https://orcid.org/0000-0003-0770-4177
Figures
Abstract
The grape genus Vitis L. includes the domesticated V. vinifera, which is one of the most important fruit crop, and also close relatives recognized as valuable germplasm resources for improving cultivars. To resolve some standing problems in the species relationships within the Vitis genus we analyzed diversity in a set of 90 accessions comprising most of Vitis species and some putative hybrids. We discovered single nucleotide polymorphisms (SNPs) in SANGER sequences of twelve loci and genotyped accessions at a larger number of SNPs using a previously developed SNP array. Our phylogenic analyses consistently identified: three clades in North America, one in East Asia, and one in Europe corresponding to V. vinifera. Using heterozygosity measurement, haplotype reconstruction and chloroplast markers, we identified the hybrids existing within and between clades. The species relationships were better assessed after discarding these hybrids from analyses. We also studied the relationships between phylogeny and morphological traits and found that several traits significantly correlated with the phylogeny. The American clade that includes important species such as V. riparia and V. rupestris showed a major divergence with all other clades based on both DNA polymorphisms and morphological traits.
Citation: Péros J-P, Launay A, Peyrière A, Berger G, Roux C, Lacombe T, et al. (2023) Species relationships within the genus Vitis based on molecular and morphological data. PLoS ONE 18(7): e0283324. https://doi.org/10.1371/journal.pone.0283324
Editor: Branislav T. Šiler, Institute for Biological Research, University of Belgrade, SERBIA
Received: November 14, 2022; Accepted: March 7, 2023; Published: July 31, 2023
Copyright: © 2023 Péros et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interest exist.
Introduction
The North temperate Vitis genus (Vitaceae family) includes a total of 60 to 70 extant species. It is divided into two subgenera, Muscadinia and Vitis that differ in chromosome number (n = 20 vs n = 19) and several morphological characters [1, 2]. Subgenus Vitis includes a single european species Vitis vinifera subsp. vinifera, which has been domesticated from V. vinifera subsp. sylvestris [3] and is now widely cultivated to produce wine, fruit juice and fresh or dry raisins. The other species are distributed in either eastern Asia or North America [4]. Some of the wild species are cultivated for fruit production for instance V. amurensis in China and M. rotundifolia in the United States. Others, such as the American species V. riparia and V. rupestris, have been used worldwide as rootstocks. However, the wild species have mostly been used as parents in crosses aiming to improve the resistance to diseases or confer adaptative traits both in scion and rootstocks cultivars [5–7].
Breeding interspecific hybrids is made easier because most Vitis species are interfertile. On the other hand, natural and past artificial hybridization could also complicate species delimitation [4, 8]. For instance, based on morphology, ecology and geography, the classifications of subg. Vitis by Planchon [9], Munson [10], Bailey [11], Galet [12] and Moore & Wen [2] are not in full agreement for the number of series, the placement of species in series and also for distinguishing hybrids from pure species. Furthermore, these classifications used different synonymous names and apart those of Planchon [9] and Galet [12] mainly concerned the North American species. Chinese Vitis species taxonomy has been more recently investigated [8, 13, 14] and also revealed problems of species boundaries and synonymies. These difficulties in the taxonomic treatment when using morphological and anatomical characters have been considered due in part to hybridization and clinal variation [1, 4, 15, 16].
Recent molecular studies allowed clarifyig the classification of subg. Vitis. It has been shown that the Eurasian species (including V. vinifera) and North American species constitute two well-separated clades each containing several subclades [17–28]. The two main clades diverged recently between the middle-late Miocene and the late Pliocene [28] and most studies agree on that partition. However, the species relationships within clades vary according to the analyzed sample, the type of data (i.e. chloroplast vs. nuclear, single genes vs. SNP arrays or NGS data) and the method used for phylogeny reconstruction. Such difficulties may result from the sharing of ancestral polymorphisms due to recent divergence and reticulation due to historical recombination [22]. For this last point, recent studies highlighted hybridization within the Asian clade [29] and within the American one [30]. The misidentification of accessions and the presence of hybrids in the studied samples may add background noise that could result in weakly supported nodes in phylogenetic trees.
Leaf morphology broadly varies among species and varieties in the Vitis genus [31]. Vitis leaves differ for instance in their degree of lobing, in the lengths and angles of their five major veins and in margin dentation. Ampelographic measurements have therefore been largely used to discriminate between Vitis vinifera cultivars [31–36], and more recently Vitis species [37–39]. Moreover, Chitwood et al [40] found evidence for an important genetic component determining leaf shape and venation patterning in grape. Therefore, besides their interest for taxonomy, some morphometric measures may be useful in the breeding of new varieties with adaptative traits for the undergoing climatic changes. In a recent work, Ma et al. [41] used a phylogenomic framework within the Vitis genus to decipher the evolution of several morphological characters, e.g. tendril architecture, leaf shape, and type of trichomes. They found several examples of convergences, and revealed in the Asian V. bryoniifolia subclade the presence of cryptic species having similar morphology and distribution.
In this study, we assessed genetic and morphological diversity in the Vitis genus throughout its geographic range within a comprehensive sample of species and putative hybrids. Our main objectives were 1) to clarify some standing problems in the Vitis phylogeny and 2) to evaluate the importance of the phylogenetic signal in morphological characters in comparison with convergence due to similar environments. Analyses helped to clarify the relationships among species and allowed hybrids identification. Morphological variation was found both within species and among clades and several traits were found autocorrelated with the phylogeny.
Materials and methods
Plant material and DNA extraction
A total of 90 Vitis accessions maintained at the INRAE repository of Vassal-Montpellier (Marseillan, France, www6.montpellier.inrae.fr/vassal_eng) were used during this study (S1 Table). This set included 13 Asian and 56 North American accessions representing 12 and 18 putative species, respectively. For Vitis vinifera we retained a set of eight cultivars (subsp. vinifera) amongst the most popular French ones and a set of eight wild accessions (subsp. sylvestris) mostly originated from French populations. Muscadinia rotundifolia was added as an outgroup with five accessions. Most of this germplasm has been studied and precisely identified [12] but seven accessions (THU2, YES, AES5, AES6, ARZ2, TRE, TRE2, S1 Table) were more recently introduced to increase the number of species or the number of accessions for some species. Total genomic DNA was extracted from young leaf tissues using the DNeasy 96 plant mini kit (Qiagen Inc., Valencia, CA, USA). DNA concentration and purity was assessed using a NanoDrop 8000 UV-Vis spectrophotometer (Thermo Scientific).
Loci sequencing
Twelve nuclear sequences (S2 Table) were amplified and sequenced in 78 accessions (S1 Table) according to Péros et al [42]. Primers were either already published or specifically designed for this study using Primer3 software [43]. GAI1 [44] and TFL1 [45] genes are involved in grape development and GAI1 has already been used in grape phylogenetics [46]. LDOX, CHI1 and DFR4 code for key enzymes in the flavonoid biosynthetic pathway in grape [47]. TC1-A and TC1-B are located in a large intron of around 1,500 bp that contains several repeat motifs. As Vitis species are highly heterozygous, several steps were needed to design primers to partly sequence this intron. The five other amplicons were selected among a large set of amplicons designed to obtain a SNP-based diversity map in the Vitis genus (INRAE project SNPGrapeMap). Polymerase chain reaction (PCR) was performed in 25 μL volumes using the following protocol for all sequences: denaturation at 94°C for 3 min, 35 cycles at 94°C for 30 s, 55°C for 1 min, and 72°C for 1 min 30 s, followed by 6 min elongation at 72°C and 10°C soak. Samples containing 1 μL of PCR products and 10 μl of de-ionized water were run on a capillary sequencer ABI Prism 3700 XL (Applied Biosystems, CA, USA). Chromatograms of the forward and reverse strands were aligned and edited with deletion of indels using the software CLC Sequence Viewer 6.0 (CLC bio then Qiagen) to generate alignment files in fasta format.
