, 104: 61-66. 10.1007/s001220200007.PubMedGoogle Scholar
- Salmaso M, Faes G, Segala C, Stefanini M, Salakhutdinov I, Zyprian E, Toepfer R, Grando MS, Velasco R: Genome diversity and gene haplotypes in the grapevine (Vitis vinifera L.), as revealed by single nucleotide polymorphisms. Mol Breed. 2004, 14: 385-395. 10.1007/s11032-004-0261-z.Google Scholar
- Soejima A, Wen J: Phylogenetic analysis of the grape family (Vitaceae) based on three chloroplast markers. Am J Bot. 2006, 93: 278-287. 10.3732/ajb.93.2.278.PubMedGoogle Scholar
- Tröndle D, Schröder S, Kassemeyer H-H, Kiefer C, Koch MA, Nick P: Molecular phylogeny of the genus Vitis (Vitaceae) based on plastid markers. Am J Bot. 2010, 97: 1168-1178. 10.3732/ajb.0900218.PubMedGoogle Scholar
- Vezzulli S, Micheletti D, Riaz S, Pindo M, Viola R, This P, Walker MA, Troggio M, Velasco R: A SNP transferability survey within the genus Vitis. BMC Plant Biol. 2008, 8: 128-10.1186/1471-2229-8-128.PubMed CentralPubMedGoogle Scholar
- Wada EB: UC Davis, Plant Biology. Systematics and evolution of Vitis. PhD thesis. 2008, Ann Arbor: ProQuest LLC, Available through Proquest UMI Microform 3336358Google Scholar
- Wen J, Nie Z-L, Soejima A, Meng Y: Phylogeny of Vitaceae based on the nuclear GAI1 gene sequences. Can J Bot. 2007, 85: 731-745. 10.1139/B07-071.Google Scholar
- Aradhya M, Wang Y, Walker MA, Prins BH, Koehmstedt AM, Velasco D, Gerrath JM, Dangle GS, Preece JE: Genetic diversity, structure, and patterns of differentiation in the genus Vitis. Plant Syst Evol. 2013, 299: 317-330. 10.1007/s00606-012-0723-4.Google Scholar
- Nie Z-L, Sun H, Chen Z-D, Meng Y, Manchester SR, Wen J: Molecular phylogeny and biogeographic diversification of Parthenocissus (Vitaceae) disjunct between Asia and North America. Am J Bot. 2010, 97: 1342-1353. 10.3732/ajb.1000085.PubMedGoogle Scholar
- Liu X-Q, Ickert-Bond SM, Chen L-Q, Wen J: Molecular phylogeny of Cissus L. of Vitaceae (the grape family) and evolution of its pantropical intercontinental disjunctions. Mol Phylogenet Evol. 2013, 66: 43-53. 10.1016/j.ympev.2012.09.003.PubMedGoogle Scholar
- Zecca G, Abbott R, Sun W-B, Spada A, Sala F, Grassi F: The timing and mode of evolution of wild grapes (Vitis). Mol Phylogenet Evol. 2012, 62: 736-747. 10.1016/j.ympev.2011.11.015.PubMedGoogle Scholar
- Garcia-Pereira MJ, Caballero A, Quesada H: Evaluating the relationship between evolutionary divergence and phylogenetic accuracy in AFLP data sets. Mol Biol Evol. 2010, 27: 988-1000. 10.1093/molbev/msp315.PubMedGoogle Scholar
- Delsuc F, Brinkmann H, Philippe H: Phylogenomics and the reconstruction of the tree of life. Nat Rev Genet. 2005, 6: 361-375.PubMedGoogle Scholar
- Burleigh JG, Bansal MS, Eulenstein O, Hartmann S, Wehe A, Vision TJ: Genome-scale phylogenetics: inferring the plant tree of life from 18,896 gene trees. Syst Biol. 2011, 60: 117-125. 10.1093/sysbio/syq072.PubMed CentralPubMedGoogle Scholar
- Maureira-Butler IJ, Pfeil BE, Muangprom A, Osborn TC, Doyle JJ: The reticulate history of Medicago (Fabaceae). Syst Biol. 2008, 57: 466-482. 10.1080/10635150802172168.PubMedGoogle Scholar
