23 Oct 2009: Meerow AW, Noblick L, Borrone JW, Couvreur TLP, Mauro-Herrera M, et al. (2009) Correction: Phylogenetic Analysis of Seven WRKY Genes across the Palm Subtribe Attaleinae (Arecaceae) Identifies Syagrus as Sister Group of the Coconut. PLOS ONE 4(10): 10.1371/annotation/c59c150a-0412-490a-8e7d-ae5622fd4434. https://doi.org/10.1371/annotation/c59c150a-0412-490a-8e7d-ae5622fd4434 View correction
The Cocoseae is one of 13 tribes of Arecaceae subfam. Arecoideae, and contains a number of palms with significant economic importance, including the monotypic and pantropical Cocos nucifera L., the coconut, the origins of which have been one of the “abominable mysteries” of palm systematics for decades. Previous studies with predominantly plastid genes weakly supported American ancestry for the coconut but ambiguous sister relationships. In this paper, we use multiple single copy nuclear loci to address the phylogeny of the Cocoseae subtribe Attaleinae, and resolve the closest extant relative of the coconut.
We present the results of combined analysis of DNA sequences of seven WRKY transcription factor loci across 72 samples of Arecaceae tribe Cocoseae subtribe Attaleinae, representing all genera classified within the subtribe, and three outgroup taxa with maximum parsimony, maximum likelihood, and Bayesian approaches, producing highly congruent and well-resolved trees that robustly identify the genus Syagrus as sister to Cocos and resolve novel and well-supported relationships among the other genera of the Attaleinae. We also address incongruence among the gene trees with gene tree reconciliation analysis, and assign estimated ages to the nodes of our tree.
This study represents the as yet most extensive phylogenetic analyses of Cocoseae subtribe Attaleinae. We present a well-resolved and supported phylogeny of the subtribe that robustly indicates a sister relationship between Cocos and Syagrus. This is not only of biogeographic interest, but will also open fruitful avenues of inquiry regarding evolution of functional genes useful for crop improvement. Establishment of two major clades of American Attaleinae occurred in the Oligocene (ca. 37 MYBP) in Eastern Brazil. The divergence of Cocos from Syagrus is estimated at 35 MYBP. The biogeographic and morphological congruence that we see for clades resolved in the Attaleinae suggests that WRKY loci are informative markers for investigating the phylogenetic relationships of the palm family.
Citation: Meerow AW, Noblick L, Borrone JW, Couvreur TLP, Mauro-Herrera M, Hahn WJ, et al. (2009) Phylogenetic Analysis of Seven WRKY Genes across the Palm Subtribe Attaleinae (Arecaceae) Identifies Syagrus as Sister Group of the Coconut. PLoS ONE 4(10): e7353. https://doi.org/10.1371/journal.pone.0007353
Editor: Simon Joly, McGill University, Canada
Received: May 23, 2009; Accepted: September 17, 2009; Published: October 6, 2009
This is an open-access article distributed under the terms of the Creative Commons Public Domain declaration which stipulates that, once placed in the public domain, this work may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose.
Funding: This study was entirely funded by the corresponding author's USDA-ARS CRIS Project's discretionary budget. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Cocos nucifera L., the coconut, is a charismatic monotypic genus forming a dominant part of littoral vegetation across the tropics. Besides its paradisiacal connotation, the coconut plays a vital role at many different economic levels . Cocos nucifera is pantropically distributed, a present day range significantly influenced both by a seed well-adapted to oceanic dispersal and the species' importance to humans –. Because of this wide geographic range, the biogeographic origins of the coconut have been one of the “abominable mysteries” of palm systematics for decades . A neotropical origin of Cocos was first proposed by de Candolle . Beccari  suggested an origin in Asia or the South Pacific, while Moore  proposed Melanesia. Harries  argued for a western Pacific origin, and later  opined for an origin in the Malesia biogeographic province (the Malay Peninsula, Indonesia, the Philippines and New Guinea), an opinion supposedly shared by a majority of coconut specialists , but weakly supported with data. Moreover, in the recent literature on the subject by coconut geneticists , , there does not appear to be a clear distinction between the deeper phylogenetic history of the genus and its far more recent domestication.
Cocos nucifera belongs to the monophyletic Cocoseae , –, one of thirteen tribes of Arecaceae subfam. Arecoideae . In addition to the coconut, this tribe also contains a number of other palms with significant economic importance, e.g., Elaeis guineensis Jacq. (African oil palm), Bactris gasipaes Kunth (peach palm), and many other species of value in local economies . The tribe was first recognized informally by Moore  as “the cocosoid palms,” denoted by fruits with bony endocarps bearing three germination pores or “eyes.” He further delimited three subgroups as the Bactris Jacq., Cocos L., and Elaeis Jacq. alliances. Uhl and Dransfield  formalized the groups as tribe Cocoeae containing 22 genera classified within five subtribes, later reduced to 20 genera in three subtribes (more or less embracing Moore's  alliances), with orthographic correction of the name to Cocoseae . In addition to the distinctive endocarp, the tribe is well-marked by its once-branched inflorescence, inconspicuous prophyll, conspicuous and often woody peduncular bract, imbricate petals of female flowers, and a triovulate gynoecium . Cocoseae now encompasses 18–19 genera of predominantly Neotropical distribution . One genus, Elaeis Jacq., has both a species endemic to tropical America and another in Africa.
Within the Cocoseae, the coconut is part of the moderate-sized subtribe Attaleinae , containing 11–12 genera. With the exception of Cocos, two endemics in Madagascar (Beccariophoenix Jum. & H. Perrier, two spp.; Voanioala J. Dransf., monotypic), and Jubaeopsis Becc. (monotypic, found in a restricted part of South Africa), the majority of the genera are Neotropical endemics .
Despite the importance of the Cocoseae, a well-resolved phylogeny for this tribe has been elusive , –, . Thus, to date, molecular systematics using plastid or nuclear markers have failed to unambiguously identify the sister genus of the coconut, determination of which is not only of biogeographic interest, but will also open fruitful avenues of inquiry regarding evolution of functional genes useful for crop improvement.
The utility of WRKY loci for determining infraspecific relationships has been demonstrated by genetic mapping in Theobroma cacao L.  and by differentiating individuals from one another within T. cacao  and C. nucifera germplasm collections , . Borrone et al.  demonstrated the utility of WRKY genes for phylogenetic inference in Malvaceae. Further information on WRKY loci, including details of their evolution and orthology, can be found in the Discussion section of this paper.
Our study focuses on reconstructing the phylogenetic relationships within the subtribe Attaleinae, and represents the most intensive sampling of the group so far in a molecular analysis. We use sequences of seven putatively independent, single copy WRKY loci originally isolated from Cocos nucifera in order to resolve the closest extant relative of the coconut, the evolutionary relationships of the other genera in the subtribe, determine how paleohistorical events shaped the evolution and biogeography of the Attaleinae, and demonstrate the utility of WRKY loci for phylogenetic inference within the Arecaceae.
Individual gene tree analyses
Excluding gaps, the number of phylogenetically informative characters (Table 1) ranged from 66 (WRKY12) to 271 (WRKY19). Maximum parsimony (MP) strict and maximum likelihood (ML) bootstrap consensus trees are available as supplemental Figures S1–4. Consistency indices (CI) were above 0.75 for all seven gene matrices investigated, while retention indices (RI) were always >0.89 (Table 1). Adding a coded indel matrix contributed little additional resolution to the trees, though bootstrap support (BP) was slightly increased for some clades (not shown). For each gene matrix, ML produced a tree identical in topology to one of the trees found by MP.
Phylogenetic incongruence among the loci.
Examination of the individual consensus trees and partitioned Bremer indices (Tables 1–2, Figs. S1, 2) indicates that incongruence is more a result of insufficient resolution at various nodes within each locus or “soft” incongruence, rather than conflicting resolution, or “hard” incongruence , though some degree of conflict at deeper nodes in the trees is evident (Figs. S1–4). P values from the incongruence length difference (ILD) tests ,  indicated that the null hypothesis of congruence could be accepted for WRKY2 and 12, and 12 and 21 (p = 0.44 and 0.68, respectively). The latter are in different major clades of the WRKY family phylogeny (1 and 2b, respectively). P values for all other pairwise combinations were never lower than 0.01. The accuracy of the ILD as an arbiter of combinability has declined steadily since Farris et al. ,  first recommended a P-value of 0.05 as the threshold for determining non-combinability. Numerous studies have concluded that P-values<0.05, and even as low as 0.001, should not preclude data set combination –. Based on these results, and the fact that many of the same monophyletic groups were resolved with each locus (Tables 1–2), we combined all seven loci for phylogenetic analysis.
Super-matrix (concatenated) analysis.
The combined sequence matrix consisted of 5831 total characters of which 974 were parsimony informative (17%). Heuristic search with parsimony found 446 equally most parsimonious trees of length = 2614, CI = 0.73 and RI = 0.87. Combining all seven WRKY loci yielded the most fully resolved trees and the highest bootstrap support (Fig. 1), both with MP and ML (the second, under-lined BP value between parentheses in the following discussion is that for ML). Partitioned Bremer (decay) indices (DI) (Table 2) indicates the relative contribution of each locus.
Numbers above branches are bootstrap support percentages. Numbers below branches are ML bootstrap support (underlined) and non-partitioned decay indices (italic). The numbers at each node refer to Table 2, which see for partitioned decay indices. Letter designations in red are area distributions of terminal taxa: A = South Africa, B = Madagascar, C = Amazonas (north, east, south), D = Chile, E = Eastern Brazil, F = Central Brazil, G = Andes, H = Central America, I = Mexico, J = West Indies, K = Southern Brazil, L = Northern South America, i.e., Caribbean coastal Venezuela and Colombia, French Guiana, Guyana, Surinam, M = Argentina-Paraguay-Uruguay, N = Colombia-Venezuela (llanos region), and O = western Amazonas.
The ML (Fig. S4) tree was essentially identical to the parsimony consensus (Fig. 1), although the terminals were more fully resolved. In general, BP was higher with ML (Table 3, Fig. S4), though generally with weak support wherever a polytomy appears in the parsimony strict consensus (Fig. 1). Bayesian analysis of the combined data matrix (Fig. 2) was also congruent with parsimony and ML, except for the lack of resolution for Bactris (Bactridinae). Most clades had posterior probability (PP) scores = 100%. The only clades with <90% were at several terminal nodes in Syagrus.
Numbers above branches are posterior probability scores, i.e., the proportion of tree within which that clade was resolved. Letter in red at nodes are ancestral area optimizations as determined by dispersal-vicariance analysis: A = South Africa, B = Madagascar, C = Amazonas (north, east, south), D = Chile, E = Eastern Brazil, F = Central Brazil, G = Andes, H = Central America, I = Mexico, J = West Indies, K = Southern Brazil, L = Northern South America, i.e., Caribbean coastal Venezuela and Colombia, French Guiana, Guyana, Surinam, M = Argentina-Paraguay-Uruguay, N = Colombia-Venezuela (llanos region), and O = western Amazonas. Ambiguous area optimizations at a node are separated by commas.