SNP array genotyping
The GrapeReseq 20 K SNPs array [48] was used to genotype 78 accessions (S1 Table) according to Illumina GoldenGate assay protocols [49] at the GenoToul platform (Toulouse, France). DNA from ‘PN40024’ corresponding to the reference genome [50] was included as control. Genotyping was performed using the Genotyping Module v1.9 of the Illumina GenomeStudio Data Analysis software. Manual curation and several filters were applied following Laucou et al [51]. We finally used the genotyping at 12,971 SNPs with only 2.04% of missing data for the 78 accessions. Among these SNPs 83.06%, 14.35% and 2.57% were discovered in V. vinifera, several American Vitis and M. rotundifolia, respectively (S3 Table).
Phylogenetic analyses
Chromatograms of forward and reverse strands were aligned and edited using the software CLC Sequence Viewer 6.0 (CLC bio then Qiagen). All positions corresponding to indels in one or several individuals were deleted. Final alignments were generated in fasta format and analyzed separately or after concanetation. Phylogenetic analyses were first conducted using MEGA7 [52]. Other analyses were performed in the R environment 3.6.0 [53] using the following R packages: apex v.1.0.3 [54], phangorn v.2.5.5 [55], ape v.5.3 [56] and adegenet v.2.1.1 [57]. Phased genotypes were obtained using the algorithm GEVIL of GEVALT software [58] implemented within SNIPlay software [59]. Haplotype networks were then constructed using the Median Joining method implemented in PopART software [60].
Ampelometric measures
Leaves from 70 accessions were collected at the end of spring at Vassal-Montpellier repository. Four to 18 fully expanded leaves per accession (median number = 8.5) at similar developmental stage were taken from the middle of shoots. Leaves were dried in a plant press between newsprint sheets. The abaxial surface of each leaf was photographed using a camera mounted on a copy stand. Parameters were obtained from these pictures using SuperAmpelo ver. 2.0 [61] and corresponded to lengths, distances, angles and calculated values (area, ratio). Median values per accession were considered and averaged for parameters measured on the right and left halves of leaves. In addition, photographs were processed using ImageJ [62] to obtain from binary images measurements of the aspect ratio AR and circularity (Circ = (4π × (area/perimeter2) [40]. The 39 parameters were grouped according to three main types of quantitative variables: length and distances (n = 16), angles (n = 10) and ratios (n = 13); they are listed in S4 Table and illustrated in S1 Fig.
OIV descriptors
Accessions were observed directly in the vineyard (INRAE repository of Vassal-Montpellier) for a total of 46 OIV descriptors [63] assessed at different development stages of leaves and shoots (S5 Table). Observations had already been recorded across several past vegetative seasons and were completed during this study to avoid missing data. These descriptors were grouped according to three types of traits: morphology (n = 15), color (n = 13), and pilosity (n = 18).
Morphometric analyses
Analyses were performed without hybrid accessions (known or identified) and for accessions having all three types of data (sequence data, leaf measures, OIV codes). This subsample comprised 56 accessions and represented 24 species or subspecies. Leaf measurements and OIV descriptors were analyzed separately using Principal Component Analysis (PCA) implemented in the R package FactoMineR v.1.42 [64]. Results were visualized using the R package factoextra v1.0.5 [65]. We also analyzed the phylogenetic signal (i.e. phylogenetic autocorrelation) present in each data set using the R package adephylo v1.1.11 [66]. The non-parametric Abouheif’s test [67] was used to detect traits with significant autocorrelation based on a randomization test with 999 permutations. Then, a phylogenetic principal component analysis (pPCA) was performed to summarize our sets of traits into a few synthetic variables showing positive (global structure) or negative (local structures) phylogenetic autocorrelation in the phylogenetic tree. This latter was that previously obtained using the alignment of the 12 loci.
Results
Species relationships based on sequence data
The final aligment for 12 loci and 78 individuals was 6,771 bp in length. MEGA7 detected 434 variables sites including 261 parcimony informative sites and 173 singletons. However, heterozygous data (IUAPC codes: K, M, R, S, W, Y) were not considered by the software and the level of polymorphism was therefore underestimated using the diploid data. The Maximum-Likehood tree with the bootstrap values calculated from 100 replicates were exported in newick format and then edited using package ape in R (Fig 1). The outgroup Muscadinia was well separated from the accessions of subgenus Vitis and among these latter the American species V. californica had a separate position. The Eurasian accessions were grouped in two sister clades corresponding to the East Asian species (EA) and to wild and cultivated V. vinifera (EU). Other North American accessions were distributed in two separate clades: NA1, sister to the Eurasian clades, and NA2. Nodes were generally poorly supported, with the exception of the V. coignetiae-V. thunbergii-V. amurensis group in the EA clade, the V. cinerea complex in the NA1 clade and the V. trealesii-V. arizonica group in the NA2 clade. Cultivated and wild accessions of V. vinifera were separated with the exception of cv. Syrah, which grouped with the wild accessions. The position of the hybrid between V. vinifera and V. labrusca (LAB, cv. Isabelle) occupied an intermediate position between EU clade and NA1, which contained the pure V. labrusca species (LABv). By contrast, the other known hybrids V. × champinii (CHA) and V. × doaniana (DOA1) were imbedded within the NA1 clade. The second putative accession of the East Asian V. thunbergii (THU2) occupied a basal position in the NA1 clade suggesting an hybrid parentage.
Vitis accessions (codes in S1 Table) were grouped in separated clades: EA (East Asia, blue), EU (Europe, spring green: V. vinifera subsp. vinifera, dark green: V. vinifera subsp. sylvestris), NA1 (North America 1, orange), NA2 (North America 2, red), V. californica (brown) and M. rotundifolia (grey). Known hybrids were figured in pink (CHA, DOA1, LAB). The tree is drawn to scale, with branch lengths measured by the number of substitutions per site. The percentage of trees in which the associated accessions clustered together is shown at nodes with threshold at 70 and 90%.
To assess genetic diversity and identify the hybrids existing in our sample, haplotype sequences were inferred from genotype data for each individual locus. The number of SNPs per locus varied from 29 (LDOX) to 106 (CHI1) for a total of 674 SNPs (Table 1) indicating that 240 polymorphic positions (674 minus 434) were not taken into account in the ML tree (Fig 1). The number of haplotypes per locus varied from 27 to 66 with similar diversity statistics among loci; in average, ThetaW, Pi and He were 0.0214 (range 0.0118–0.0387), 0.0106 (range 0.0063–0.0177) and 0.92 (range 0.863–0.953), respectively (Table 1). Using a custom R script, we calculated the number of differences between the two haplotypes for each accession. The values averaged over the 12 loci were figured as boxplots for each group identified in the ML tree (S2 Fig). Outliers comprised, as expected, the recognized hybrids V. × labrusca (LAB), V. × champinii (CHA) and V. × doaniana (DOA1) and other outliers might therefore be also suspected to be hybrids: V. coriacea in the NA1 group, one accession of V. rupestris and V. girdiana in the NA2 group, and accessions assigned to V. yeshanensis, V. thunbergii and V. piazeskii in the East Asian group (EA).
To clarify the origin of these putative hybrids we analyzed the haplotype networks for each of the 12 loci. From the network obtained with locus 2351A (Fig 2), it clearly appeared that V. girdiana resulted from hybridization between V. vinifera and an American species, whereas both V. thunbergii (THU2) and V. yeshanensis were hybrids between V. vinifera and Asian species. The high heterozygosity of V. × labrusca (cv. Isabelle) was due to a well-documented artificial cross between V. labrusca and V. vinifera. High heterozygosity in V. × champinii and V × doaniana resulted from natural hybridization between species belonging to the two distinct American clades NA1 and NA2. Networks obtained for other loci were shown in S3 Fig. Sequences such as GAI1 (S3A Fig), CHI1 (S3B Fig) and 4275A (S3C Fig) indicated that both hybrids shared one haplotype with V. candicans whereas their second haplotype belonged to an American species of clade NA2, most probably V. rupestris for V. × champinii and V. riparia for V. × doaniana.