- Doyle JJ, Doyle JL, Brown AHD: Incongruence in the diploid B-genome species complex of Glycine (leguminosae) revisited: Histone H3-D alleles versus chloroplast haplotypes. Mol Biol Evol. 1999, 16: 354-362. 10.1093/oxfordjournals.molbev.a026116.PubMedGoogle Scholar
- Sang T, Zhong Y: Testing hybridization hypotheses based on incongruent gene trees. Syst Biol. 2000, 49: 422-434. 10.1080/10635159950127321.PubMedGoogle Scholar
- Bruen TC, Philippe H, Bryant D: A simple and robust statistical test for detecting the presence of recombination. Genetics. 2006, 172: 2665-2681.PubMed CentralPubMedGoogle Scholar
- Lanier HC, Knowles LL: Is recombination a problem for species-tree analyses?. Syst Biol. 2012, 61: 691-701. 10.1093/sysbio/syr128.PubMedGoogle Scholar
- Huson DH, Bryant D: Application of Phylogenetic Networks in Evolutionary Studies. Mol Biol Evol. 2006, 23: 254-267. Software at [http://www.splitstree.org]PubMedGoogle Scholar
- Rokas A, Carroll SB: Bushes in the tree of life. PLoS Biol. 2006, 4: 1899-1904.Google Scholar
- Brandley MC, Warren DL, Leache AD, McGuire JA: Homoplasy and clade support. Syst Biol. 2009, 58: 184-198. 10.1093/sysbio/syp019.PubMedGoogle Scholar
- Wolfe KH, Li W-H, Sharp PM: Rates of nucleotide substitution vary greatly among plant mitochondrial, chloroplast, and nuclear DNA. Proc Natl Acad Sci USA. 1987, 84: 9054-9058. 10.1073/pnas.84.24.9054.PubMed CentralPubMedGoogle Scholar
- Degnan JH, Rosenberg NA: Discordance of species trees with their most likely gene trees. PLoS Genet. 2006, 2 (e68): 62-768.Google Scholar
- Wiens JJ: Polymorphism in systematic and comparative biology. Annu Rev Ecol Syst. 1999, 30: 327-362. 10.1146/annurev.ecolsys.30.1.327.Google Scholar
- This P, Lacomb T, Cadle-Davidson M, Owens C: Wine grape (Vitis vinifera L.) color associates with allelic variation in the domestication gene VvmybA1. Theor Appl Genet. 2007, 114: 723-730. 10.1007/s00122-006-0472-2.PubMedGoogle Scholar
- Philippe H, Brinkmann H, Lavrov DV, Littlewood DTJ, Manuel M, Wörheide G, Baurein D: Resolving Difficult Phylogenetic Questions: Why More Sequences Are Not Enough. PLoS Biol. 2011, 9: e1000602-10.1371/journal.pbio.1000602.PubMed CentralPubMedGoogle Scholar
- Brown RW: Paleocene flora of the Rocky Mountains and Great Plains. US Geol Surv Prof Pap. 1962, 375: 1-120.Google Scholar
- Mai DH: Neue früchte und samen aus Paläozänen ablagerungen Mitteleuropas. Feddes Repertorium. 1987, 98: 197-229.Google Scholar
- Xiang Q-Y, Soltis DE, Soltis PS, Manchester SR, Crawford DJ: Timing of the Eastern Asian-Eastern North American floristic disjunction: molecular clock corroborates paleontological estimates. Mol Phylogenet Evol. 2000, 15: 462-472. 10.1006/mpev.2000.0766.PubMedGoogle Scholar
- Wikström N, Savolainen V, Chase MW: Evolution of the Angiosperms: calibrating the family tree. Proc R Soc Lond B. 2001, 268: 2211-2220. 10.1098/rspb.2001.1782.Google Scholar
- Magallón S, Castillo A: Angiosperm diversification through time. Am J Bot. 2009, 96: 349-365. 10.3732/ajb.0800060.PubMedGoogle Scholar
- Nie Z-L, Sun H, Manchester SR, Meng Y, Luke Q, Wen J: Evolution of the intercontinental disjunctions in six continents in the Ampelopsis clade of the grape family (Vitaceae). BMC Evol Biol. 2012, 12: 17-10.1186/1471-2148-12-17. http://www.biomedcentral.com/1471-2148/12/17,PubMed CentralPubMedGoogle Scholar