The monophyly of all genera of the Attaleinae is supported by the combined WRKY data matrix (Fig. 1), with the exception of Syagrus, which is paraphyletic with Lytocaryum. The very well-supported subtribe (100/100% BP; support for the Attaleinae crown node was validated by running successive bootstrap analyses with just Elaeis oleifera and then the Bactris spp. as the designated outgroup) consists of three clades. The first represents the African genera (91/95% BP), within which Becarriophoenix is sister to a Jubaeopsis/Voanioala clade (84/84% BP). Partitioned Bremer support for this clade (Table 2) is positive (6 loci) or neutral (1 locus) for all seven loci, and universally positive for its sister relationship to the rest of the Attaleinae. The African clade is sister to the American clade, which resolves as two monophyletic groups. The better supported of the two (95/90% BP) consists of Cocos strongly supported as sister to Syagrus inclusive of Lytocaryum . Only WRKY2 is incongruent with this resolution (Table 2). Four of seven loci resolved Lytocaryum as nested within Syagrus (Figs. S1–4). The second, less well-supported (72/96% BP), clade of American genera unites Butia and Jubaea in a clade (100/100% BP) that is sister to a fairly well-supported clade (88/96% BP) of Allagoptera (100/100% BP), Parajubaea (100/100% BP) and Polyandrococos. Allagopetra and Polyandrococos form a clade with moderate to high support (84/97%). A monophyletic Attalea (100/100% BP) is sister to the rest of this clade. Within Attalea, monophyletic subg. Attalea (100/100% BP) and Scheelea (75/81% BP) subclades are resolved, while the Orbignya group appears paraphyletic to Scheelea (Fig. 1). Substructure in the Syagrus clade includes a “rain forest” group (86/98% BP), uniting species from Amazonas, the Andean foothills, the Caribbean and the Atlantic rain forest of Brazil (Fig. 1), to which Lytocaryum has a weaker sister relationship. An “Eastern Brazilian” clade is resolved as sister to the rain forest group. Finally, the outlying clade in the genus (76/84% BP) unites three clustering (multiple stems) species with the solitary-stemmed S. macrocarpa.
Reconciling gene and species trees.
Reconciling individual gene trees to the species trees (generated from the combined analysis) necessitated significant costs in terms of deep coalescence events (Table 4). Locus WRKY21 had the lowest cost, followed by WRKY2 (these loci also were the most highly congruent based on the ILD). The greatest number of deep coalescence events, 82, occurred with WRKY12.
Twenty-five heuristic searches in GeneTree ,  found a total of 190 species trees. The strict consensus of these trees (Fig. S5, nodes collapsed to the generic level) resolves all genera as monophyletic, except for a paraphyletic Syagrus in which Lytocaryum is embedded (some of the gene trees used as input resolved Lytocaryum as sister to Syagrus). The American and African clades are resolved. In the African clade, Jubaeopsis and Voanioala are sisters, with Beccariophoenix subtending, as with MP, ML and BA (Figs. 1, 2). In the American clade, Jubaea is sister to Butia, Polyandrococos and Allagoptera also form a clade, but there is no further internal resolution. The GeneTree results must be approached with caution, as the five trees used for each locus represent only a small fraction of the total number of fully-resolved trees found for each.
Divergence time estimates.
The maximum-clade-credibility tree (Fig. 3) produced by BEAST  was identical in the topology of the Attaleinae to the maximum likelihood tree (Fig. S4) and the majority rule consensus of the 12,500 trees sampled from the mixed model (partitioned) MrBayes analysis (Fig. 2), with the exception of a few taxa that terminate zero length branches, and some lower PP at some terminal clades. Table 5 provides the age estimates of the important nodes, and mean dates are mapped on a chronogram derived from the maximum-clade-credibility tree (Fig. 3). The most recent common ancestor (MCRA) of the African and American clades is estimated at 43.7 MYBP, with a 95% highest posterior density (HPD) range of 27–50 (Table 5). Crown age for the American Attaleinae is 38.4 MYBP (23.9–44.4 95% HPD), while that of the modern African genera is 28.6 MYPB (10.4–32.8 95% HPD). Crown clade ages for the two main clades of American Attaleinae are ca. 33 and 35 MYBP, respectively, the latter node being the MCRA of Syagrus and Cocos nucifera. Syagrus (including Lytocaryum) would appear to be the oldest genus in the American Attaleinae, with a crown age of 27 MYBP (15.4–30.5 95% HPD), while the equally speciose Attalea is relatively young at 13 MYBP (8.6–20.5 95% HPD). The three main clades of Syagrus s. s. share crown node ages of ca. 16.6–17.5 MYBP.
Numbers are mean node ages; blue bars at nodes represent range of 95% HPD for the age estimate. Triangles indicate collapsed generic, sub-generic or species nodes. * = calibration point.
Dispersal-vicariance optimization places Madagascar (B) and Eastern Brazil (E) as the ancestral distribution for the Attaleinae, with subsequent vicariance between the two hemispheres (Fig. 2), and one dispersal to South Africa (Jubaeopsis). Both major clades of the American Attaleinae remained restricted to eastern Brazil until an Eocene/early Oligocene range extension to coastal Chile or Southern Brazil by the MCRA of Allagoptera, Butia, Jubaea, Polyandrococos and Parajubaea, and Miocene dispersals to Amazonas and the West Indies (Attalea). Range extensions of Syagrus, first into contiguous Southern Brazil and Argentina-Paraguay, did not occur until the mid-Miocene (Fig. 2, 3), with even later dispersal of the genus into Amazonas, Northern South America, and the West Indies. Despite its subsequent pantropical distribution, the crown node of Cocos nucifera is unambiguously positioned in eastern Brazil.
As the results of our molecular analysis were examined, it became clear that a survey of leaf anatomical characters within Attaleinae in progress by one of us (L. Noblick) lent support to one of the clades resolved by the WRKY genes. Allagoptera, Parajubaea and Polyandrococos all have have nonvascular bundles attached to their hypodermis layers (Fig. 4), but only Parajubaea and Allagoptera have very distinctive, “girder”-like narrow vascular bundles that span the distance between the upper and lower surface of the leaflets Fig. 4A–F). These are absent in Polyandrococos (Fig. 4G). However, both Parajubaea and Polyandrococos have an irregular, undulating abaxial surface often with hairy depressions, which is less visible in some species of Allagoptera. (Fig. 4A, B)
A. Allagopetra arenaria. B. A. brevicalyx. C. A. campestris. D. A. leucocalyx. E. Parajubaea cocoides. F. P. torallyi. G. Polyandrococos caudescens. Black arrows with white outlines (A–F) point to “girder”-like vascular bundles in Allagoptera and Parajubaea. Curved arrows indicate depressions in abaxial surface (E–G) characteristic of Parajubaea and Polyandrococos. Scale bars = ca. 0.5 mm.
WRKY transcription factors are predominantly plant-specific proteins, broadly distributed across the genome . A single WRKY locus is found in the ancient eukaryote Giardia lamblia Kofoid & Christiansen and the mycetozoan Dictylostelium discoideum Raper, but are absent from the genomes of fungi and animals . WRKY genes, members of the WRKY-GCM1 superfamily , , contain a highly conserved DNA binding domain about 60 amino acids in length composed of the conserved WRKYGQK sequence followed by a C2H2- or C2HC-type zinc finger motif. Eulgem et al.  classified them into groups and subgroups based upon the number and type of WRKY domains, additional amino acid motifs, and phylogeny resolved from 58 loci isolated from Arabidopsis (DC.) Heynh. Group 1 WRKY genes were defined by the presence of two WRKY domains, each of the C2H2-type zinc finger motif, with only the C-terminal WRKY domain actively binding DNA. Group 2 WRKY genes contained only a single WRKY domain, and were classified into subgroups a-e based upon additional amino acid motifs found outside the WRKY domain. Group 3 WRKY genes were defined by the presence of the C2HC-type zinc finger motif in the DNA-binding WRKY domain. Further annotation of Arabidopsis WRKY genes  and phylogenies of Oryza L. and Arabidopsis WRKY gene families have resulted in several minor modifications to the original classification scheme , –. WRKY proteins are involved in several diverse pathways , – including regulation of starch, anthocyanin, and sesquiterpene anabolism; seed development , ; trichome development; embryogenesis; and plant responses to both abiotic and biotic stresses .
A feature common to WRKY genes is interruption of the coding region of the highly conserved, DNA-binding functional WRKY domain (the C-terminal WRKY domain in Group 1 WRKY genes and the single domains of Groups 2 and 3) with an intron. The size and sequence of the intron varies in each gene, but its position is highly conserved within each group/subgroup , , , . Variability present in the intron distinguishes among diverse WRKY loci isolated from a single species and, in many cases, allows for the design of primers specific to each , . The ancient origin and evolutionary expansion of the WRKY family was confirmed by discovery of a single WRKY gene each in D. discoideum, G. lamblia, and Chlamydamonas reinhardtii Dangeard, several from the moss Physcomitrella patens (Hedw.) B.S.G. and the fern Ceratopteris richardii Brongn.  to over 100 in Oryza sativa L. . This expansion was due primarily to large-scale duplications of entire genomic regions as a result of separate polyploid events , – throughout plant evolutionary time, rather than via tandem repeats, with a great deal of microsynteny retained even between evolutionarily distant plant species , . Rapid diversification of WRKY genes predates the divergence of monocots and dicots , .
WRKY loci and orthology/paralogy
Paralogy is the leading concern when using nuclear genes, especially members of multigene families. Paralogous sequences from gene duplication events due to unequal crossing over, replicative transpostition or ancient polyploidization events, when unrecognized, may lead to erroneous phylogenetic inferences , . Several lines of evidence, both direct and indirect, argue against paralogy as an issue with WRKY loci, although it cannot be ruled out conclusively without linkage mapping of the loci. In Oryza sativa, of the 102 WRKY loci described from subsp. indica, 99 are unigenes; of the 98 copies characterized in subsp. japonica, 97 are unique . For the seven loci chosen for this study, there was little indication, either by direct sequencing multiple individuals from a single species or from cloning, of paralogous copies for any of the individual WRKY loci within a single species, although allelic variation was detected (see Materials and Methods). Certainly, WRKY loci occurring in different groups, either in the Euglem et al.  classification or the more recent modification of Zhang and Wang  represent orthologs. Two of our loci belong to Group 2c (WRKY16 and WRKY19). These two loci were placed in different clusters of group 2c and each had greater identity with orthologs from Theobroma, Persea and Oryza than with each other. One locus (WRKY21) belongs to group 2b. Four of the loci belong to Group 1 (two WRKY domains). Two of these (WRKY6 and WRKY7) resolve in two distant clusters based on either the C- or N-terminus domains. Two loci (WRKY2, WRKY12), tended to cluster close together based on their conserved domains, thus raising the possibility that they could represent recent paralogs. Nonetheless, primers designed for each of these three amplifies a single product in the vast majority of species that we sampled, suggesting that their divergence occurred before the diversification of the Cocoseae. Further evidence of the independence of these loci is the fact that the sequences of all seven amplified from a single species could not be aligned except over portions of their highly conserved WRKY domains.
Phylogeny of the Attaleinae
Gunn's  analysis with the nuclear gene prk from across all 20 genera of Cocoseae supported recognition of a “spiny” clade (monophyletic Bactridinae and Elaeis; the monotypic genus Amazonian Barcella Drude resolved as sister to the rest of the tribe), and a “non-spiny” clade, i.e., the Attaleinae. Asmussen et al.'s  four plastid gene analysis of the entire family also resolved these two clades of Cocoseae with 91% BP (spiny clade) and 71% BP (non-spiny) , but with less than 50% BP for the tribe as a whole. The monophyly of subtribe Attaleinae is unquestionable (Figs. 1–2). WRKY loci resolve the African genera as a distinct clade sister to the American genera (Figs. 1–2), and Voanioala and Jubaeopsis as sister genera. Gunn's  analyses either resolved this clade as paraphyletic (BA and ML), or as a polytomy with the American clade (parsimony). These three genera all depart from the n = 16 chromosome number that is characteristic of the American clade . Beccariophoenix madagascariensis has n = 18, while Jubaeopsis (n = 80–100) and Voanioala (n = 298) have been suggested as autopolyploids . The WRKY resolution (Figs. 1–2) suggests that Madagascar rather than mainland Africa was the point of origin of the clade, with subsequent vicariance to South Africa. This is further supported by the fact that both Jubaeopsis and Voanioala are polyploids, while Beccariophoenix has a chromosome number much closer to the American Attaleinae.
The “Cocos alliance” resolved by WRKY genes consist of only Cocos nucifera, Syagrus and Lytocaryum (embedded in Syagrus), positioned as sister to the “Attalea alliance” of all other American genera. As in Gunn's  study, the sister relationship of Jubaea and Butia is very well supported, congruent with their similar leaf  and root ,  anatomy, but the novel clade positioning of Attalea as sister to Butia/Jubaea has high support only with ML and BA (Figs. 1–2).