Haplotypes of known (CHA, DOA1, LAB) and putative (GIR, THU2, YES) hybrids (colored in pink) occupied separate position in the network. Other Vitis haplotypes were colored as in Fig 1 considering the clade to which they belong.
We constructed another ML tree without the known or newly identified hybrid accessions (Fig 3), with the aim to better depict the relationships among pure species. Branch supports were not greatly improved but some changes in accession placement were observed. Within the Eurasian clade, V. piazeskii (PIA) now occupied a sister position in the EA clade and the two subspecies of V. vinifera were separated although without strong support. In the NA1 clade, the V. cinerea var. cinerea and var. berlandieri accessions constituted a well-separated branch with V. cordifolia (COD) in a sister position. Another clade was represented by V. aestivalis and V. labrusca, whereas V. candicans (CAN) was a distant species at the base of the NA1 clade. This part of the tree might be obscured by a possible hybridization between V. aestivalis (AES) and V. candicans (CAN) either with V. simpsonii (SIM) or more probably V. coriacea (COR) as suggested by the high heterozygosity in this latter accession (S2 Fig). The arizonica-rupestris-riparia clade within the NA2 clade also included V. longii (LON, syn. V. acerifolia) and V. novo mexicana (MEX). V. rubra (RUB, syn. V. palmata) and V. monticola (MON) occupied a basal position in this clade. The position of V. californica was not changed after the exclusion of hybrids and was always clearly separated from other accessions of the subgenus Vitis.
Six Vitis hybrids were not included in this analysis (CHA, DOA1, LAB, GIR, THU2, YES). Same legend than Fig 1.
Species relationships based on SNP array
The full data set included 70 accessions genotyped at 12,971 SNPs. The distance tree constructed with the NJ algorithm confirmed the accessions placement of in the four clades identified using the sequence data but with a higher number of highly supported nodes (Fig 4). Hybrids that had a vinifera parent clearly appeared at the base of the vinifera clade. In addition to LAB (cv. Isabelle), GIR, THU2 and YES, they included two accessions without sequence data: another V × labrusca hybrid (LAB2, cv. Concord) and a falsely identified putative accession of V. aestivalis (AES6). Since around 83% of the array SNPs originated in V. vinifera (S3 Tab), branch lengths were very long for V. vinifera accessions and hybrids with this species. On the contrary, branch lengths were short in the other clades especially in NA2 that was not represented in the discovery panel. The East Asian clade also was not represented in the discovery panel but some SNPs discovered in V. vinifera might be expected to exist in the East Asian accessions. As another consequence of this ascertainment bias, the mean observed heterozygosity was high and highly variable in the European clade, but low and less variable in the other clades, especially in NA2. Hybrids with V. vinifera again corresponded to outliers in the box-plots (S4 Fig), whereas V. × champinii (CHA) and V. × doaniana (DOA1), although not identified as outliers, presented the highest values of heterozygosity within the NA1 clade.
Colors used for hybrids and clades as in Fig 1. The percentage of trees in which the associated accessions clustered together is shown at nodes with threshold at 70 and 90%.
We then excluded the hybrids to construct another distance tree (Fig 5A). Differences in branch length were clearly visible but clade arrangement was not evident especially for the species occupying basal positions in the clades. To reduce the effect of ascertainment bias and therefore better visualize the relationships between non-vinifera accessions, we constituted three other data set with decreasing fractions of vinifera SNPs sampled at random. The new SNP sets had 5,000 (69.4%), 1,000 (31.3%) and 500 (18.5%) vinifera SNPs, respectively. Decrease in the percentage of vinifera SNPs did not change the species relationships but allowed visualizing better the relationships in non-vinifera accessions by modifying branch lengths (Fig 5B–5D). V. amurensis appeared basal in the Eurasian clade (EA+EU). For American taxa, except the clearly differentiated clades such as aestivalis or cinerea-berlandieri in NA1 or arizonica-rupestris-riparia in NA2, the position of V. californica, V. candicans, V. rubra and V. monticola were not fully congruent with the relationships depicted using sequence data (Fig 3).
a) 12,971 SNPs with 10,774 V. vinifera SNPs, b) 7,197 SNPs with 5,000 V. vinifera SNPs), c) 3,197 SNPs with 1,000 V. vinifera SNPs, and d) 2,697 SNPs with 500 V. vinifera SNPs. Hybrids are not included. Vitis clades colored as in Fig 1. The percentage of trees in which the associated accessions clustered together is shown at nodes with threshold at 70 and 90%.
Polymorphim in chloroplast DNA
A total of 23 SNPs in the array corresponded to polymorphisms in chloroplast DNA (Ibanez, pers. comm.). Only 11 of these loci were informative (biallelic) in our sample of 70 accessions and allowed defining 19 chlorotypes with 1 to 8 SNP differences and corresponding to 1 to 16 accessions (S5A Fig). The hybrids THU2 and YES shared their chlorotype with accessions from the European clade and with V. californica, therefore they issued most probably from a hybridization between an Asian species as female and V. vinifera as male. The GIR and LAB (cv. Isabelle) hybrids shared their chlorotype with American accessions and therefore V. vinifera was again the male parent. The CHA and DOA1 hybrids had a chlorotype differing from V. candicans so their female parent was more probably V. rupestris for CHA and V. riparia for DOA1 according to the nuclear results. Using the mlg.filter function in package poppr we collapsed chlorotypes having only one difference, which resulted in eight chlorotypes (S5B Fig). This allowed suggesting an American species as female parent for AES6 and V. vinifera for LAB2 (cv. Concord).
PCA analysis of morphological traits
In the PCA with the 39 leaf measures (Fig 6), the two first axes explained 60.7% of the variance. PC1 (35% of the variance) reflected the effect of many length variables (Fig 6A) and separated the accessions according to their leaf size (Fig 6B). Species with large leaves such as V. coignetiae (COI), V. romanetii (ROM), V. riparia (RIP) and some cultivars of V. vinifera (e.g. Carignan, CAR) were distinguished from species with small leaves such as V. monticola (MON) and V. rupestris (RUP). The variables most contributing to PC2 (25.7% of the variance) were angles and calculated ratio; both separated the accessions according to leaf form. For instance, the distance SPSP’ and the angle pi both allowed distinguishing species with a large sinus opening: V. rupestris (RUP) and V. riparia (RIP). The angles between the three main veins N1,N2 and N3 and especially their sums grouped accessions having lobes (mostly the sampled vinifera accessions). Finally, some calculated variables such as the ratio OMETN2N3 (omega+eta/lengthN2+lengthN3) also appeared useful to define the leaf form and was only correlated with OMETOSOI (omega+eta/OS+OI) which depended on sinuses depth. In addition to these general trends, a rather large variation was noticeable within species having several accessions. Another PCA was performed, keeping only the angles and calculated variables (except leaf area) to better interpret the differences in leaf form among species (S6 Fig). V. rupestris was easily distinguished on PC1 (38.1% of the variance) from all other species due to its reniform leaves (high leaf width relative to length AR, high values of mu and lambda) with large open petiolar sinus (high values of pi). Other species were separated along PC2 (23.0% of the variance) according to variables measuring lobbing and serration such as Circularity (Circ.) and sinus depth (RS, RI, OMETOSOI). V. vinifera and some Asian species such as V. piasezki and V. amurensis) were lobed and serrated in contrast to other Asian species such as V. armata, V. pentagona, V. balanseana, whose had entire cordiform leaves.
a) correlation plot for variables classified according to their type (length, angle and ratio, see S4 Table) b) plot for Vitis accessions grouped by clades identified using molecular data and colored as in Fig 1. Hybrids were excluded.