- Gong F, Karsai I, Liu Y-S: Vitis Seeds (Vitaceae) from the late Neogene Gray fossil site, Northeastern Tennessee, U.S.A. Rev Palaeobot Palyno. 2010, 162: 71-83. 10.1016/j.revpalbo.2010.05.005.Google Scholar
- Manchester SR: Biogeographical relationships of North American Tertiary floras. Ann Missouri Bot Gard. 1999, 86: 472-522. 10.2307/2666183.Google Scholar
- Press F, Siever R: Earth. 1985, New York: W.H. Freeman and Company, 4Google Scholar
- Bowen GJ, Beerling DJ, Koch PL, Zachos JC, Quattlebaum T: A humid climate state during the Palaeocene/Eocene thermal maximum. Nature. 2004, 432: 495-499. 10.1038/nature03115.PubMedGoogle Scholar
- Qian H, Ricklefs RE: Large-scale processes and the Asian bias in species diversity of temperate plants. Nature. 2000, 407: 180-182. 10.1038/35025052.PubMedGoogle Scholar
- Wen J, Ickert-Bond S, Nie Z-L, Li R: Timing and modes of evolution of eastern Asian- North American biogeographic disjunctions in seed plants. Darwin’s Heritage Today: Proceedings of the Darwin 2010 Beijing International Conference. Edited by: Long M, Gu H, Zhou Z. 2010, Beijing: Higher Education Press, 252-269.Google Scholar
- Tiffney BH, Manchester SR: The use of geological and paleontological evidence in evaluating plant phylogeographic hypotheses in the northern hemisphere Tertiary. Int J Plant Sci. 2001, 162 (Suppl 6): S3-S17.Google Scholar
- Xiang Q-Y, Soltis DE, Soltis PS: The Eastern Asian and Eastern and Western North American floristic disjunction: congruent phylogenetic patterns in seven diverse genera. Mol Phylogenet Evol. 1998, 10: 178-190. 10.1006/mpev.1998.0524.PubMedGoogle Scholar
- Hartl DL, Clark AG: Principles of population genetics. 1989, Sunderland: SinaurGoogle Scholar
- DeLattin G: Über den Ursprung und die Verbreitung der Reben. Der Züchter. 1939, 11: 217-225.Google Scholar
- Wen J: Evolution of East Asian and Eastern North American disjunct distributions in flowering plants. Annu Rev Ecol Syst. 1999, 30: 421-455. 10.1146/annurev.ecolsys.30.1.421.Google Scholar
- Tiffney BH: The Eocene North Atlantic land bridge: its importance in Tertiary and modern phytogeography of the northern hemisphere. J Arnold Arbor. 1985, 66: 243-73.Google Scholar
- Hopkins DM: Cenozoic history of the Bering land bridge. Science. 1959, 129: 1519-1528. 10.1126/science.129.3362.1519.PubMedGoogle Scholar
- Marincovich L, Gladenkov AY: Evidence for an early opening of the Bering Strait. Nature. 1999, 397: 149-151. 10.1038/16446.Google Scholar
- Hewitt GM: The genetic legacy of the Quaternary ice ages. Nature. 2000, 405: 907-913. 10.1038/35016000.PubMedGoogle Scholar
- Planchon JE: Monographie des Ampelidae varies. Monographia Phanerogamerum. 1887, 5: 305-368.Google Scholar
- Munson TV: Foundations of American grape culture. 1909, Denison: TV Munson and SonGoogle Scholar
- Bailey LH: The species of grapes peculiar to North America. Gentes Herbarum. 1934, 3: 151-244.Google Scholar
- This P, Lacomb T, Thomas MR: Historical origins and genetic diversity of wine grapes. Trends Genet. 2006, 22: 511-519. 10.1016/j.tig.2006.07.008.PubMedGoogle Scholar