The close relationship of Lytocaryum and Syagrus, inferred by previous studies with plastid  and nuclear  sequences, and evidenced morphologically as well ,  is here further corroborated. As in Gunn , one of our loci (WRKY7) resolves a sister relationship specifically with S. romanzoffiana (Table 1; Fig. S1). Syagrus ruschiana, which is sister to all of the other “rain forest” clade species, is, like Lytocaryum, a plant of the Brazilian Serra do Mar, and also bears fruits that split at the tip with thin exocarp and mesocarp, much like Lytocaryum. The sum of the evidence to date suggests that Lytocaryum should be transferred to Syagrus, but the low bootstrap support for its exact position relative to Syagrus in our combined tree (Fig. 1) might argue for awaiting further data.
The robust monophyly of Attalea, formerly split into as many as six genera largely on the basis of male floral characteristics –, is indisputably supported by the WRKY sequences (Figs. 1–2) – only two of the seven gene trees (WRKY7, WRKY16) fail to resolve the monophyly of this genus. The only Caribbean species in the genus, A. crassispatha, endemic to Haiti, is robustly sister to both a distinct Scheelea and two Orbignya subclades (Fig. 2). Glassman  placed this species in Orbignya. Within the monophyletic Scheelea clade, A. anisitsiana resolves consistently with two collections of A. phalerata, to which some workers have assigned it to synonomy , . Similarly, A. cohune and A. guacuyule, the latter considered synonomous with the former in some accounts , , resolve robustly as sister species.
The origins of the coconut
With three plastid genes , Cocos resolved as sister to Attalea H. B. K. in a weakly supported clade. In Gunn's study with the low copy nuclear gene prk , Cocos was positioned as sister to Parajubaea Burret, but only with ML. With MP or BA, its exact relationships were unresolved. Even a total evidence supermatrix/supertree approach, which included all molecular markers commonly used in palm systematics (13 in total), plus both morphological and RFLP datasets, did not help resolve the exact position of Cocos within the Attaleinae , placing the genus as sister to Parajubaea in a clade with only 52% BP. The WRKY consensus positions C. nucifera as sister to the South American genus Syagrus, and the clade has very strong support (Figs. 1–2), last sharing a common ancestor ca. 35 MYBP (Fig. 2, Table 5), though the crown node age of modern Cocos is estimated at only ca. 11 MYPB. The taxonomic history of Cocos and Syagrus has long been intertwined , , . In the first edition of Genera Palmarum , this relationship was explicitly noted, but was expunged from the second edition , based on Gunn's  weak resolution of Cocos as sister to Parajubaea with ML. Our analysis presents the strongest evidence to date for a close phylogenetic relationship between Cocos and Syagrus, and that the biogeographic ancestry of the coconut, regardless of its subsequent etnhobotanical history, is firmly rooted in South America. Although only two loci on their own explicitly support the sister relationship to Syagrus (WRKY7: Figs. S1, S3; WRKY19: Figs. S2, S4), only a single locus (WRKY2) in the combined analysis has incongruent partitioned Bremer support at the crown node for Cocos–Syagrus/Lytocaryum (node 45 in Fig. 1 and Table 2). The crown node of the six coconut genotypes included in our analysis terminates a branch 110 steps in length, one of the longest on the trees, raising the specter of long branch attraction (LBA) . Likelihood, which is less sensitive to LBA  also robustly supports this resolution, as do both our partitioned and non-partitioned Bayesian analyses. We tried several parsimony approaches that have been suggested to test for the presence of LBA : i.e., outgroup exclusion, and removing all or some Syagrus and/or Lytocaryum sequences from the alignment. Cocos still resolved as sister to Syagrus (inc. Lytocaryum) when outgroups were removed. When Syagrus and Lytocaryum were removed from the alignment, Cocos resolved as the first branch after the African clade and thus sister to the remaining American Attaleinae. With Syagrus excluded, Cocos resolved as sister to Lytocaryum. With even just a single species of Syagrus included in the analysis (we successively included just one species from each of the three main Syagrus subclades), Cocos resolved as its sister. We believe that the relationship between Cocos and Syagrus resolved by seven concatenated WRKY gene alignments is real, considering our substantial depth of sampling of the Attaleinae, and that the number of steps as well as the appreciable time duration (>20 MY) between the stem and crown nodes of Cocos reflects intervening extinction events.
This resolution might indicate that an Atlantic dispersal of the progenitors of the coconut is likely, perhaps along now submerged mid- to south Atlantic stepping stones, since the diversity of Syagrus is concentrated in Eastern Brazil (Figs. 1–2). But as this event precedes the Andean orogeny (Fig. 2), a Pacific coast dispersal from a broadly distributed lowland rain forest ancestor of Syagrus and Cocos cannot be ruled out. The resolution of the South Pacific coconut variety ‘Niu Leka’ as sister to all other cultivars would support the latter scenario as well, and is congruent with a Pacific Ocean dispersal scenario for coconut .
Unfortunately, beyond the sister relationship of ‘Niu Leka’ to all other varieties, the well-resolved clade of coconut cultivars, while demonstrating the utility of WRKY loci at infra-specific levels, can not be interpreted strictly because SSR studies indicate that three of the six individuals (‘Atlantic Tall’, ‘Pacific Tall’, and ‘Red Spicata’) used in this paper were introgressed with other cultivars . However, the degree of resolution suggests that WRKY loci could be successfully applied to a phylogeographic study of Cocos nucifera.
The Allagopetra-Parajubaea-Polyandrococos clade
The most surprising clade is that which unites Parajubaea with Allagoptera and Polyandrococos . Gunn's  analysis of prk sequences embedded Polyandrococos within Allagoptera, and Dransfield et al.  formally transferred the monotypic genus into Allagoptera. Polyandrococos is sister to Allagoptera, rather than embedded within it as Gunn  resolved with prk, thereby rendering Dransfield et al.'s  transfer of Polyandrococos to Allagoptera equivocal. Both Allagoptera and Polyandrococos have spicate inflorescences and share similar pinnae phylotaxy . The Andean Parajubaea, still uncertainly consisting of 1–3 species ,  has been variously treated as synonymous with Jubaea , Allagoptera  and Polyandrococos . Parajubaea was aligned with Butia, Jubaea and Syagrus by Uhl and Dranfield . Leaf anatomy (Fig. 4) supports the relationship among Allagoptera, Parajubaea and Polyandrococos.
Biogeographic and paleohistorical implications
Plant species exchange between Africa and South America continued after direct connections were severed in the late Cretaceous, at least through the early Paleocene, possibly via the Walvis Ridge/Rio-Grande Rise and Sierra Leone Ridges –. It was during the Paleocene that palms appear to have diversified greatly in Northern South America , , and the late Paleocene is where the stem age of the subtribe falls. In the early Eocene, fossil palm pollen was reduced in abundance in South America, perhaps due to climatic changes that occurred at the Paleocene/Eocene boundary but with little apparent loss of diversity . A date of ca. 43.7 MYBP (Table 5, Fig. 3) for the MRCA of the African and American Attaleinae is similar to Gunn's (2004), estimate with three of her four fossil calibration points (∼43 MYPB; any further comparison of our age estimates to Gunn (2004) is difficult due to the incongruent resolution of generic relationships between our trees and hers). This predates the terminal Eocene global cooling event, which for reasons unknown had little effect on the South American flora, including palms, but resulted in massive extinctions in west Africa, very notably of palms . However, the broad 95% HPD range (27.2–50.3 MYBP) at this node makes paleohistorical interpretation difficult. Thus, modern African representation of the Attaleinae is restricted to three relicts: a monotypic genus on the mainland (Jubaeopsis), and two Madagascar endemic genera (Beccariophoenix, 2 spp., and Voanioala, monotypic) that are both high polyploids, Voanioala with the highest chromosome number known in the monocotyledons .
Establishment of the two major clades of American Attaleinae on the other hand does appear congruent with the terminal Eocene cooling event (38.4 MYBP Table 5, Fig. 3), with subsequent cladogenesis in the Oligocene. Area optimization (Fig. 2) indicates that the ancestral American Attaleinae were restricted to Eastern Brazil during this time period. The MCRA of Allagoptera (including Polyandrococos), Butia, Jubaea and Parajubaea had dispersed to southern Brazil, the Andes and coastal Chile by the early Oligocene (Figs. 2, 3). The latter two may have been long distance dispersal events, as Jubaea consists of a single species of limited range in the central Chilean coast range, while Parajubaea is known only from cultivation in Andean Ecuador and Colombia (P. cocoides) and in the wild from the Andes of southern Bolivia (P sunkha and P. torallyi) , . Alternatively, these rare taxa may represent relicts of a once more broadly distributed lineage that suffered massive extinction. A vicariant event splitting the the MRCA of Jubaea/Butia and Parajubaea/Allagoptera (including Polyandrococos) in the proto-Andean region at this time is congruent with a Pacific marine incursion known as the Western Andean Portal (WAP) or Guayaquil gap, which a number of studies have proposed took place from the Eocene to the mid Miocene –, effectively disrupting exchange between the northern and southern Andean regions. The late Oligocene-early Miocene divergence between the Andean Parajubaea and the eastern Brazilian Allagoptera (including Polyandrococos) corresponds to the flooding of western Amazonia caused by the uplift of the Eastern Cordillera of the Central Andes in the early Miocene and subsequent Caribbean marine incursion to the north –. An enormous system of long-lived lakes and wetlands (“Lake Pebas” or the “Pebas Sea”) was formed that provided a significant barrier to dispersal from eastern South America to the Andes – that lasted until the Late Miocene. The divergence of Jubaea and Butia, however, appears to have occurred considerably later (14.5 MYBP), perhaps as the central Andes rose to sufficient elevation to obstruct dispersal or fragment a broad ancestral range of their MRCA.
Based on the estimated dates, much of the subsequent diversification in the Attaleinae can be attributed to the Andean uplift from late Miocene through the Pliocene –, and Pleistocene fluctuations in the extent and location of rain forest and seasonally dry climates in South America , . Species level divergences in Attalea, Butia and Syagrus are concentrated within the last 10 MYBP, much as Richardson et al.  determined for Inga Mill. (Fabaceae).
Other than the pantropical Cocos nucifera, the only two species of Attaleinae found in the Caribbean are both endemics: Attalea crassispatha (9 MYBP) in Haiti and Syagrus amara (8.4 MYBP) in the Lesser Antilles with essentially contemporaneous late Miocene divergence dates from their respective congenors.
WRKY genes in combination present the most fully resolved and well-supported phylogeny of the palm subtribe Attaleinae, and indicate a likely sister relationship between Cocos and Syagrus. Branch age estimates of our phylogeny are, in many cases, congruent with known paleohistorical events in South America. The biogeographic and morphological congruence that we see in the resolution of the larger genera such as Attalea and Syagrus (Figs. 1–2) suggest that WRKY loci are informative markers for investigating the phylogenetic relationships of Cocoseae, and should be tested further in the tribe, and perhaps other tribes of Arecoideae as well.
Materials and Methods
DNA was isolated from living accessions of 75 taxa (Table S1), mostly in cultivation at the Montgomery Botanical Center and the USDA-ARS-National Germplasm Repository: 72 from all genera of subtribe Attaleinae, and, as outgroups, two species of Bactris (subtribe Bactridinae) and Elaeis oleifera (Elaeidinae). The strategy for larger genera was to sample evenly from all recognized geographic or morphological groups. Multiple individuals were sampled for several species in several genera as a consistency check.
DNA was extracted from silica gel dried leaf samples using the BIO101 kit as described in Mauro-Herrera et al. . The quantity of DNA isolated was assessed with a GeneQuant pro RNA/DNA calculator (Amersham Pharmacia Biotech, Piscataway, N.J.). Isolated DNA was stored at −80°C until use.