For the traits coded using OIV scales, the two first PCA axes explained 31.7% of the variance (19.7% for PC1 and 12% for PC2) and the contribution of most variables was weak (Fig 7A). However, the different genetic clades and also species within clades were clearly distinguished (Fig 7B). PC1 separated species from the NA2 clade whereas PC2 tended to separate NA1 to EU and EA clades. Opening of the shoot tip (M1 = OIV001), surface relief (M102 = OIV102) and several OIV descriptors measuring organ pilosity were the most contributing traits to PC1, the most villous species of clade NA1 (V. candicans, V. coriacea, V. simpsonii) and the glabrous species of clade NA2 such as V. riparia and V arizonica were at opposite sides of PC1. In contrast to the previous analyses with leaf measures, accessions from the same species or species subgroup within clades showed lower variation and clustered in the PCA plot (Fig 8B). This was the case, for instance, for V. cinerea var. cinerea accessions or for the group of Asian species having spines on internodes (V. armata, V. romanetii, V. davidii) although this latter characteristic was not encoded by OIV.
a) correlation plot for OIV codes classified according to their type (M: Morphology, C: Color, P: Pilosity, see S5 Table) b) plot for Vitis accessions grouped by clades identified using molecular data and colored as in Fig 1. Hybrids were excluded.
a) eight leaf measurements, b) ten OIV codes. The first global principal component of the pPCA (PC1) is shown. Positive and negative values are represented using black and white symbols, respectively, with a size proportional to the absolute value. In the phylogenetic tree clades are colored as in Fig 1. Hybrids were excluded.
Searching for a phylogenetic signal
We found 21 and 4 leaf measures with significant autocorrelation at p<0.01 and 0.05, respectively, on a total of 39 measures (listed in S4 Table). The results of pPCA indicated that the variation in these traits corresponded to a global signal (S7A Fig) and the main component (PC1) isolated V. riparia from the other species (S8A Fig). We then selected the measures having the highest loadings in the analysis (quantile (a,0.1) | a>quantile (a,0.9)) and drew a plot (Fig 8A) using only the eight selected variables, some of which being correlated (S9A Fig). These traits were characteristics of leaf form: distance between F and N2 extremity (FN2), angle between veins 1 and 3 (ANGN1N3: omega + eta), ratios of leaf width measured at vein extremities (RN2N3, RN2N4), angle mu and ratio between major and minor axis (AR).
We found 30 and 5 OIV descriptors with significant autocorrelation at p<0.01 and 0.05, respectively, on a total of 46 descriptors (S6 Table). The results of pPCA indicated that the variation corresponded to a global signal (S7B Fig) and the main component (PC1) isolated the whole NA2 clade + M. rotundifolia from all the other species (S8b Fig). We then selected the variables as above and drew another plot using the 11 variables having the highest loadings (Fig 7B), some of which being correlated (S9B Fig). The most structured traits corresponded to all three categories: morphology (M1 = OIV001: opening of the shoot tip; M102 = OIV102: surface relief, M73 = OIV073: undulation between veins), coloring (C2 = OIV002: distribution of anthocyanin coloration on prostrate hairs of the shoot; C7 = OIV007: color of the dorsal side of internodes; C9 = OIV009: color of the dorsal side of nodes) and villosity (P4 = OIV004: density of prostrate hairs on the shoot tip; P5 = OIV005: density of erect hairs on the shoot tip, P53 = OIV053: density of prostrate hairs between main veins on lower side of blade (4th leaf); P56 = OIV56: density of erect hairs on main veins on lower side of blade (4th leaf)).
Discussion
The Vitis genus comprises several clades
In the last decade, numerous efforts have been made to obtain a comprehensive description of species relationships within the Vitis genus. Phylogenetic studies have concerned both the chloroplast [17–19, 26, 28] and nuclear [18–25, 27, 47] genome using different sets of markers and accessions. All these works agreed with the sister position of subg. Muscadinia relative to subg. Vitis and most of them concluded that subg. Vitis may be divided in several clades corresponding to the continental origin of the species. However, incongruent results have been frequently observed for the placement of either clades in subg. Vitis or species within clades. These discrepancies have been attributed to non-exclusive reasons such as recent divergence, natural hybridization or false identification of accessions. Another additional explanation might be the type of markers or the manner they were obtained. For instance, previously widely used markers such as Simple Sequence Repeats (SSR) were prone to homoplasy [68]. More recent technologies also have some biases, due to the procedures used to select SNPs in the case of SNP arrays [69, 70] and to the part of reduced genome examined by genotyping-by-sequencing [71]. Using a very large number of markers issued from whole-genome sequencing did not always guaranty a more precise description of the species relationships. For instance in grape, V. cinerea and V. aestivalis were not recovered as monophyletic in a whole-plastome tree [26]. As another example, the relationships among American species were poorly described within the ML tree obtained after the complete nuclear genome sequencing of 472 Vitis accessions [27, S3 Fig]. Quality of sequences, parametrization in the bioinformatic process, use of a reference sequence from V. vinifera and high heterozygosity in nuclear data could introduce additional difficulties. Furthermore, to assess species relationships even in a recent divergent genus, using several thousand or millions of markers do not seem necessary in contrast for example to genetic association studies [72].
In this work, we first used Sanger sequencing to discover polymorphims in a limited number of loci. ML tree based on these polymorphims confirmed i) the outgroup position of M. rotundifolia, ii) the peculiar position of V. californica, iii) the existence of two other clades in North America, iv) a clade comprising all Asian species and v) the derived position of V. vinifera in Eurasia. In addition, accessions of the same species generally clustered together. We then used an SNP array comprising around thirty times more SNPs than the Sanger data and obtained a similar phylogeny with a higher support for the main nodes. We demonstrated that the ascertainment bias due to the discovery panel had large effects on branch lengths with a lower discrimination of species in clades not included in the panel (e.g. North American clade 2). Moreover, some species (e.g. V. rubra, V. monticola, V. candicans) having a sister position in clades according to the ML tree (Sanger data) exhibited an ambiguous position in the distance tree (SNP array data). This could be due to the absence in the array of the polymorphisms specific of these more distant species. We therefore considered that the ML tree obtained with sequence data better reflected the species relationships.
Hybridization in Vitis genus
Natural hybrids in the genus Vitis have long been recognized based on morphological and agronomic traits, especially in North America [73]. For instance, Foex [74] suggested, as early as 1882, that V. × champinii was a hybrid combining the vigor of V. candicans and the rooting ability and hardiness of V. rupestris. Several parentages proposed in the past have indeed been supported by recent molecular data: Zecca et al [30] using data from the 9K SNP array [75] obtained evidence that V. × champinii and V. × doaniana resulted from natural crosses between V. candicans and, respectively, V. rupestris and V. acerifolia. In the present work, we reached a similar conclusion based on haplotype networks of a few nuclear loci for V. × champinii, whereas fo V. × doaniana the second parent might be V. riparia or the close species V. longii (syn. V. acerifolia). Moreover, cpDNA polymorphisms allowed us to infer the direction of crosses. Their hybrid origin might not be deduced form their position in the ML-tree since they clustered with one of their parent (V. candicans) within the same clade (NA1). By contrast, artificial hybrids between V. vinifera and V. labrusca (also called V. labruscana) occupied intermediate positions in the ML and distance trees. Our cpDNA data showed that the female parent is V. labrusca for cv. Isabelle and that the chlorotype of cv. Concord originated from V. vinifera. This latter result confirms the recent study of Wen et al. [76] who also pointed out a probable backcross with V. labrusca as already suggested by Huber et al. [77].
The previous examples indicated that the position of hybrids in phylogenetic trees is not predictable and depends on which parent the hybrid shares the most derived characters with [78]. We therefore used information not only from the trees but also from haplotype networks and the heterozygosity values to define the most probable hybrids in our sample. A total of eight accessions were confirmed or identified as hybrids and have been eliminated from downstream analyses. We considered this step as essential to provide an accurate view of the relationships between the pure species. However, we acknowledge the fact that possible hybrids within clades or complex interspecific hybrids may have escaped from our scrutiny at this stage.
Zecca et al [30] detected a very interesting gene flow between the M. rotundifolia outgroup and two species from clade NA1 V. candicans (syn. V. mustagensis) and V. coriacea (syn. V. shuttleworthii), which are both currently sympatric with M. rotundifolia. It had been also previously documented that this species hybridized with several American Vitis species and more readily when used as a male than a female parent [79]. The existence of introgression between the two subgenera in North America would bring the outgroup closer to American species in phylogenetic trees. This could have provided a false support for the American origin of subg. Vitis suggested in several studies [22, 23, 25].