- McGovern PE: Ancient wine. The search for the origins of viticulture. 2003, Princeton: Princeton University PressGoogle Scholar
- Moore MO: Treatment from the Jepson manual. Vitaceae. [http://ucjeps.berkeley.edu/interchange/I_treat_indexes.html]
- De Queiroz K: Different species problems and their resolution. BioEssays. 2005, 27: 1263-1269. 10.1002/bies.20325.PubMedGoogle Scholar
- Nixon KC, Wheeler QD: An amplification of the phylogenetic species concept. Cladistics. 1990, 6: 211-223. 10.1111/j.1096-0031.1990.tb00541.x.Google Scholar
- Comeaux BL: Nomenclatural clarification of the name Vitis simpsonii Munson (Vitaceae). SIDA. 1987, 12: 279-286.Google Scholar
- Genetic Resources Information Network. [http://www.ars-grin.gov/npgs/acc/acc_queries.html]; [http://www.ars-grin.gov/cgi-bin/npgs/html/tax_search.pl?Vitis]
- Doyle JJ, Doyle JL: DNA isolation from leaves. Phytochemical Bulletin. 1987, 19: 11-15.Google Scholar
- Lodhi MA, Ye GN, Weeden NF, Reisch BI: A simple and efficient method for DNA extraction from grapevine cultivars and Vitis species. Plant Mol Biol Rep. 1994, 12: 6-13. 10.1007/BF02668658.Google Scholar
- Wheeler DL, Barrett T, Benson DA, Bryant SH, Canese K, Chetvernin V, Church DM, DiCuccio M, Edgar R, Federhen S, Geer LY, Kapustin Y, Khovayko O, Landsman D, Lipman DJ, Madden TL, Maglott DR, Ostell J, Miller V, Pruitt KD, Schuler GD, Sequeira E, Sherry ST, Sirotkin K, Souvorov A, Starchenko G, Tatusov RL, Tatusov TA, Wagner L, Yaschenko E: Database resources of the National Center for Biotechnology Information. Nucl Acids Res. 2006, 36 (suppl 1): D13-D21.Google Scholar
- Labate JA, Baldo AM: Tomato SNP discovery by EST mining and resequencing. Mol Breed. 2005, 16: 343-349. 10.1007/s11032-005-1911-5.Google Scholar
- Rozen S, Skaletsky HJ: Primer3: A software component for picking PCR primers. [https://sourceforge.net/project/showfiles.php?group_id=112461]; public WWW server at [http://Fokker.wi.mit.edu/primer3/input.htm]
- Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ: Basic local alignment search tool. J Mol Biol. 1990, 215: 403-410.PubMedGoogle Scholar
- Filatov DA, Proseq: A software for preparation and evolutionary analysis of DNA sequence data sets. Mol Ecol Notes. 2002, 2: 621-624. 10.1046/j.1471-8286.2002.00313.x.Google Scholar
- Cornish-Bowden A: Nomenclature for incompletely specified bases in nucleic acid sequences: recommendations. Nucleic Acids Res. 1985, 13: 3021-3030. 10.1093/nar/13.9.3021.PubMed CentralPubMedGoogle Scholar
- Chenna R, Sugawara H, Koike T, Lopez R, Gibson TJ, Higgins DG, Thompson JD: Multiple sequence alignment with the Clustal series of programs. Nucl Acids Res. 2003, 31: 3497-3500. 10.1093/nar/gkg500.PubMed CentralPubMedGoogle Scholar
- Baldo AM, Francis DM, Caramante M, Robertson LD, Labate JA: AlleleCoder: a PERL script for coding co-dominant polymorphism data for PCA. Plant genetic resources: characterization and utilization. 2011, 9: 528-530. 10.1017/S1479262111000839.Google Scholar
- Posada D, Crandal KA: The effect of recombination on the accuracy of phylogeny estimation. J Mol Evol. 2002, 54: 396-402.PubMedGoogle Scholar