WRKY gene isolation
By using a degenerate primer pair , , 21 WRKY sequences were isolated from Cocos nucifera . Of these, 12 were of sufficient size (>400 bp) to potentially yield a significant number of phylogenetically informative base substitutions, and specific primers were designed to maximize read length [the original primers of Mauro-Herrera et al.  were designed for short fragments flanking single nucleotide polymorphisms (SNPs)]. Ultimately, seven loci (Table S2) were selected that amplified a single product and which were able to be directly sequenced. The original sequences from C. nucifera were subjected to BLAST analyses against the non-redundant database at GenBank  and conceptual amino acid sequences of our eight loci were aligned with Arabidopsis thaliana, Theobroma cacao, Oryza sativa, and Persea americana WRKY proteins using ClustalX  to determine to which WRKY subgroup they belonged and to eliminate likely paralogs. Loci were chosen that had greater identity with orthologs from unrelated species than with orthologs from Cocos nucifera.
DNA Amplification and sequencing
Amplifications contained: 0.200 nM forward and reverse primer, 200 µM dNTPs, 1 mg/ml bovine serum albumin, 1 x amplification buffer with 2 mM MgSO4, 0.025 U/µl reaction volume Taq DNA Polymerase (New England Biolabs, Ipswich, MA) and 10 ng of template DNA brought to a total volume of 15 µls with nuclease-free H20. Amplifications were conducted using PTC-225 thermalcyclers (MJ Research, Waltham, MA). Conditions were: 95°C, 2 min; [95°C, 30 s; 57–64°C, 60 s; 72°C, 60 s] x 35 cycles; 72°C, 10 min; 4 C, hold. Amplification success was determined by agarose gel electrophoresis in 1.2% agarose, 0.5 x TBE buffer, ethidium bromide-stained, and visualized with UV light. Amplifications were treated with Exonuclease I and Shrimp Alkaline Phophatase to remove any unincorporated PCR primers and dNTP's, ethanol precipitated, and resuspended in sterile H2O. Direct sequencing was done in both directions on 1–2 µl of the treated amplification product with either the forward or reverse primer used for the initial amplification. All sequencing was done by capillary electrophoresis on an ABI 3100 or 3730 Genetic Analyzer using the BigDye Terminator Cycle Sequencing Ready Reaction Kit v3.1 (Applied Biosystems, Foster City, CA).
In the preliminary screening, seven WRKY loci produced a single band migrating at the expected mobility upon gel electrophoresis for each of three taxa amplified. Direct sequencing of the amplification products of most samples (Table S1) gave clean, clear signals with little or no noise for 60–100% of the taxa sampled. In some cases, double peaks gave evidence of allelic variation, and these were coded as ambiguities.
All sequences have been deposited in GenBank (Table S1) and assigned the following accession number ranges: WRKY2: FJ956927 - FJ956996, WRKY6: FJ957069 - FJ957142, WRKY7: FJ957143 - FJ957215, WRKY12: FJ957216 - FJ957283, WRKY16: FJ957284 - FJ957353, WRKY19: FJ957354 - FJ957428, WRKY21: FJ956997 - FJ957068.
Cloning was necessary for several taxa for six of the seven loci (Table S1). For pre-cloning PCR, AmpliTaq (Applied Biosystems, Foster City, CA) was used instead of NEB Taq polymerase. PCR products were cloned into pGEM-T vector, and the vector was transformed into JH109 High Efficiency Competent Cells following the instructions of the manufacturer (Promega, Madison, WI). Colonies were transferred to 96-well plates and incubated overnight at 37°C in SOC media with 100 µg/ml ampicillin. Transformed cells were lysed by resuspending the pelleted cells in 50 µls of 10 mM Tris-HCl pH 8.0. One µl of this was used as templates for PCR to confirm insert size on an agarose gel, and the PCR product was also used for the cycle sequencing reaction.
In most of the cases where cloning was necessary, clones showed allelic variation at SNPs or microsatellite repeats within introns, but consistently resolved as sisters with 100% BP when incorporated into the aligned matrix. For these, which constituted the majority, we used consensus sequences in our final alignments. Several species of Attalea, Butia and particularly Syagrus exhibited small indel polymorphisms among the clones, often evidence of interspecific hybridization, which has been reported in these three genera –. We noted that low frequency clones resolved at times with other species in their respective genera. For the present study, these clones were dropped. We plan to investigate this issue further with broader sampling.
Sequences were aligned using MAFFT ,  with subsequent manual editing in Sequencher™ 4.8 (Gene Codes Corporation, Ann Arbor, MI). The aligned lengths (Table 1) ranged from 658 nt (WRKY16) to 1277 nt (WRKY21).
Aligned sequences for the seven WRKY loci were analyzed separately and in combination using MP with PAUP* v. 4.0b10 , and with two model-based approaches, ML, utilizing TreeFinder  and, for the combined analysis only, BA, with MrBayes v. 3.1.2 , . Best fit nucleotide substitution model was determined for each gene region with KAKUSAN v.3 , which also generates input files for these two programs. Best fit models were evaluated using the corrected Akaike Information Criterion (AICc) ,  for ML and, for the BA, the Bayesian Information Criterion (BIC) . Significance of model fit statistics was determined by Chi-square analysis. For ML and BA, a mixed model, retaining each partition's best fit nucleotide substitution model, was applied.
MP tree searches were heuristic, conducted under the Fitch (equal) weights  criterion with 1000 rounds of random addition, saving no more than 10 minimum length trees per search for swapping using tree branch reconnection. Tree branches were collapsed if the minimum length = 0. Gaps were coded as missing characters in the initial analyses, but were also coded with the program SeqState , using the simple coding (SIC) of Simmons and Ochoterena . Before combining the data sets, we performed an incongruence length difference test (ILD = partition homogeneity test in PAUP* , ) on all pairwise combinations of loci to assess the degree of congruence between them. One hundred heuristic searches were conducted, each with 10 random addition replications, saving no more than 10 trees from each for TBR branch swapping. Internal support was determined by bootstrapping  (1000 heuristic replicates with simple addition, TBR branch-swapping, saving 10 trees per replicate). The cut-off BP value was 50%. A BP value >85% was considered good support, 75–85% was designated moderate support, and ≤ less than 75% as weak. For the combined analysis, both partitioned and non-partitioned Bremer (decay) indices  using TreeRot v. 3.0  were also calculated (Table 2). One hundred heuristic searches with random addition sequence were implemented for each constraint statement postulated by TreeRot, saving no more than 10 trees per search. A minimum DI = 2 was considered to represent good support for a clade , .
Treefinder  scripts generated by KAKUSAN  were used in part to conduct the maximum likelihood analysis. The first single search, using the best fit proportional model generated by KAKUSAN , was used to generate optimum substitution model and substitution rates in Treefinder , as well as a single tree. The optimum model and rates were specified for a second, unrestricted search in Treefinder , with the single tree produced from the first as a starting tree. One thousand replicates were run, with a search depth of two. ML bootstrap support was generated with 500 replicates, applying the same model and rates.
Two parallel runs were performed in MrBayes, each consisting of four chains, one “cold” and three incrementally heated. Elaeis oleifera was designated as outgroup. Two and one half million Markov chain Monte Carlo (MCMC) generations were run, with convergence diagnostics calculated every 1000th generation for monitoring the stabilization of log likelihood scores. Convergence was also evaluated for each of the two parallel runs using Tracer v.1.4 , with half (1.25 million) of the generations as burn-in. Effective sample sizes (ESS) of all parameters were >100. Trees in each chain were sampled every 100th generation. A 50% majority rule consensus tree was generated from the combined sampled trees of both runs after discarding the first 50% (12,500 trees, 6250 each run).
Gene tree incongruence testing
Different loci used to infer phylogenetic relationships of the same taxa often result in incongruent topologies , , due to such factors as horizontal gene transfer, gene duplication or loss, or deep coalescence , the latter leading to incomplete lineage sorting , . As we believe that our seven loci are orthologous , deep coalescence events are the likely reasons for “hard” incongruence among our gene trees. We attempted to resolve a species tree from the seven gene trees with a method more powerful than concatenation of the matrices. We first tried the program BEST v. 2.2 , which uses a Bayesian hierarchical model to estimate the phylogeny of a group of species using multiple estimated gene tree distributions . When, after 25 million generations (six weeks on a Pentium Xeon with 2 GB RAM), log likelihood scores had not stabilized, we concluded that our computing resources were insufficient for efficient use of the BEST program. We then turned to GeneTree v. 1.3.0 , which estimates a reconciled species tree with the lowest “cost” from multiple individual gene tree topologies , costs being defined as inferred gene duplications, losses, or deep coalescence events, depending on the gene family history. We generated new, fully dichotomized trees by running 500 random addition heuristic search replicates in PAUP* as described above, saving the 10 shortest trees from each for TBR swapping to a final limit of 1000. We then determined pair-wise tree-to-tree symmetric distances in PAUP* for each gene (as well as the trees from the combined analysis), then ran principle coordinate analysis (PCA) on the normalized pair-wise distance matrices using Multivariate Statistical Package (MVSP v. 3.13, Kovach Computing Services, Anglesey, Wales). Five trees that appeared broadly distributed in tree space were selected for each locus and the combined analysis as gene trees and species trees, respectively. We tested the fit of the gene trees to the species trees, and also generated de novo species trees by conducting 25 heuristic searches in GeneTree with random tree starting points, no constraints, and gene tree bootstrapping (the latter is the only way to get GeneTree to sample all of the trees provided to develop a species tree). A strict consensus tree was generated in PAUP* from the species trees found by GeneTree.
Divergence time estimation
Molecular dating was carried out using BEAST 1.4.8 . The method implemented in BEAST ,  simultaneously estimates divergence times, tree topology and rates, thereby providing a clear advantage over previous relaxed clock methods  that estimate tree topology and divergence dates separately –. For this study we relied on a fossil fruit from northern Colombia assigned to the Attaleineae and dated to ∼60 MYBP . Though the authors named this fossil as “cf. Cocos” , we believe that to assume homology of the impression to modern Cocos nucifera would be rash (the presence of germination pores in the endocarp–the most diagnostic fruit character for all of tribe Cocoseae , was not able to be confirmed in this fossil impression). It was thus conservatively placed at the stem node of the Attaleinae.
The ML tree, found with TreeFinder and rendered ultrametric using the program r8s , was used as starting tree for the BEAST runs. The likelihood ratio test (LRT) implemented in the program HyPhy  was used to assess whether a molecular clock could be applied to any of the loci, setting the optimal substitution model, rates and base frequencies determined by KAKUSAN . Based on the results of the LRTs, a global molecular clock was rejected for all seven WRKY loci, however HyPhy detected evidence of local clock-rate rates in portions of each tree. As the dataset deviated from a strict molecular clock model, a lognormal non-correlated relaxed clock model was specified in BEAST, and a general time- reversible substitution model with gamma-distributed rate heterogeneity (GTR+G) was invoked.
In order to accommodate for calibration uncertainty, we applied a normal distribution as a prior to the calibration node within the BEAST analysis with a mean of 60 myr and standard deviation of 1.5 (effectively enclosing dates from 558-62 MYBP). Although normal prior distributions are used when dating trees under an indirect approach , we prefer this type of distribution because it does not place a strict minimum age on the calibration. Indeed, the actual dating of the fossil is also subject to uncertainty and by allowing the ages to vary around the mean of the distribution appears as a more realistic choice.
A total of 10 different runs of 10 million generations each were undertaken on the online cluster of the Computational Biology Service Unit from Cornell University (http://cbsuapps.tc.cornell.edu/beast.aspx). This cluster imposes a time limit of three days (72 hours) per analysis but allows several runs of the same analysis simultaneously. Analyses were undertaken by sampling every 1000th generation. Tracer v.1.4  was used to check for convergence of the model likelihood and parameters between each run until reaching stationarity. The resulting log and tree files were then combined using LogCombiner. Results were considered reliable once the ESS of all parameters was above 100 (see results for the total number of generations). Branches with posterior probabilities (PP) below 0.8 were considered as weak, between 0.8 and 0.95 as moderate, and above 0.95 as strong.