Species relationships within clades
For V. vinifera (clade EU), the accessions included in our work only covered a little part of the species diversity. The cultivars of subsp. vinifera analyzed represented mostly the diversity group from Western Europe and have a wine usage [51, 80]. The studied wild accessions of subsp. sylvestris also originated from the same region except the Tunisian accession (LAS). In Western Europe, the differentiation of the two subspecies is probably limited due to the existence of a gene flow [81]. However, the two subspecies were clearly separated using the sequence data without the hybrids (Fig 3). The SNP array data without hybrids also separated the two subspecies, although the Tunisian accession clustered or not with the cultivated accessions depending on the set and number of SNPs used (Fig 5).
The East Asian clade is sister to the European clade. The sequence data allowed to cluster some Asian species: i) V. amurensis, V. coignetiae and V. thunbergii that constitute a northern clade, ii) V. flexuosa and V. balanseana representing the V. bryoniifolia clade recently studied by Ma et al [41] and iii) V. armata, V. romanetii and V. davidii included by Galet [12] in the Spinosae serie. Several discrepancies were observed between the sequence and array data. For instance, V. amurensis was found sister to other Asian taxa in the distance tree but not in the ML tree. Moreover, in the distance tree V. armata did not cluster with the two other species of the Spinosae serie, which was identified using sequence data. These differences could be due to the lack of Asian taxa within the discovery panel for the SNP array.
Our results for American species are largely congruent with those of Klein et al [24] who also described two distinct groups: Clade I and II corresponding to our clades NA2 and NA1, respectively. These authors discussed in details the characters of each of the clades in relation with previous classification, morphological characters and phylogeography. NA1 clade includes species from eastern regions like V. aestivalis, V. cinerea, V. cordifolia (syn. V. vulpina), V. labrusca, V. coriacea (syn. V. shuttleworthi), V. simpsonii and V. candicans (syn. V. mustangensis). Both sequence and array data identified a V. cinerea clade including the two varieties var. cinerea and var. berlandieri (syn. var. helleri). These varieties were clearly distinguished especially in the distance tree with a strong support (Fig 7D). Gene flow between them is probably limited due to the strict adaptation of var. berlandieri to dry climate and lime soils [82]. Some authors proposed to classify var. berlandieri and var. cinerea in two different species based on morphological differences [83]. However, a recent population analysis showed that the two varieties are in the species boundaries and that var. cinerea probably contains several genetic groups [84]. Interestingly, our present study identified V. cordifolia as sister to V. cinerea.
Further work is required to better understand the relationships among V. aestivalis, V. simpsonii, V. coriacea and V. candicans. These species share similar climatic niches in the southeastern United States [85] where some morphological similarities and possible hybridization caused taxonomic ambiguities [86]. For instance, our V. coriacea accession exhibited a high heterozygosity and shared many polymorphisms with V. simpsonii and V. aestivalis. Furthermore, the position of the three species in our phylogenetic trees changed according to the data set. Sequence data grouped V. aestivalis and V. simpsonii whereas the SNP array data grouped V. simpsonii and V. coriacea. V. simpsonii has sometimes been named V. cinerea var. floridana [87]. By contrast, accessions with this latter name did not group with other V. cinerea accessions in recent phylogenetic trees based on chloroplast [26] and nuclear [30] polymorphisms. We therefore suggest that V. cinerea var. floridana does not belong to V. cinerea; it is more probably a pure species or a hybrid close to V. aestivalis and V. coriacea.
NA2 clade includes the species: V. arizonica, V. riparia, V. longii (syn. V. acerifolia) and V. rupestris, thus confirming previous results [21, 22, 24]. In addition, we found a very close proximity between V. arizonica and V. treleasii. These two species have been assigned to the same series in previous classifications based on morphology [10, 12]. We identified V. monticola as a sister species in clade NA2 in congruence with previous molecular studies [21, 24]. We obtained conflicting results for V. rubra (syn. V. palmata). Sequence data identify this species as sister to the clade NA2 but SNP array indicated a close genetic relationship with V. candicans from clade NA1. This species could be a relict and its isolated position in the distance tree may be due to the sharing of ancestral polymorphisms with M. rotundifolia, which was represented in the discovery panel.
Our results also confirmed the particular phylogenetic position of V. californica, already pointed in several contributions [e.g. 17–19, 26], this species appearing very different from all other American species. In contrast to Ma et al [25], our nuclear data are congruent with plastome data [e.g. 18, 26]. V. californica was probably isolated for a long time in northwestern America and then entered in contact and hybridized with V. girdiana [30, 87] as well as, more recently, with V. vinifera [87]. V. californica shared characters with the core subg. Vitis from Eurasia: low to high susceptibility to mildew diseases and reddening leaves at fall whereas all other American species display low to high resistance to mildews and have yellowing leaves at fall.
Adaptation and morphological diversification in subgenus Vitis
Subgenus Vitis experienced rapid radiation in East Asia [25] and also a rapid and probably more recent radiation in North America [26]. These radiations have been associated with adaptation to diverse ecological conditions implying diversification for morphological characters. Similar adaptation has been observed in both continents, leading to two main growth habits: climbers with large leaves, tall plant stature with strong stems and tendrils (e.g. V. riparia, V. coignetiae) and shrubs with tiny leaves, small plant stature, many branches and weak climbing ability (e.g. V. rupestris, V. pentagona). A striking example of convergence was the assigment by Galet of V. labrusca and V. coignetiae to the Labruscae series, as both species have large and entire leaves but are phylogenetically very distinct (e.g. Fig 3). Recent studies [41, 88] also pointed out the importance of environment-driven evolutionary convergences driven in Vitis.
In this context, it is not surprising that our analyses of leaf measurements and OIV traits poorly reflected the species relationships found using DNA polymorphisms. As expected, the size and form of leaves did not allow recovering the differents clades although reniform leaves were only found in clade NA2 (V. rupestris). PCA analyses of OIV traits tend to isolate NA2 and also to separate NA1 from the EU and EA clades but with large variation within species. We then searched for a phylogenetic signal (i.e. “a tendency for closely related species to possess similar trait values due to descent from a common ancestor” [89, 90]) in our data using a method that took simultaneously into account phylogeny and morphological traits. Several traits were found autocorrelated with the phylogeny and the main result pointed again the NA2clade as the most distinct within the core Vitis subg. Vitis. Diversification in North America has been therefore very marked with three clades (V. califonica, NA1 and NA2) probably isolated during a long time. They adapted to different environments, which are very contrasted in Northern America [73, 91]. Assessing the relative roles of common ancestry and adaptation in determining morphological traits appears therefore very challenging at the subgenus level [92].
However, leaf measurements and OIV codes are still very useful for identification of species and varieties within species. Moreover, more precise methods for quantifying leaf shape are now available and have been recently applied to Vitis species. These methods performed very well for identification purposes even at the clonal level within V. riparia and V. rupestris [38]. The digital reconstruction of grapevine leaves allowed also recently identifying numerous QTL associated with leaf shape [93]. Taking into account the variation along the shoot has recently been found to outperform the classic measures of individual leaves [94].
Conclusions
In this work, we showed that a relatively low number of SNP markers obtained from sequences of a few nuclear genes are sufficient to recover the main divisions in the Vitis genus. Our phylogenetic results were largely similar to other recent phylogenies evidenced for this genus and confirmed the continental distribution of genetically distinct clades. In addition, using haplotype reconstruction of nuclear genes and a few chloroplast markers, we identified hybrids in our sample and demonstrated how they might affect the position of pure species in phylogenetic trees. Despite an evident effect of ascertainment bias, genotyping with an SNP array provided congruent results with the sequence data. There was recently a renewed interest for morphological variation in Vitis, not only for classification purposes but also to decipher evolution and adaptation in the genus knowing the genetic relationships between species [25, 29, 41, 80]. Here we contributed to this topic by searching a phylogenetic signal in morphological characters. The signal was generally weak but confirmed that the American clade NA2 is the most genetically divergent in subgenus Vitis exhibiting original morphological traits. Further research would aim to identify the ancestral species of clades and to follow their evolutionary history during their geographic extension [85]. For that purpose we suggest to combine different types of data: nuclear and plastid polymorphisms, including indels [42], morphological characters in particular for the leaf form and biochemical characters such as those involved in the flavonoid pathway [95].