- Drummond AJ, Rambaut A: BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol. 2007, 7: 214-222. 10.1186/1471-2148-7-214.PubMed CentralPubMedGoogle Scholar
- Kluge AG: A concern for evidence and a phylogentic hypothesis of relationships among Epicrates (Bovidae, Serpentes). Syst Zool. 1989, 38: 7-25. 10.2307/2992432.Google Scholar
- Simmons MP, Ochoterena H: Gaps as characters in sequencebased phylogenetic analyses. Syst Biol. 2000, 49: 369-381.PubMedGoogle Scholar
- Müller K: SeqState - primer design and sequence statistics for phylogenetic DNA data sets. Appl Bioinf. 2005, 4: 65-69. 10.2165/00822942-200504010-00008.Google Scholar
- Simmons M, Müller K, Norton A: The relative performance of indel-coding methods in simulations. Mol Phylogenet Evol. 2007, 44: 724-740. 10.1016/j.ympev.2007.04.001.PubMedGoogle Scholar
- Egan AN, Crandall KA: Incorporating gaps as phylogenetic characters across eight DNA regions: ramifications for North American Psoraleeae (Leguminosae). Mol Phylogenet Evol. 2008, 46: 532-546. 10.1016/j.ympev.2007.10.006.PubMedGoogle Scholar
- Wiens JJ: Combining data sets with different phylogenetic histories. Syst Biol. 1998, 47: 568-581. 10.1080/106351598260581.PubMedGoogle Scholar
- Huson DH: SplitsTree: analyzing and visualizing evolutionary data. Bioinformatics. 1998, 14: 68-73. 10.1093/bioinformatics/14.1.68.PubMedGoogle Scholar
- Huson DH, Bryant DF: Application of phylogenetic networks in evolutionary studies. Mol Biol Evol. 2006, 23: 254-276.PubMedGoogle Scholar
- Gradstein FM, Ogg JG, Smith AG, Agterberg FP, Bleeker W, Cooper RA, Davydov V, Gibbard P, Hinnov LA, House MR, Lourens L, Luterbacher HP, McArthur J, Melchin MJ, Robb LJ, Shergold J, Villeneuve M, Wardlaw BR, Ali J, Brinkhuis H, Hilgen FJ, Hooker J, Howarth RJ, Knoll AH, Laskar J, Monechi S, Plumb KA, Powell J, Raffi I, RÖhli U, et al: A geologic time scale 2004. 2005, Cambridge: Cambridge University PressGoogle Scholar
- International Stratigraphic Chart. http://www.stratigraphy.org/index.php/ics-chart-timescale,
- Drummond AJ, Marc A, Suchard MA, Dong Xie D, Rambaut A: Bayesian phylogenetics with BEAUti and the BEAST1.7. Mol Biol Evol. doi: 10.1093/molbev/mss075.Google Scholar
- Figtree. http://tree.bio.ed.ac.uk/software/figtree/],
- Drummond AJ, Ho SYW, Phillips MJ, Rambaut A: Relaxed phylogenetics and dating with confidence. PLoSBiology. 2006, 4: e88-Google Scholar
- HIV sequence database. [http://www.hiv.lanl.gov/content/sequence/findmodel/findmodel.html]
- Bruno WJ, Socci ND, Halpern AL: Weighted neighbor joining: a likelihood-based approach to distance-based phylogeny reconstruction. Mol Biol Evol. 2000, 17: 189-197. 10.1093/oxfordjournals.molbev.a026231.PubMedGoogle Scholar
- Yang Z: PAML 4: a program package for phylogenetic analysis by maximum likelihood. Mol Biol Evol. 2007, 24: 1586-1591. 10.1093/molbev/msm088.PubMedGoogle Scholar
- Posada D, Crandall KA: Modeltest: testing the model of DNA substitution. Bioinformatics. 1998, 14: 817-818. 10.1093/bioinformatics/14.9.817.PubMedGoogle Scholar
- Donoghue MJ: A critique of the biological species concept and recommendations for a phylogenetic alternative. Bryologist. 1985, 88: 172-181. 10.2307/3243026.Google Scholar