All 10 runs in BEAST reached stationarity within the first 10,000 generations (because the ML starting tree was already near optimum), and all parameter estimates were consistent between runs. Runs were thus combined, after removing a burnin of 100 trees each (10,000 generations), into a single run of ca. 100 million generations. All parameters, including age estimates, reached acceptable ESS values and were thus deemed reliable (ESS>100). The tree files were combined after a burnin of 100 trees for each run. As the resulting combined tree file was too large to analyze in TreeAnnotator, combining of the different runs was redone. It was thus re-sampled at a lower frequency of every 5,000th tree resulting in a file containing 20,000 trees sampled from the posterior and used to generate a maximum clade credibility phylogram in TreeAnnotator (Fig. 3).
The results presented in this paper are based on a non-partitioned analysis. A partitioned analysis (using the closest approximations in BEAST of the models applied in ML and BA, Table 3) never reached stationarity. However, in the past we have observed that partitioned and non partitioned data returned identical age estimations (Couvreur, unpubl. data).
The biogeographic patterns inferred from our gene trees were assessed using the dispersal-vicariance method of analysis  as modeled by the program DIVA version 1.2 . The program uses vicariance (i.e., allopatric speciation) in its optimization of ancestral distributions but takes into consideration dispersal and extinction events and indicates their direction , . The most parsimonious reconstructions minimize such events. Unlike other biogeographic inference methods based on a strict vicariance model –, DIVA does not restrict widespread distributions to terminals or limit ancestral distributions to single areas . By allowing for dispersal and extinction as well as vicariance events within its model, DIVA does not impose adherence of area scenarios to a rigid “area cladogram.” It is thus much more amenable for biogeographic analysis within regions that have a complex paleogeological history, which a strict vicariance model cannot adequately address. Ancestral area optimizations in DIVA become less certain as the root node of the tree is approached. A weakness of the program is its assignment of nearly every area occupied by the terminal taxa in the tree to the more basal nodes, unless some type of constraint is imposed. Thus, the maximum areas allowed for ancestral nodes set to the minimum (two) to reduce ambiguities at the more basal nodes of the tree , , . An exact optimization (versus heuristic) was invoked by allowing the maximum number of alternative reconstructions to be held at any node. The fifteen areas assigned to the 75 terminal taxa in our matrix were: A = South Africa, B = Madagascar, C = Amazonas (north, east, south), D = Chile, E = Eastern Brazil, F = Central Brazil, G = Andes, H = Central America, I = Mexico, J = West Indies, K = Southern Brazil, L = Northern South America, i.e., Caribbean coastal Venezuela and Colombia, French Guiana, Guyana, Surinam, M = Argentina-Paraguay-Uruguay, N = Colombia-Venezuela (llanos region), and O = western Amazonas. For Cocos nucifera, which is pantropical, we assigned only those areas of these fifteen where the species is currently found.
Leaflet anatomical sections
Leaflets of Allagopetra, Parajubaea and Polyandrococos were hand sectioned with a double edged razor blade on a cutting board after folding the leaflet back and forth on itself various times to facilitate sectioning. Sections were not stained. Dried material was boiled for 5–10 minutes in water with Aerosol Detergent (Stepahn Co., Northfield, IL).
Strict consensus trees from parsimony analysis for loci WRKY2-12 (four loci).
(1.31 MB TIF)
Strict consensus trees from parsimony analysis for loci WRKY16, 19, and 21 (three loci).
(1.00 MB TIF)
Maximum likelihood bootstrap consensus trees for each of four WRKY loci: WRKY2, 6, 7, and 12.
(0.87 MB TIF)
Maximum likelihood bootstrap consensus trees for each of three WRKY loci: WRKY16, 19, and 21, as well as the combined analysis (all seven loci).
(0.94 MB TIF)
Strict consensus tree of 190 lowest “cost” reconciled species trees found by 25 heuristic searches by the program GeneTree using gene tree bootstrapping.
(0.08 MB TIF)
All 75 taxa used in the study are listed with voucher specimens and GenBank accession numbers for the WRKY sequences.
(0.12 MB DOC)
We gratefully acknowledge the receipt of leaf material of Jubaeopsis caffra from Huntington Botanical Garden, and DNA samples from Jean-Christophe Pintaud. We also thank William J. Baker for providing pertinent information from papers in press or in preparation, and Simon Joly, Richard Winkworth, and an anonymous reviewer for helpful comments on an earlier draft of this manuscript.
Conceived and designed the experiments: AWM LN JWB DNK. Performed the experiments: AWM LN JWB MMH WJH DNK KN NO. Analyzed the data: AWM TLPC KN. Contributed reagents/materials/analysis tools: AWM WJH RJS. Wrote the paper: AWM LN JWB TLPC.
- 1. Balick MJ, Beck HT (1990) Useful palms of the world. 724. New York: Columbia University Press. MJ BalickHT Beck1990Useful palms of the world.724New YorkColumbia University Press
- 2. Harries H (1978) The evolution, dissemination and classification of Cocos nucifera L. Bot Rev 44: 265–319.H. Harries1978The evolution, dissemination and classification of Cocos nucifera L.Bot Rev44265319
- 3. Harries HC (1995) Evolution of crop plants. In: Smartt J, Simmonds NW, editors. New York: Longman Scientific and Technical. pp. 389–395.HC Harries1995Evolution of crop plants.J. SmarttNW SimmondsNew YorkLongman Scientific and Technical389395
- 4. Gruezo WS, Harries HC (1984) Self-sown, wild-type coconuts in the Philippines. Biotropica 16: 140–147.WS GruezoHC Harries1984Self-sown, wild-type coconuts in the Philippines.Biotropica16140147
- 5. Gunn B (2004) The phylogeny of the Cocoeae (Arecaceae) with emphasis on Cocos nucifera. Ann Missouri Bot Gard 91: 505–522.B. Gunn2004The phylogeny of the Cocoeae (Arecaceae) with emphasis on Cocos nucifera.Ann Missouri Bot Gard91505522
- 6. de CandolleA (1886) Origin of cultivated plants (1959 English translation). New York: Hafner. 468 p.CandolleA de1886Origin of cultivated plants (1959 English translation).New YorkHafner468
- 7. Beccari O (1963) ‘I’he origin and dispersal of Cocos nucifera [reprint]. Principes 7: 57–69.O. Beccari1963‘I’he origin and dispersal of Cocos nucifera [reprint].Principes75769
- 8. Moore HE Jr (1973) The major groups of palm and their distribution. Gentes Herb 11: 27–111.HE Moore Jr1973The major groups of palm and their distribution.Gentes Herb1127111
- 9. Baudouin L, Lebrun P (2009) Coconut (Cocos nucifera L.) DNA studies support the hypothesis of an ancient Austronesian migration from Southeast Asia to America. Genet Res Crop Evol 56: 257–262.L. BaudouinP. Lebrun2009Coconut (Cocos nucifera L.) DNA studies support the hypothesis of an ancient Austronesian migration from Southeast Asia to America.Genet Res Crop Evol56257262
- 10. Asmussen CB, Baker WJ, Dransfield J (2000) Phylogeny of the palm family (Arecaceae) based on rps16 intron and trnL-trnF plastid DNA sequences. In: Wilson KL, Morrison DA, editors. Systematics and evolution of monocots. Collingwood: CSIRO Publishing. pp. 525–537.CB AsmussenWJ BakerJ. Dransfield2000Phylogeny of the palm family (Arecaceae) based on rps16 intron and trnL-trnF plastid DNA sequences.KL WilsonDA MorrisonSystematics and evolution of monocotsCollingwoodCSIRO Publishing525537
- 11. Asmussen CB, Dransfield J, Deickmann V, Barfod AS, Pintaud J-C, Baker WJ (2006) A new subfamily classification of the palm family (Arecaceae): evidence from plastid DNA. Bot J Linn Soc 151: 15–38.CB AsmussenJ. DransfieldV. DeickmannAS BarfodJ-C PintaudWJ Baker2006A new subfamily classification of the palm family (Arecaceae): evidence from plastid DNA.Bot J Linn Soc1511538
- 12. Hahn WJ (2002a) A molecular phylogenetic study of the Palmae (Arecaceae) based on atpB, rbcL, and 18S nrDNA sequences. Syst Biol 51: 92–112.WJ Hahn2002aA molecular phylogenetic study of the Palmae (Arecaceae) based on atpB, rbcL, and 18S nrDNA sequences.Syst Biol5192112
- 13. Hahn WJ (2002b) A phylogenetic analysis of the Arecoid line of palms based on plastid DNA sequence data. Mol Phyl Evol 23: 189–204.WJ Hahn2002bA phylogenetic analysis of the Arecoid line of palms based on plastid DNA sequence data.Mol Phyl Evol23189204
- 14. Dransfield J, Uhl NW, Asmussen CB, Baker WJ, Harley MM, et al. (2008) Genera palmarum: the evolution and classification of palms. Kew: Kew Publishing. 744 p.J. DransfieldNW UhlCB AsmussenWJ BakerMM Harley2008Genera palmarum: the evolution and classification of palms.KewKew Publishing744
- 15. Janick J, Paull RE (2008) Encyclopedia of fruits and nuts. Cambridge: CABI. 800 p.J. JanickRE Paull2008Encyclopedia of fruits and nuts.CambridgeCABI800
- 16. Uhl NW, Dransfield J (1987) Genera palmarum: a classification of palms based on the work of H. E. Moore, Jr. Lawrence, KS: International Palm Society and L. H. Bailey Hortorium610.NW UhlJ. Dransfield1987Genera palmarum: a classification of palms based on the work of H. E. Moore, Jr. Lawrence, KS: International Palm Society and L. H. Bailey Hortorium610
- 17. Dransfield J, Uhl NW, Asmussen CB, Baker WJ, Harley MM, et al. (2005) A new phylogenetic classification of the palm family, Arecaceae. Kew Bull 60: 559–569.J. DransfieldNW UhlCB AsmussenWJ BakerMM Harley2005A new phylogenetic classification of the palm family, Arecaceae.Kew Bull60559569
- 18. Henderson A, Galeano G, Bernal R (1995) Field Guide to the Palms of the Americas. Princeton: Princeton University Press. 363 p.A. HendersonG. GaleanoR. Bernal1995Field Guide to the Palms of the AmericasPrincetonPrinceton University Press363
- 19. Baker WJ, Savolainen V, Asmussen-Lange CB, Chase MW, Dransfield J, et al. (2009) Complete generic-level phylogenetic analyses of palms (Arecaceae) with comparisons of supertree and supermatrix approaches. Syst Biol 58: 240–256.WJ BakerV. SavolainenCB Asmussen-LangeMW ChaseJ. Dransfield2009Complete generic-level phylogenetic analyses of palms (Arecaceae) with comparisons of supertree and supermatrix approaches.Syst Biol58240256
- 20. Borrone JW (2004) The isolation, characterization, and application of WRKY genes as useful molecular markers in tropical trees (PhD dissertation). Miami: Florida International University. JW Borrone2004The isolation, characterization, and application of WRKY genes as useful molecular markers in tropical trees (PhD dissertation).MiamiFlorida International University
- 21. Borrone JW, Kuhn DN, Schnell RJ (2004) Isolation, characterization, and development of WRKY genes as useful genetic markers in Theobroma cacao. Theor Appl Genet 109: 495–507.JW BorroneDN KuhnRJ Schnell2004Isolation, characterization, and development of WRKY genes as useful genetic markers in Theobroma cacao.Theor Appl Genet109495507