Supporting information
S2 Fig. Boxplots of mean number of SNPs between haplotypes.
Vitis accessions were grouped in clades and colored as in Fig 1. Wild and cultivated of V. vinifera (EU) are grouped. Known hybrids (CHA, DOA1, LAB), colored in pink, showed high values as well as other accessions that were therefore considered as putative hybrids.
https://doi.org/10.1371/journal.pone.0283324.s002
(PDF)
S3 Fig. Haplotypes networks constructed for eleven loci.
a) GAI, b) CHI1, c) 4275A, d) TFL1, e) LDOX, f) DFR4, g) TC1A, h) TC1B, i) 255A, j)1526A and k) 4194A. Haplotypes are colored as in Fig 1.
https://doi.org/10.1371/journal.pone.0283324.s003
(PDF)
S4 Fig. Boxplots of mean heterozygosity calculated using the SNP array data.
Six hybrids having a V. vinifera parentage showed very high values. In the Eastern Asian clade, the high value of V. amurensis questioned its origin. Accessions were grouped in clades colored as in Fig 1. Wild and cultivated of V. vinifera (EU) are grouped.
https://doi.org/10.1371/journal.pone.0283324.s004
(PDF)
S5 Fig. Chlorotype sharing among Vitis accessions.
a) assignment in 19 chlorotypes (MLG = Multi Locus Genotype), b) assignment in eight chlorotypes when one marker difference was allowed. Accessions grouped in clades colored as in Fig 1.
https://doi.org/10.1371/journal.pone.0283324.s005
(PDF)
S6 Fig. PCA of leaf measurements (angle and ratio).
a) correlation plot for variables classified according to their type (angle and ratio) b) plot for Vitis accessions, grouped in clades colored as in Fig 1. Hybrids were excluded.
https://doi.org/10.1371/journal.pone.0283324.s006
(PDF)
S8 Fig. Phylogenetic signals in morphological variables.
a) leaf measurements, b) OIV codes.
https://doi.org/10.1371/journal.pone.0283324.s008
(PDF)
S9 Fig. Correlation plots for variables with significant phylogenetic signal.
a) leaf measurements, b) OIV codes.
https://doi.org/10.1371/journal.pone.0283324.s009
(PDF)
S2 Table. List of genes and primers for sequencing.
https://doi.org/10.1371/journal.pone.0283324.s011
(PDF)
S3 Table. Discovery panel and SNPs from the Illumina chip.
https://doi.org/10.1371/journal.pone.0283324.s012
(PDF)
S4 Table. Leaf traits measured and their phylogenetic signal.
https://doi.org/10.1371/journal.pone.0283324.s013
(PDF)
S6 Table. Phylogenetic signal in OIV descriptors.
https://doi.org/10.1371/journal.pone.0283324.s015
(PDF)
S1 Data. Archive of molecular and morphological data analyzed.
https://doi.org/10.1371/journal.pone.0283324.s016
(ZIP)
Acknowledgments
We thank the Montpellier Supagro grape collection and the INRAE Vassal-Montpellier Grapevine Biological Resources Center (INRAE-GBRC) for providing samples for DNA extraction and for leaf measurements using SuperAmpelo. We also thank INRAE-GCRB for providing OIV data for part of the accessions. We thank Cédric Moisy for the processing of leaf images using ImageJ. We also thank two anonymous reviewers for their useful suggestions and Philippe Chatelet for the English editing.
References
- 1. Moore MO. Classification and systematics of Eastern North American Vitis L. (Vitaceae) North of Mexico. Sida. 1991;14:339–367.
- 2.
Moore MO, Wen J. Vitaceae Jussieu. Grape family. In: Flora of North America Editorial Committee, editor. Flora of North America North of Mexico, Magnoliophyta: Vitaceae to Garryaceae. Volume 12. New York and Oxford; 2016; pp.3–22.
- 3. Zohary D, Spiegel-Roy P. Beginnings of fruit growing in the Old World. Science. 1975;187:319–327. pmid:17814259
- 4. Levadoux L, Boubals D, Rives M. Le genre Vitis et ses espèces. Ann Amélior Plantes. 1962;12:19–44.
- 5.
Mullins MG, Bouquet A, Williams LE. The grapevine and its relatives. In: Mullins MG, Bouquet A, Williams LE, editors. Biology of the grapevine. Cambridge: Cambridge University Press; 1992. pp.17–27.
- 6. Sawler J, Reisch B, Aradhya MK, Prins B, Zhong GY, Schwaninger H, et al. Genomics assisted ancestry deconvolution in grape. PLoS One. 2013;8(11):e80791. pmid:24244717
- 7. Migicovsky Z, Sawler J, Money D, Eibach R, Miller AJ, Luby JJ, et al. Genomic ancestry estimation quantifies use of wild species in grape breeding. BMC Genomics. 2016;17:478. pmid:27357509
- 8. Wan Y, Schwaninger H, Li D, Simon CJ, Wang Y, Zhang C. A review of taxonomic research on Chinese wild grapes. Vitis. 2008;47:81–88.
- 9.
Planchon J-E. Ampelideae. DC Prodromus systematis Naturalis, Monographiae Phanerogamarum. Paris: Masson. 1887; pp. 305–654.
- 10.
Munson TV. Foundations of American grape culture. Denison: Munson and Son; 1909. or 10. Munson TV. Foundations of American grape culture. New York: Orange Judd Company; 1909.
- 11. Bailey LH. The species of grape peculiar to North America. Gentes Herberum 1934; 3 151–244.
- 12.
Galet P (1988) Les vitacées. In: Cépages et vignobles de France. Tome 1. Les vignes américaines. Montpellier: Pierre Galet; 1988.
- 13. Fengqin Z, Fangmei L, Dabin G. Studies on germplasm resources of wild grape species (Vitis spp.) in China. Proceedings of the 5th International Symposium on Grape Breeding, 12–16 September 1989, St. Martin/Pfalz, FRG. Vitis (Special Issue) 1990; 50–57.
- 14.
Chen ZD, Ren H & Wen J (2007) Vitaceae. In: Wu CY, Hong D-Y, Raven PH, editors. Flora of China, vol. 12. Beijing: and St. Louis: Science Press and Missouri Botanical Garden Press. 2007; 12, 173–222. Available from: http://flora.huh.harvard.edu/china/mss/volume12/Vitaceae.pdf]
- 15. Barrett HC, Carmer SG, Rhodes AM. A taximetric study of interspecific variation in Vitis. Vitis. 1969;8:177–187.
- 16. Comeaux BL, Nesbitt WB, Fantz PR. Taxonomy of the native grapes of North Carolina. Castanea. 1987;52:197–215.
- 17. Trondle D, Schroder S, Kassemeyer HH, Kiefer C, Koch MA, Nick P. Molecular phylogeny of the genus Vitis (Vitaceae) based on plastid markers. Am J Bot. 2010;97(7):1168–1178. pmid:21616868
- 18. Péros J-P, Berger G, Portemont A, Boursiquot J-M, Lacombe T. Genetic variation and biogeography of the disjunct Vitis subg. Vitis (Vitaceae). J Biogeogr. 2011;38:471–486.
- 19. Zecca G, Abbott JR, Sun WB, Spada A, Sala F, Grassi F. The timing and the mode of evolution of wild grapes (Vitis). Mol Phylogenet Evol. 2012;62(2):736–747.