- Cracraft J: Species concepts and speciation analysis. Curr Ornithol. 1983, 1: 159-187.Google Scholar
- Stamatakis A: RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics. 2006, 22: 2688-2690. 10.1093/bioinformatics/btl446.PubMedGoogle Scholar
- Stamatakis A, Hoover P, Rougemont J: A fast bootstrapping algorithm for the RAxML web-servers. Syst Biol. 2008, 57: 58-771. 10.1080/10635150801898912.Google Scholar
- Huson DH, Richter DC, Rausch C, Dezulian T, Franz M, Rupp R: Dendroscope: An interactive viewer for large phylogenetic trees. BMC Bioinforma. 2007, 8: 460-10.1186/1471-2105-8-460.Google Scholar
- Huelsenbeck JP, Ronquist F: MRBAYES: Bayesian inference of phylogeny. Bioinformatics. 2001, 17: 754-755. 10.1093/bioinformatics/17.8.754.PubMedGoogle Scholar
- Rambaut A, Drummond AJ: Tracer V171. [http://beast.bio.ed.ac.uk/Main_Page]
- Goloboff PA, Farris JS, Nixon KC: TNT, a free program for phylogenetic analysis. Cladistics. 2008, 24: 774-786. 10.1111/j.1096-0031.2008.00217.x.Google Scholar
- Fitch WM: Toward defining the course of evolution: minimum change for a specified tree topology. Syst Zool. 1971, 20: 406-416. 10.2307/2412116.Google Scholar
- Giribet G: TNT: tree analysis using new technology. Syst Biol. 2005, 54: 176-178. 10.1080/10635150590905830.Google Scholar
- Nixon KC: The parsimony ratchet, a new method for rapid parsimony analysis. Cladistics. 1999, 15: 407-414. 10.1111/j.1096-0031.1999.tb00277.x.Google Scholar
- Goloboff PA: Analyzing large data sets in reasonable times: solutions for composite optima. Cladistics. 1999, 15: 415-428. 10.1111/j.1096-0031.1999.tb00278.x.Google Scholar
- Ronquist F: Ancestral areas and parsimony. Syst Biol. 1994, 43: 267-274.Google Scholar
- Li H-L: Floristic relationships between Eastern Asia and Eastern North America. Trans Am Philos Soc. 1952, 42: 371-429. 10.2307/1005654.Google Scholar
- Thorne RF: Major disjunctions in the geographic ranges of seed plants. Q Rev Biol. 1972, 47: 365-411. 10.1086/407399.Google Scholar
- Boufford DE, Spongberg SA: Eastern Asian- Eastern North American phytogeographical relationships – a history from the time of Linnaeus to the twentieth century. Ann Missouri Bot Gard. 1983, 70: 423-439. 10.2307/2992081.Google Scholar
- Tiffney BH: Perspectives on the origin of the floristic similarity between Eastern Asia and Eastern North America. J Arnold Arbor Harv Univ. 1985, 66: 73-94.Google Scholar
- Fennel JL: Inheritance studies with the tropical grape. J Hered. 1948, 39: 54-66.Google Scholar
- Comeau BL: A new Vitis (Vitaceae) from Vera Cruz, Mexico. Sida. 1987, 12: 273-277.Google Scholar
- Wan Y, Schwaninger HR, Li D, Simon CJ, Wang Y, He P: The eco-geographic distribution of wild grape Germplasm in China. Vitis. 2008, 47: 77-80.Google Scholar
- Hsu J: Late Cretaceous and Cenozoic vegetation in China, emphasizing their connections with North America. Ann Missouri Bot Gard. 1983, 70: 490-508. 10.2307/2992084.Google Scholar
- Milne RI, Abbott RJ: The origin and evolution of Tertiary relict floras. Adv Bot Res. 2002, 38: 281-314.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.