- 22. Mauro-Herrera M, Meerow AW, Borrone JW, Kuhn DN, Schnell RJ (2006) Ten informative markers developed from WRKY sequences in coconut (Cocos nucifera). Mol Ecol Notes 6: 904–906.M. Mauro-HerreraAW MeerowJW BorroneDN KuhnRJ Schnell2006Ten informative markers developed from WRKY sequences in coconut (Cocos nucifera).Mol Ecol Notes6904906
- 23. Mauro-Herrera M, Meerow AW, Borrone JW, Kuhn DN, Schnell RJ (2007) Usefulness of WRKY gene-derived markers for assessing genetic population structure: An example with Florida coconut cultivars. Scientia Hort 115: 19–26.M. Mauro-HerreraAW MeerowJW BorroneDN KuhnRJ Schnell2007Usefulness of WRKY gene-derived markers for assessing genetic population structure: An example with Florida coconut cultivars.Scientia Hort1151926
- 24. Borrone JW, Meerow AW, Kuhn DN, Whitlock BA, Schnell RJ (2007) The potential of the WRKY gene family for phylogenetic reconstruction: an example from the Malvaceae. Mol Phyl Evol 44: 1141–1154.JW BorroneAW MeerowDN KuhnBA WhitlockRJ Schnell2007The potential of the WRKY gene family for phylogenetic reconstruction: an example from the Malvaceae.Mol Phyl Evol4411411154
- 25. Wendel JF, Doyle JJ (1998) Phylogenetic incongruence: Window into genome history and molecular evolution. In: Soltis DE, Soltis PS, Doyle JJ, editors. Molecular Systematics of Plants II: DNA Sequencing. Norwell: Kluwer Academic Publishers. pp. 265–296.JF WendelJJ Doyle1998Phylogenetic incongruence: Window into genome history and molecular evolution.DE SoltisPS SoltisJJ DoyleMolecular Systematics of Plants II: DNA SequencingNorwellKluwer Academic Publishers265296
- 26. Farris JS, Källersjö M, Kluge AG, Bult C (1994) Testing significance of incongruence. Cladistics 10: 315–319.JS FarrisM. KällersjöAG KlugeC. Bult1994Testing significance of incongruence.Cladistics10315319
- 27. Farris JS, Källersjö M, Kluge AG, Bult C (1995) Constructing a significance test for incongruence. Syst Biol 44: 570–572.JS FarrisM. KällersjöAG KlugeC. Bult1995Constructing a significance test for incongruence.Syst Biol44570572
- 28. Cunningham CW (1997a) Is congruence between data partitions a reliable predictor of phylogenetic accuracy? Empirically testing an iterative procedure for choosing among phylogenetic methods. Syst Biol 46: 464–478.CW Cunningham1997aIs congruence between data partitions a reliable predictor of phylogenetic accuracy? Empirically testing an iterative procedure for choosing among phylogenetic methods.Syst Biol46464478
- 29. Cunningham CW (1997b) Can three incongruence tests predict when data should be combined? Mol Biol Evol 14: 733–740.CW Cunningham1997bCan three incongruence tests predict when data should be combined?Mol Biol Evol14733740
- 30. Davis JI, Simmons MP, Stevenson DW, Wendel JF (1998) Data decisiveness, data quality, and incongruence in phylogenetic analysis: an example from the monocotyledons using mitochondrial atp A sequences. Syst Biol 47: 282–310.JI DavisMP SimmonsDW StevensonJF Wendel1998Data decisiveness, data quality, and incongruence in phylogenetic analysis: an example from the monocotyledons using mitochondrial atp A sequences.Syst Biol47282310
- 31. Desalle R, Brower VZ (1997) Process partitions, congruence, and the independence of characters: Inferring relationships among closely related Hawaiian Drospohila from multiple gene regions. Syst Biol 46: 751–764.R. DesalleVZ Brower1997Process partitions, congruence, and the independence of characters: Inferring relationships among closely related Hawaiian Drospohila from multiple gene regions.Syst Biol46751764
- 32. Flynn JJ, Nedbal MA (1998) Phylogeny of the Carnivora (Mammalia): congruence vs. incompatibility among multiple data sets. Mol Phy Evol 9: 414–26.JJ FlynnMA Nedbal1998Phylogeny of the Carnivora (Mammalia): congruence vs. incompatibility among multiple data sets.Mol Phy Evol941426
- 33. Messenger SL, Mcguire JA (1998) Morphology, molecules, and the phylogenetics of cetaceans. Syst Biol 47: 90–124.SL MessengerJA Mcguire1998Morphology, molecules, and the phylogenetics of cetaceans.Syst Biol4790124
- 34. Sidall ME (1997) Prior agreement: Arbitration or arbitrary? Syst Biol 46: 765–769.ME Sidall1997Prior agreement: Arbitration or arbitrary?Syst Biol46765769
- 35. Sullivan J (1996) Combining data with different distributions of among-site variation. Syst Biol 45: 375–380.J. Sullivan1996Combining data with different distributions of among-site variation.Syst Biol45375380
- 36. Yoder AD, Irwin JA, Payseur BA (2001) Failure of the ILD to determine data combinability for Slow Loris phylogeny. Syst Biol 50: 408–424.AD YoderJA IrwinBA Payseur2001Failure of the ILD to determine data combinability for Slow Loris phylogeny.Syst Biol50408424
- 37. Page RDM (1998) GeneTree: comparing gene and species phylogenies using reconciled trees. Bioinformatics 14: 819–820.RDM Page1998GeneTree: comparing gene and species phylogenies using reconciled trees.Bioinformatics14819820
- 38. Page RDM, Charleston MA (1997) From gene to organismal phylogeny: Reconciled trees and the gene tree/species tree problem. Mol Phyl Evol 7: 231–240.RDM PageMA Charleston1997From gene to organismal phylogeny: Reconciled trees and the gene tree/species tree problem.Mol Phyl Evol7231240
- 39. Drummond AJ, Rambaut A (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol 7: 214.AJ DrummondA. Rambaut2007BEAST: Bayesian evolutionary analysis by sampling trees.BMC Evol Biol7214
- 40. Zhang Y, Wang L (2005) The WRKY transcription factor superfamily: its origin in eukaryotes and expansion in plants. BMC Evol Biol 5: 1.Y. ZhangL. Wang2005The WRKY transcription factor superfamily: its origin in eukaryotes and expansion in plants.BMC Evol Biol51
- 41. Babu MM, Iyer LM, Balaji S, Aravind L (2006) The natural history of the WRKY-GCM1 zinc fingers and the relationship between transcription factors and transposons. Nucl Acids Res 34: 6505–6520.MM BabuLM IyerS. BalajiL. Aravind2006The natural history of the WRKY-GCM1 zinc fingers and the relationship between transcription factors and transposons.Nucl Acids Res3465056520
- 42. Yamasaki K, Kigawa T, Inoue M, Watanabe S, Tateno M, et al. (2008) Structures and evolutionary origins of plant-specific transcription factor DNA-binding domains. Plant Phys Biochem 46: 394–401.K. YamasakiT. KigawaM. InoueS. WatanabeM. Tateno2008Structures and evolutionary origins of plant-specific transcription factor DNA-binding domains.Plant Phys Biochem46394401
- 43. Eulgem T, Rushton PJ, Robatzek S, Somssich LE (2000) The WRKY superfamily of plant transcription factors. Trends Pl Sci 5: 199–206.T. EulgemPJ RushtonS. RobatzekLE Somssich2000The WRKY superfamily of plant transcription factors.Trends Pl Sci5199206
- 44. Dong J, Chen C, Chen Z (2003) Expression profiles of the Arabidopsis WRKY gene superfamily during plant defense response. Pl Mol Biol 51: 21–37.J. DongC. ChenZ. Chen2003Expression profiles of the Arabidopsis WRKY gene superfamily during plant defense response.Pl Mol Biol512137
- 45. Wu K-L, Guo Z-J, Wang H-H, Li J (2005) The WRKY family of transcription factors in rice and Arabidopsis and their origins. DNA Res 12: 9–26.K-L WuZ-J GuoH-H WangJ. Li2005The WRKY family of transcription factors in rice and Arabidopsis and their origins.DNA Res12926
- 46. Xie Z, Zhang Z-L, Zou X, Huang J, Ruas P, et al. (2005) Annotations and functional analyses of the rice WRKY gene superfamily reveal positive and negative regulators of abscisic acid signaling in aleurone cells. Pl Phys 137: 176–189.Z. XieZ-L ZhangX. ZouJ. HuangP. Ruas2005Annotations and functional analyses of the rice WRKY gene superfamily reveal positive and negative regulators of abscisic acid signaling in aleurone cells.Pl Phys137176189
- 47. Ülker B, Somssich LE (2004) WRKY transcription factors: from DNA binding towards biological function. Curr Opin Pl Biol 7: 491–498.B. ÜlkerLE Somssich2004WRKY transcription factors: from DNA binding towards biological function.Curr Opin Pl Biol7491498
- 48. Garcia D, Gerald JNF, Berger F (2005) Maternal control of integument cell elongation and zygotic control of endosperm growth are coordinated to determine seed size in Arabidopsis. Pl Cell 17: 52–60.D. GarciaJNF GeraldF. Berger2005Maternal control of integument cell elongation and zygotic control of endosperm growth are coordinated to determine seed size in Arabidopsis.Pl Cell175260
- 49. Luo M, Dennis ES, Berger F, Peacock WJ, Chaudhury A (2005) MINISEED3 (MINI3), a WRKY family gene, and HAIKU2 (IKU2), a leucine-rich repeat (LRR) KINASE gene, are regulators of seed size in Arabidopsis. Proc Nat Acad Sci U S A 102: 17531–17536.M. LuoES DennisF. BergerWJ PeacockA. Chaudhury2005MINISEED3 (MINI3), a WRKY family gene, and HAIKU2 (IKU2), a leucine-rich repeat (LRR) KINASE gene, are regulators of seed size in Arabidopsis.Proc Nat Acad Sci U S A1021753117536
- 50. Bowers JE, Chapman BA, Rong J, Paterson AH (2003) Unravelling angiosperm genome evolution by phylogenetic analysis of chromosomal duplication events. Nature 422: 433–438.JE BowersBA ChapmanJ. RongAH Paterson2003Unravelling angiosperm genome evolution by phylogenetic analysis of chromosomal duplication events.Nature422433438
- 51. Cannon SB, Mitra A, Baumgarten A, Young ND, May G (2004) The roles of segmental and tandem gene duplication in the evolution of large gene families in Arabidopsis thaliana. BMC Pl Biol 4: 10.SB CannonA. MitraA. BaumgartenND YoungG. May2004The roles of segmental and tandem gene duplication in the evolution of large gene families in Arabidopsis thaliana.BMC Pl Biol410
- 52. Thomas BC, Pedersen B, Freeling M (2006) Following tetraploidy in an Arabidopsis ancestor, genes were removed preferentially from one homeolog leaving clusters enriched in dose-sensitive genes. Genome Res 16: 934–946.BC ThomasB. PedersenM. Freeling2006Following tetraploidy in an Arabidopsis ancestor, genes were removed preferentially from one homeolog leaving clusters enriched in dose-sensitive genes.Genome Res16934946
- 53. Rossberg M, Theres K, Acarkan A, Herroro R, Schmitt T, et al. (2001) Comparative sequence analysis reveals extensive microlinearity in the Lateral Suppressor regions of the tomato, Arabidopsis, and Capsella genomes. Pl Cell 13: 979–988.M. RossbergK. TheresA. AcarkanR. HerroroT. Schmitt2001Comparative sequence analysis reveals extensive microlinearity in the Lateral Suppressor regions of the tomato, Arabidopsis, and Capsella genomes.Pl Cell13979988
- 54. Grover CE, Kim H, Wing RA, Paterson AH, Wendel JF (2004) Incongruent patterns of local and global genome size evolution in cotton. Genome Res 14: 1474–1482.CE GroverH. KimRA WingAH PatersonJF Wendel2004Incongruent patterns of local and global genome size evolution in cotton.Genome Res1414741482