- 20. Aradhya M, Wang Y, Walker MA, Prins BH, Koehmstedt AM, Velasco D, et al. Genetic diversity, structure, and patterns of diffrentiation in the genus Vitis. Pl Syst Evol. 2013;299:317–30.
- 21. Miller AJ, Matasci N, Schwaninger H, Aradhya MK, Prins B, Zhong GY, et al. Vitis phylogenomics: hybridization intensities from a SNP array outperform genotype calls. PLoS One. 2013;8(11):e78680.
- 22. Wan Y, Schwaninger HR, Baldo AM, Labate JA, Zhong GY, Simon CJ. A phylogenetic analysis of the grape genus (Vitis L.) reveals broad reticulation and concurrent diversification during neogene and quaternary climate change. BMC Evol Biol. 2013;13:141. pmid:23826735
- 23. Liu XQ, Ickert-Bond SM, Nie ZL, Zhou Z, Chen LQ, Wen J. Phylogeny of the Ampelocissus-Vitis clade in Vitaceae supports the New World origin of the grape genus. Mol Phylogenet Evol. 2016;95:217–228.
- 24. Klein LL, Miller AJ, Ciotir C, Hyma K, Uribe-Convers S, Londo J. High-throughout sequencing data clarify evolutionary relationships among North American Vitis species and improve identification in USDA Vitis germplasm collections. Amer J Bot. 2018;105(2):1–12.
- 25. Ma Z-Y, Wen J, Icker-Bond S, Nie Z-L, XChen L-Q, Liu X-Q. Phylogenomics, biogeography, and adaptative radiation of grapes. Mol Phylogenet Evol. 2018;129:258–267.
- 26. Wen J, Harris AJ, Kalburgi Y, Zhang N, Xu Y, Zheng W et al. (2018) Chloroplast phylogenomics of the New World grape species (Vitis, Vitaceae). Journal of Systematics and Evolution 56: 297–308.
- 27. Liang Z, Duan S, Sheng J, Zhu S, Ni X, Shao J, et al. Whole-genome resequencing of 472 Vitis accessions for grapevine diversity and demographic history analyses. Nature communications. 2019;10(1):1190. pmid:30867414
- 28. Zecca G, Grassi F, Tabidze V, Pipia I, Kotorashvili A, Kotaria N, et al. Dates and rates in grape’s plastomes: evolution in slow motion. Curr Genet. 2020;66:123–140. pmid:31201544
- 29. Ma Z-Y, Wen J, Tian J-P, Jamal A, Chen L-Q, Liu X-Q. Testing reticulate evolution of four Vitis species from East Asia using restriction-site associated DNA sequencing. J Syst Evol. 2018;56:331–339.
- 30. Zecca G, Labra M, Grassi F. (2020). Untangling the evolution of American wild grapes: admixed species and how to find them. Frontiers in Plant Science. 10. 1814. pmid:32117355
- 31.
Galet P. Précis d’ampélographie pratique. 7th edition. Montpellier: Galet P; 1998.
- 32. Boursiquot J-M, Vignau L, Boulet JC. Researches on ampelometry utilization. Riv Vitic Enol. 1989;1:37–52.
- 33. Martinez M-C, Grenan S. A graphic reconstruction method of an average vine leaf. Agronomie. 1999;19:491–507.
- 34. Preiner D, Safner T, Karolan Kontic J, Markovic Z, Simon S, Maletic E. Analysis of phyllometric parameters efficiency in discrimination of Croation native V. vinifera cultivars. Vitis. 2014;53:215–217.
- 35. Zinelabidine LH, Laiadi Z, Benmehaia R, Gago P, Boso S, Santiago JL, et al. Comparative ampelometric and genetic analysis of grapevine cultivars from Algeria and Morocco. Aust J Grape Wine Res. 2014;20:324–333.
- 36. Alba V, Bergamini C, Genghi R, Gasparro M, Prniola R, Antonacci D. Ampelometric leaf trait and SSR loci selection for a multivariate statistical approach in Vitis vinifera L. biodiversity management. Mol Biotechnol. 2015;57:709–719.
- 37. Chitwood DH, Klein LL, O’Hanlon R, Chacko S, Greg M, Kitchen C, et al. Latent developmental and evolutionary shapes embedded within the grapevine leaf. New Phytol. 2016;210(1):343–355. pmid:26580864
- 38. Klein LL, Caito M, Chapnick C, Kitchen C, O’Hanlon R, Chitwood DH, et al. Digital morphometrics of two North American grapevines (Vitis: Vitaceae) quantifies leaf variation between species, within species, and among individuals. Front Plant Sci. 2017;8:373.
- 39. Min Z, LI R, Zhao X, Li R, Zhang Y, Liu MH, et al. Morphological variability in leaves of Chinese wild Vitis species. Sci Hort. 2018;238:138–146.
- 40. Chitwood DH, Ranjan A, Martinez CC, Headland LR, Thiem T, Kumar R, et al. A modern ampelography: a genetic basis for leaf shape and venation patterning in grape. Plant Physiol. 2014;164(1):259–272. pmid:24285849
- 41. Ma Z-Y, Wen J, Tian J-P, Gui LL, Liu X-Q. Testing morphological trait evolution and assessing species delimitations in the grape genus using a phylogenomic framework. Mol Phylogenet Evol. 2020;148: 106809. pmid:32224125
- 42. Péros J-P, Launay A, Berger G, Lacombe T, This P. MybA1 gene diversity across the Vitis genus. Genetica. 2015;143:373–384.
- 43.
Rosen S, Skaletsky H (2000) Primer3 on the www for general users and for biologist programmers. In: Krawetz S, Misener S, editors. Bioinformatics methods and protocols. Totowa: Humana Press. pp. 365–386.
- 44. Boss PK, Thomas MR. Association of dwarfism and floral induction with a grape ’green revolutioné mutation. Nature. 2002;416:847–850. pmid:11976683
- 45. Fernandez L, Torregrosa L, Segura V, Bouquet A, Martinez-Zapater JM. Transposon-induced gene activation as a mechanism generating cluster shape somatic variation in grapevine. Plant J. 2010;61(4):545–557. pmid:19947977
- 46. 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.
- 47. Bogs J, Ebadi A, McDavid D, Robinson SP. Identification of the flavonoid hydroxylases from grapevine and their regulation during fruit development. Plant Physiol. 2006;140:279–291. pmid:16377741
- 48. Le Paslier M-C, Choisne N, Scalabrin S, Bacilieri R, Berard A, Bounon R et al. The GrapeReSeq 18k Vitis genotyping chip. IX International Symposium on Grapevine Physiology & Biotechnology, La Serena, Chili. 2013.
- 49. Shen R, Fan JB, Campbell D, Chang W, Chen J, Doucet D, et al. High-throughput SNP genotyping on universal bead arrays. Mutat Res. 2005;573(1–2):70–82. pmid:15829238
- 50. Jaillon O, Aury JM, Noel B, Policriti A, Clepet C, Casagrande A, et al. The grapevine genome sequence suggests ancestral hexaploidization in major angiosperm phyla. Nature. 2007;449:463–467. pmid:17721507
- 51. Laucou V, Launay A, Bacilieri R, Lacombe T, Adam-Blondon AF, Bérard A, et al. Extended diversity analysis of cultivated grapevine Vitis vinifera with 10K genome-wide SNPs. PLoS One. 2018;13(2): e0192540.
- 52. Kumar S, Stecher G, Tamura K (2016). MEGA7: Molecular evolutionary genetics Analysis version 7.0 for bigger datasets. Mol Biol Evol. 2016;33:1870–1874.
- 53.
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2019. https://www.R-project.org/.