- 55. Xiong Y, Liu T, Sun S, Li J, Chen M (2005) Transcription factors in rice: a genome-wide comparative analysis between monocots and eudicots. Pl Mol Biol 59: 191–203.Y. XiongT. LiuS. SunJ. LiM. Chen2005Transcription factors in rice: a genome-wide comparative analysis between monocots and eudicots.Pl Mol Biol59191203
- 56. Hughes CE, Eastwood RJ, Bailey CD (2006) From feast to famine? Selecting nuclear DNA sequence loci for plant species-level phylogeny reconstruction. Phil Tran Roy Soc Lond B Biol Sci 361: 211–225.CE HughesRJ EastwoodCD Bailey2006From feast to famine? Selecting nuclear DNA sequence loci for plant species-level phylogeny reconstruction.Phil Tran Roy Soc Lond B Biol Sci361211225
- 57. Martin AP, Burg TM (2002) Perils of paralogy: using HSP70 genes for inferring organismal phylogenies. Syst Biol 51: 570–587.AP MartinTM Burg2002Perils of paralogy: using HSP70 genes for inferring organismal phylogenies.Syst Biol51570587
- 58. Ross CA, Liu Y, Shen QJ (2007) The WRKY gene family in rice (Oryza sativa). J Integ Plant Biol 49: 827–842.CA RossY. LiuQJ Shen2007The WRKY gene family in rice (Oryza sativa).J Integ Plant Biol49827842
- 59. Read RW (1966) New chromosome counts in the Palmae. Principes 10: 55–61.RW Read1966New chromosome counts in the Palmae.Principes105561
- 60. Johnson MAT, Kenton AY, Bennett MD, Brandham PE (1989) Voanioala gerardii has the highest known chromosome number in the monocotyldons. Genome 32: 328–333.MAT JohnsonAY KentonMD BennettPE Brandham1989Voanioala gerardii has the highest known chromosome number in the monocotyldons.Genome32328333
- 61. Glassman S (1970) A conspectus of the palm genus Butia Becc. Fieldiana Bot 32: 127–172.S. Glassman1970A conspectus of the palm genus Butia Becc.Fieldiana Bot32127172
- 62. Seubert E (1998a) Root anatomy of palms. IV. Arceoideae, part 1. General remarks and descriptions of the roots. Feddes Rep 109: 89–127.E. Seubert1998aRoot anatomy of palms. IV. Arceoideae, part 1. General remarks and descriptions of the roots.Feddes Rep10989127
- 63. Seubert E (1998a) Root anatomy of palms. IV. Arceoideae, part 2. Systematic implications. Feddes Rep 109: 231–247.E. Seubert1998aRoot anatomy of palms. IV. Arceoideae, part 2. Systematic implications.Feddes Rep109231247
- 64. Henderson A, Balick M (1991) Attalea crassispatha, a rare and endemic Haitian palm. Brittonia 43: 198–194.A. HendersonM. Balick1991Attalea crassispatha, a rare and endemic Haitian palm.Brittonia43198194
- 65. Pintaud J-C (2008) An overview of the taxonomy of Attalea (Arecaceae). Rev Peru Biol 15 (supl 1): 55–63.J-C Pintaud2008An overview of the taxonomy of Attalea (Arecaceae).Rev Peru Biol15 (supl 1)5563
- 66. Glassman S (1999) A taxonomic treatment of the palm subtribe Attaleinae (tribe Cocoeae). Illinois Biol Monogr 59: 1–414.S. Glassman1999A taxonomic treatment of the palm subtribe Attaleinae (tribe Cocoeae).Illinois Biol Monogr591414
- 67. Glassman S (1987) Revisions of the palm genus Syagrus Mart. and other selected genera in the Cocos alliance. Illinois Biol Monogr 56: 1–230.S. Glassman1987Revisions of the palm genus Syagrus Mart. and other selected genera in the Cocos alliance.Illinois Biol Monogr561230
- 68. Bergsten J (2005) A review of long-branch attraction. Cladistics 21: 163–193.J. Bergsten2005A review of long-branch attraction.Cladistics21163193
- 69. Mauro-Herrera M, Meerow AW, Perera L, Russell J, Schnell RJ (2009) Ambiguous genetic relationships among coconut (Cocos nucifera L.) cultivars: the effects of outcrossing, sample source and size, and method of analysis. Genet Resour Crop Evol 56: M. Mauro-HerreraAW MeerowL. PereraJ. RussellRJ Schnell2009Ambiguous genetic relationships among coconut (Cocos nucifera L.) cultivars: the effects of outcrossing, sample source and size, and method of analysis.Genet Resour Crop Evol56DOI 10.1007/s10722-009-9463-x. DOI 10.1007/s10722-009-9463-x.
- 70. Moraes MR (1996) Novelties of the genera Parajubaea and Syagrus (Palmae) from interandean valleys of Bolivia. Novon 6: 85–92.MR Moraes1996Novelties of the genera Parajubaea and Syagrus (Palmae) from interandean valleys of Bolivia.Novon68592
- 71. Moraes MR, Henderson A (1990) The genus Parajubaea (Palmae). Brittonia 42: 92–99.MR MoraesA. Henderson1990The genus Parajubaea (Palmae).Brittonia429299
- 72. Wendland H (1878) Index général. Les palmiers, histoire iconographique. Paris: J Rothschild. 247 p. H. Wendland1878Index général.Les palmiers, histoire iconographiqueParisJ Rothschild247In: de Kerchove de Denterghem O, ed.
- 73. Kuntze CEO (1891) Revisio generum plantarum 2: 726.CEO Kuntze1891Revisio generum plantarum2726
- 74. Barbosa Rodrigues J (1901) Palmae-Polyandrococos. Contr Jard Bot Rio de Janeiro 1: 7–17.J. Barbosa Rodrigues1901Palmae-Polyandrococos.Contr Jard Bot Rio de Janeiro1717
- 75. MacDougal I, Douglas RA (1988) Age progressive volcanism in the Tasmanitid seamounts. Earth Planet Sci Let 89: 207–220.I. MacDougalRA Douglas1988Age progressive volcanism in the Tasmanitid seamounts.Earth Planet Sci Let89207220
- 76. Morley RJ (2003) Interplate dispersal paths for megathermal angiosperms. Persp Pl Ecol Evol Syst 6: 5–20.RJ Morley2003Interplate dispersal paths for megathermal angiosperms.Persp Pl Ecol Evol Syst6520
- 77. Theide J (1977) Subsidence of aseismic ridges: evidence from sediments on Rio Grande Rise (southwestern Atlantic Ocean). Amer Assoc Petr Geol Bull 61: 929–940.J. Theide1977Subsidence of aseismic ridges: evidence from sediments on Rio Grande Rise (southwestern Atlantic Ocean).Amer Assoc Petr Geol Bull61929940
- 78. Herngreen GFW, Chlonova AF (1981) Cretaceous microfloral provinces. Pollen et Spores 23: 441–555.GFW HerngreenAF Chlonova1981Cretaceous microfloral provinces.Pollen et Spores23441555
- 79. Morley RJ (2000) Origin and evolution of tropical rain forests. West Sussex: John Wiley and Sons. 378 p.RJ Morley2000Origin and evolution of tropical rain forests.West SussexJohn Wiley and Sons378
- 80. Nuttal CP (1990) A review of the Tertiary non-marine molluscan faunas of the Pebasian and other inland basins of north-western South America. Bull Brit Mus Nat Hist Geol 45: 165–371.CP Nuttal1990A review of the Tertiary non-marine molluscan faunas of the Pebasian and other inland basins of north-western South America.Bull Brit Mus Nat Hist Geol45165371
- 81. Hoorn C (1993) Marine incursions and the influence of Andean tectonics on the Miocene depositional history of northwestern Amazonia: Results of a palynostratigraphic study. Palaeogeogr Palaeocl 105: 267–309.C. Hoorn1993Marine incursions and the influence of Andean tectonics on the Miocene depositional history of northwestern Amazonia: Results of a palynostratigraphic study.Palaeogeogr Palaeocl105267309
- 82. Hoorn C, Guerrero J, Sarmiento GA, Lorente MA (1995) Andean tectonics as a cause of changing drainage patterns in Miocene northern South America. Geology 23: 237–240.C. HoornJ. GuerreroGA SarmientoMA Lorente1995Andean tectonics as a cause of changing drainage patterns in Miocene northern South America.Geology23237240
- 83. Steinmann M, Hungerbühler D, Seward D, Winkler W (1999) Neogene tectonic evolution and exhumation of the southern Ecuadorian Andes; a combined stratigraphy and fission-track approach. Tectonophysics 307: 255–276.M. SteinmannD. HungerbühlerD. SewardW. Winkler1999Neogene tectonic evolution and exhumation of the southern Ecuadorian Andes; a combined stratigraphy and fission-track approach.Tectonophysics307255276
- 84. Hungerbühler D, et al. (2002) Neogene stratigraphy and Andean geodynamics of southern Ecuador. Earth Sci Rev 57: 75–124.D. Hungerbühler2002Neogene stratigraphy and Andean geodynamics of southern Ecuador.Earth Sci Rev5775124
- 85. Santos C, Jaramillo C, Bayona G, Rueda M, Torres V (2008) Late Eocene marine incursion in north-western South America. Palaeogeogr Palaeocl 264: 140–146.C. SantosC. JaramilloG. BayonaM. RuedaV. Torres2008Late Eocene marine incursion in north-western South America.Palaeogeogr Palaeocl264140146
- 86. Lundberg JG, Marshall LG, Guerrero J, Horton B, Malabarba MCSL, et al. (1998) The stage for Neotropical fish diversification: A history of tropical South American rivers. In: Malabarba LR, Reis RE, Vari RP, Lucena ZM, Lucena CAS, editors. Phylogeny and classification of neotropical fishes. Porto Alegre: Edipucrs. pp. 13–48.JG LundbergLG MarshallJ. GuerreroB. HortonMCSL Malabarba1998The stage for Neotropical fish diversification: A history of tropical South American rivers.LR MalabarbaRE ReisRP VariZM LucenaCAS LucenaPhylogeny and classification of neotropical fishesPorto AlegreEdipucrs1348
- 87. Wesselingh FP, Räsänen ME, Irion G, Vonhof HB, Kaandorp R, et al. (2002) Lake Pebas: A palaeoecological reconstruction of a Miocene, ong-lived lake complex in western Amazonia. Cenozoic Res 1: 35–81.FP WesselinghME RäsänenG. IrionHB VonhofR. Kaandorp2002Lake Pebas: A palaeoecological reconstruction of a Miocene, ong-lived lake complex in western Amazonia.Cenozoic Res13581
- 88. Wesselingh FP (2006) Miocene long-lived lake Pebas as a stage of mollusc radiations, with implications for landscape evolution in western Amazonia. Scripta Geol 133: 1–17.FP Wesselingh2006Miocene long-lived lake Pebas as a stage of mollusc radiations, with implications for landscape evolution in western Amazonia.Scripta Geol133117
- 89. Wesselingh FP, Salo JA (2006) A Miocene perspective on the evolution of the Amazonian biota. Scripta Geol 133: 439–458.FP WesselinghJA Salo2006A Miocene perspective on the evolution of the Amazonian biota.Scripta Geol133439458
- 90. Antonelli A, Nylander JAA, Persson C, Sanmartín I (2009) Tracing the impact of the Andean uplift on Neotropical plant evolution. Proc Nat Acad Sci USA 106: 9749–9754.A. AntonelliJAA NylanderC. PerssonI. Sanmartín2009Tracing the impact of the Andean uplift on Neotropical plant evolution.Proc Nat Acad Sci USA10697499754
- 91. van der Hammen T, Werner JH, van Dommelen H (1973) Palynological record of the upheaval of the northern Andes: a study of the Pliocene and Lower Quaternary of the Colombian Eastern Cordillera and the early evolution of its high-Andean biota. Rev Palynol Paleobot 16: 1–122.T. van der HammenJH WernerH. van Dommelen1973Palynological record of the upheaval of the northern Andes: a study of the Pliocene and Lower Quaternary of the Colombian Eastern Cordillera and the early evolution of its high-Andean biota.Rev Palynol Paleobot161122
- 92. van der Hammen T (1979) History of the flora, vegetation and climate in the Cordillera Oriental during the last five million years. In: Larsen H, Holm-Nielsen LB, editors. Tropical Botany. London: Academic Press. pp. 25–32.T. van der Hammen1979History of the flora, vegetation and climate in the Cordillera Oriental during the last five million years.H. LarsenLB Holm-NielsenTropical BotanyLondonAcademic Press2532
- 93. Garzione CN, Hoke GD, Libarkin JC, Withers S, MacFadden B, et al. (2008) Rise of the Andes. Science 320: 1304–1307.CN GarzioneGD HokeJC LibarkinS. WithersB. MacFadden2008Rise of the Andes.Science32013041307