- 54. Jombart T, Archer F, Schliep K, Kamvar Z, Harris R, Paradis E, et al. apex: phylogenetics with multiple genes. Mol Ecol Resour. 2017;17:19–26. pmid:27417145
- 55. Schliep K. phangorn: phylogenetic analysis in R. Bioinformatics. 2011;27:592–593. pmid:21169378
- 56. Paradis E, Schliep K. ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics. 2019;35:526–528. pmid:30016406
- 57. Jombart T. adegenet: a R package for the multivariate analysis of genetic markers. Bioinformatics. 2008;24:1403–1405. pmid:18397895
- 58. Davidovich O, Kimmel G, Shamir R. GEVALT: an integrated software tool for genotype analysis. BMC Bioinform. 2007;8:36. pmid:17270038
- 59. Dereeper A, Nicolas S, Le Cunff L, Bacilieri R, Doligez A, Peros JP, et al. SNiPlay: a web-based tool for detection, management and analysis of SNPs. Application to grapevine diversity projects. BMC Bioinform. 2011;12:134. pmid:21545712
- 60. Leigh JW, Bryant D. POPART: full-feature software for haplotype network construction. Methods Ecol Evol. 2015;6:1110–6.
- 61. Soldavini C, Stefanini M, Dallaserra M, Policarpo M, Schneider A. SuperAmpelo, a software for ampelometric and ampelometric descriptions in Vitis. Acta Hortic. 2009;827:253–257.
- 62. Abramoff MD, Magalhaes PJ, Ram SJ. Image processing with ImageJ. Biophotonics Int. 2004;11:36–42.
- 63.
Organisation Internationale de la Vigne et du Vin. OIV descriptor list for grape varieties and Vitis species. 2nd Edition. 2009: OIV Paris. Available from: http://oiv/Standards%20and%20technical%20documents/Description%20of%20grape%20varieties
- 64. Le S, Josse J, Husson F. FactoMineR: an R package for multivariate analysis. J Stat Soft. 2008;25:1–18.
- 65. Kassambara A, Mundt F. factoextra: extract and visualize the results of Multivariate Data Analyses. R package version 1.0.5. 2017. Available from: https://CRAN.R-project.org/package=factoextra
- 66. Jombart T, Balloux F, Dray S. adephylo: new tools for investigating the phylogenetic signal in biological traits. Bioinformatics. 2010;26(15):1907–1909. pmid:20525823
- 67. Abouheif E. A method for testing the assumption of phylogenetic independence in comparative data. Evol Ecol Res. 1999;1:895–909.
- 68. Garza JC, Freimer NB. Homoplasy for size at microsatellite loci in humans and chimpanzees. Genet Res. 1996;6:211–217. pmid:8963898
- 69. Lachance J, Tishkoff SA. SNP ascertainment bias in population genetic analyses: why it is important, and how to correct it. BioEssays. 2013;35(9):780–786. pmid:23836388
- 70. Albrechten A, Nielsen FC, Nielsen R. Ascertainment biases in SNP chips affect measures of population divergence. Mol Biol Evol. 2010;27:2634–2547.
- 71. Arnold B, Corbett-Detig RB, Hartl D, Bomblies K. RADseq underestimates diversity and introduces genealogical biases due to nonrandom haplotype sampling. Mol Ecol. 2013;22:3179–3190. pmid:23551379
- 72. Myles S, Boyko AR, Owens CL, Brown PJ, Grassi F, Aradhya MK, et al. Genetic structure and domestication history of the grape. Proc Natl Acad Sci U S A. 2011;108(9):3530–3535. pmid:21245334
- 73.
Viala P. Une mission viticole en Amérique. Montpellier: C. Coulet; 1889.
- 74. Foex G. Les hybrides de V. vinifera par V. rupestris et le Vitis champinii. La Vigne Américaine. 1882;6(1):19–22.
- 75. Myles S, Chia J-M, Hurwitz B, Simon CJ, Zhong GY, Buckler E, et al. Rapid genomic characterization of the genus Vitis. PLoS One. 2010;5(1):e8219.
- 76. Wen J, Herron SA, Yang X, Liu B-B, Zuo Y-J, Harris AJ, et al. Nuclear and chloroplast sequences resolve the enigmatic origin of the Concord grape. Front Plant Sci. 2020;11:263. pmid:32256506
- 77. Huber F, Rockel F, Schwander F, Mail E, Eibach R, Cousins P, et al. A view into American grapevine history: Vitis vinifera cv. ’Sémillon’ is an ancestor of ’Catawba’ and ‘Concord’. Vitis. 2016;55:53–56.
- 78. Funk VA. Phylogenetic patterns and hybridization. Ann Missouri Bot Gard. 1985;72:681–715.
- 79. Detjen LR. The limits in hybridization of Vitis rotundifolia with related species and genera. Tech Bull n°17. North Carolina Agric Exp Stat; 1919, 26 pp.
- 80. Bacilieri R, Lacombe T, Le Cunff L, Di Vecchi-Staraz M, Laucou V, Genna B, et al. Genetic structure in cultivated grapevines is linked to geography and human selection. BMC Plant Biol. 2013;13:25. pmid:23394135
- 81. Riaz S, De Lorenzis G, Velasco D, Koehmstedt A, Maghradze D, Bobokashvili Z, et al. Genetic diversity analysis of cultivated and wild grapevine (Vitis vinifera L.) accessions around the Mediterranean basin and Central Asia. BMC Plant Biol. 2018;18(1):137.
- 82. Barrett HC. Vitis cinerea as a source of desirable characters in grape breeding. Proc Amer Soc Hort Sci. 1957;70:165–168.
- 83. Gandhi KN, Brown LE. A nomenclature note on Vitis cinerea and V. berlandieri (Vitaceae). SIDA. 1989;13(4):506–508.
- 84. Péros JP, Cousins P, Launay A, Cubry P, Walker A, Prado E, et al. Genetic diversity and population structure in Vitis species illustrate phylogeographic patterns in eastern North America. Mol Ecol. 2021;30(10):2333–48.
- 85. Callen ST, Klein LL, Miller AJ. Climatic niche characterization of 13 North American Vitis species. Am J Enol Vitic. 2016;67(3):339–49.
- 86. Comeaux BL, Fantz PR. Nomenclatural clarification of the name Vitis simpsonii Munson (Vitaceae). SIDA. 1987;12:279–86.
- 87. Dangl GS, Mendum ML, Yang J, Walker MA, Preece JE. Hybridization of cultivated Vitis vinifera with wild V. californica and V. girdiana in California. Ecol Evol. 2015;5(23):5671–84.
- 88. Ickert-Bond SM, Harris AJ, Lutz JR, Wen J. A detailed study of leaf micromorphology and anatomy of New World Vitis L. subgenus Vitis within a phylogenetic and ecological framework reveals evolutionary convergence. J Syst Evol. 2018;56:309–330.
- 89. Little SA, Kembel SW, Wilf P. Paleotemperature proxies from leaf fossils reinterpreted in light of evolutionary history. Plos One. 2010; 5(12):e15161. pmid:21203554
- 90. Givnish TJ. Comparative studies of leaf form: assessing the relative roles of selective pressures and phylogenetic constraints. New Phytol. 1987; 106 (Suppl.):131–160.
- 91.
Heinitz CC, Uretsky J, Dodson Peterson JC, Huerta-Acosta KG, Walker MA. Crop wild relatives of grape (Vitis vinifera L.) throughout North America. In: Greene S, Williams K, Khoury C, Kantat MB, Marek L, editors. North American Crop Wild Relatives. 2: Springer International Publishing; 2019. p. 329–51.
- 92. Chitwood DH, Sinha NR. Evolutionary and environmental forces sculpting leaf development. Curr Biol. 2016; 26: R297–R306. pmid:27046820
- 93. Demmings EM, Williams BR, Lee C-R, Barba P, Yang S, Hwang C-F et al. Quantitative trait locus analysis of leaf morphology indicates conserved shape loci in grapevine. Front Plant Sci. 2019; 10:1373. pmid:31803199
- 94. Bryson AE, Wilson Brown M, Mullins J, Dong W, Bahmani K, Bornowski N et al. Composite modeling of leaf shape along shoots discriminates Vitis species better than individual leaves. Appl Plant Sci. 2020; 8(12):e11404.
- 95. Moore MO, Giannasi DE. Foliar flavonoids of eastern North American Vitis (Vitaceae) north of Mexico. Plant Syst Evol. 1994;193:21–36.