- 94. Haffer J (1987) Quaternary history of tropical America. In: Whitmore TC, Prance GT, editors. Biogeography and Quaternary history in tropical America. Oxford: Clarendon Press. pp. 1–18.J. Haffer1987Quaternary history of tropical America.TC WhitmoreGT PranceBiogeography and Quaternary history in tropical AmericaOxfordClarendon Press118
- 95. Prance GT (1987) Biogeography of Neotropical plants. In: Whitmore TC, Prance GT, editors. Biogeography and Quaternary history in tropical America. Oxford: Clarendon Press. pp. 46–55.GT Prance1987Biogeography of Neotropical plants.TC WhitmoreGT PranceBiogeography and Quaternary history in tropical AmericaOxfordClarendon Press4655
- 96. Richardson JE, Pennington RT, Pennington TD, Hollingsworth PM (2001) Rapid diversification of a species-rich genus of neotropical rain forest trees. Science 293: 2242–2245.JE RichardsonRT PenningtonTD PenningtonPM Hollingsworth2001Rapid diversification of a species-rich genus of neotropical rain forest trees.Science29322422245
- 97. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, et al. (1997) GappedBLAST and PSI-BLAST: a new generation of protein database search programs. Nucl Acids Res 25: 3389–3402.SF AltschulTL MaddenAA SchafferJ. ZhangZ. Zhang1997GappedBLAST and PSI-BLAST: a new generation of protein database search programs.Nucl Acids Res2533893402
- 98. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG (1997) The ClustalX windows interface: flexible strategies for mutiple sequence alignment aided by quality analysis tools. Nucl Acids Res 24: 4876–4882.JD ThompsonTJ GibsonF. PlewniakF. JeanmouginDG Higgins1997The ClustalX windows interface: flexible strategies for mutiple sequence alignment aided by quality analysis tools.Nucl Acids Res2448764882
- 99. Balick MJ, Pinheiro CUB, Anderson AB (1987) Hybridization in the babassu palm complex: I. Orbignya phalerata x O. eichleri. Am J Bot 74: 1013–1032.MJ BalickCUB PinheiroAB Anderson1987Hybridization in the babassu palm complex: I. Orbignya phalerata x O. eichleri.Am J Bot7410131032
- 100. Glassman S (1972) A new hybrid in the palm genus Syagrus Mart. Fieldiana Bot 32: 241–257.S. Glassman1972A new hybrid in the palm genus Syagrus Mart.Fieldiana Bot32241257
- 101. Noblick LR (1994) Palms of Bahia. Acta Hort 360: 85–94.LR Noblick1994Palms of Bahia.Acta Hort3608594
- 102. Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucl Acids Res 30: 3059–3066.K. KatohK. MisawaK. KumaT. Miyata2002MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform.Nucl Acids Res3030593066
- 103. Katoh K, Kuma K, Toh H, Misawa K, Miyata T (2005) MAFFT version 5: improvement in accuracy of multiple sequence alignment. Nucl Acids Res 33: 511–518.K. KatohK. KumaH. TohK. MisawaT. Miyata2005MAFFT version 5: improvement in accuracy of multiple sequence alignment.Nucl Acids Res33511518
- 104. Swofford DL (2002) PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods) version 4.10. Sunderland: Sinauer Associates. DL Swofford2002PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods) version 4.10.SunderlandSinauer Associates
- 105. Jobb GG. Jobb(2008) TREEFINDER October 2008 version (www.treefinder.de, Munich). (2008) TREEFINDER October 2008 version (www.treefinder.de, Munich).
- 106. Huelsenbeck JP, Ronquist F (2001) MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17: 754–755.JP HuelsenbeckF. Ronquist2001MRBAYES: Bayesian inference of phylogeny.Bioinformatics17754755
- 107. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574.F. RonquistJP Huelsenbeck2003MrBayes 3: Bayesian phylogenetic inference under mixed models.Bioinformatics1915721574
- 108. Tanabe AS (2007) Kakusan: a computer program to automate the selection of a nucleotide substitution model and the configuration of a mixed model on multilocus data. Mol Ecol Notes 7: 962–964.AS Tanabe2007Kakusan: a computer program to automate the selection of a nucleotide substitution model and the configuration of a mixed model on multilocus data.Mol Ecol Notes7962964
- 109. Akaike H (1973) Information theory and an extension of the maximum likelihood principle. In: Petrov BN, Csaki F, editors. Second International Symposium on Information Theory. Budapest: Akademia Kiado. pp. 267–281.H. Akaike1973Information theory and an extension of the maximum likelihood principle.BN PetrovF. CsakiSecond International Symposium on Information TheoryBudapestAkademia Kiado267281
- 110. Shono H (2000) Efficiency of the finite correction of Akaike's Information Criteria. Fisheries Sci 66: 608–610.H. Shono2000Efficiency of the finite correction of Akaike's Information Criteria.Fisheries Sci66608610
- 111. Schwarz G (1978) Estimating the dimension of a model. Ann Stat 6: 461–46.G. Schwarz1978Estimating the dimension of a model.Ann Stat646146
- 112. Fitch WM (1971) Toward defining the course of evolution: minimum change for a specific tree topology. Syst Zool 20: 406–416.WM Fitch1971Toward defining the course of evolution: minimum change for a specific tree topology.Syst Zool20406416
- 113. Müller KF (2005) SeqState - primer design and sequence statistics for phylogenetic DNA data sets. Appl Bioinformatics 4: 65–69.KF Müller2005SeqState - primer design and sequence statistics for phylogenetic DNA data sets.Appl Bioinformatics46569
- 114. Simmons MP, Ochoterena H (2000) Gaps as characters in sequence based phylogenetic analyses. Syst Biol 49: 369–381.MP SimmonsH. Ochoterena2000Gaps as characters in sequence based phylogenetic analyses.Syst Biol49369381
- 115. Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791.J. Felsenstein1985Confidence limits on phylogenies: an approach using the bootstrap.Evolution39783791
- 116. Bremer K (1988) The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 198–213.K. Bremer1988The limits of amino acid sequence data in angiosperm phylogenetic reconstruction.Evolution42198213
- 117. Sorenson MD, Franzosa EAMD SorensonEA Franzosa(2007) TreeRot, version 3 (Boston University, Boston). (2007) TreeRot, version 3 (Boston University, Boston).
- 118. Meerow AW, van der Werff H (2004) Pucara (Amaryllidaceae) reduced to synonymy with Stenomesson on the basis of nuclear and plastid DNA spacer sequences, and a new related species of Stenomesson. Syst Bot 29: 511–517.AW MeerowH. van der Werff2004Pucara (Amaryllidaceae) reduced to synonymy with Stenomesson on the basis of nuclear and plastid DNA spacer sequences, and a new related species of Stenomesson.Syst Bot29511517
- 119. Meerow AW, Francisco-Ortega J, Kuhn DN, Schnell RJ (2006) Phylogenetic relationships and biogeography within the Eurasian clade of Amaryllidaceae based on plastid ndhF and nrDNA ITS sequences: lineage sorting in a reticulate area? Syst Bot 31: 42–60.AW MeerowJ. Francisco-OrtegaDN KuhnRJ Schnell2006Phylogenetic relationships and biogeography within the Eurasian clade of Amaryllidaceae based on plastid ndhF and nrDNA ITS sequences: lineage sorting in a reticulate area?Syst Bot314260
- 120. Rambaut A, Drummond AJA. RambautAJ Drummond(2007) Tracer v1.4, Available from http://beast.bio.ed.ac.uk/Tracer. (2007) Tracer v1.4, Available from http://beast.bio.ed.ac.uk/Tracer.
- 121. Hughes CE, Eastwood RJ, Bailey CD (2006) From feast to famine? Selecting nuclear DNA sequence loci for plant species-level phylogeny reconstruction. Phil Tran Roy Soc Lond B Biol Sci 361: 211–225.CE HughesRJ EastwoodCD Bailey2006From feast to famine? Selecting nuclear DNA sequence loci for plant species-level phylogeny reconstruction.Phil Tran Roy Soc Lond B Biol Sci361211225
- 122. Rokas A, Williams BL, King N, Carroll SB (2003) Genome-scale approaches to resolving incongruence in molecular phylogenies. Nature 425: 798–804.A. RokasBL WilliamsN. KingSB Carroll2003Genome-scale approaches to resolving incongruence in molecular phylogenies.Nature425798804
- 123. Maddison WP (1997) Gene trees in species trees. Syst Biol 46: 523–536.WP Maddison1997Gene trees in species trees.Syst Biol46523536
- 124. Maddison WP, Knowles LL (2006) Inferring phylogeny despite incomplete lineage sorting. Syst Biol 55: 21–30.WP MaddisonLL Knowles2006Inferring phylogeny despite incomplete lineage sorting.Syst Biol552130
- 125. Liu L (2008) BEST: Bayesian estimation of species trees under the coalescent model. Bioinformatics 24: 2542–2543.L. Liu2008BEST: Bayesian estimation of species trees under the coalescent model.Bioinformatics2425422543
- 126. Liu L, Pearl DK (2007) Species Trees from Gene Trees: reconstructing Bayesian posterior distributions of a species phylogeny using estimated gene tree distributions. Syst Biol 56: 504–514.L. LiuDK Pearl2007Species Trees from Gene Trees: reconstructing Bayesian posterior distributions of a species phylogeny using estimated gene tree distributions.Syst Biol56504514
- 127. Drummond AJ, Ho SYW, Phillips MJ, Rambaut A (2006) Relaxed phylogenetics and dating with confidence. PloS Biol 4: 699–710.AJ DrummondSYW HoMJ PhillipsA. Rambaut2006Relaxed phylogenetics and dating with confidence.PloS Biol4699710
- 128. Ho SYW, Phillips MJ, Drummond AJ, Cooper A (2005) Accuracy of rate estimation using relaxed-clock models with a critical focus on the early metazoan radiation. Mol Biol Evol 22: 1355–1363.SYW HoMJ PhillipsAJ DrummondA. Cooper2005Accuracy of rate estimation using relaxed-clock models with a critical focus on the early metazoan radiation.Mol Biol Evol2213551363
- 129. Sanderson M (1997) A nonparametric approach to estimating divergence times in the absence of rate constancy. Mol Biol Evol 14: 1218–1231.M. Sanderson1997A nonparametric approach to estimating divergence times in the absence of rate constancy.Mol Biol Evol1412181231
- 130. Sanderson MJ (2002) Estimating absolute rates of molecular evolution and divergence times: A penalized likelihood approach. Mol Biol Evol 19: 101–109.MJ Sanderson2002Estimating absolute rates of molecular evolution and divergence times: A penalized likelihood approach.Mol Biol Evol19101109
- 131. Thorne JL, Kishino H, Painter IS (1998) Estimating the rate of evolution of the rate of molecular evolution. Mol Biol Evol 15: 1647–1657.JL ThorneH. KishinoIS Painter1998Estimating the rate of evolution of the rate of molecular evolution.Mol Biol Evol1516471657
- 132. Gomez-N C, Jaramillo C, Herrera F, Wing SL, Callejas R (2009) Palms (Arecaceae) from a Paleocene rain forest of northern Colombia. Am J Bot 96: 1300–1312.C. Gomez-NC. JaramilloF. HerreraSL WingR. Callejas2009Palms (Arecaceae) from a Paleocene rain forest of northern Colombia.Am J Bot9613001312
- 133. Sanderson MJ (2003) r8s: inferring absolute rates of molecular evolution and divergence times in the absence of a molecular clock. Bioinformatics 19: 301–302.MJ Sanderson2003r8s: inferring absolute rates of molecular evolution and divergence times in the absence of a molecular clock.Bioinformatics19301302
- 134. Kosakovsky SL, Pond SD, Frost W, Muse SV (2005) HyPhy: hypothesis testing using phylogenies. Bioinformatics 21: 676–679.SL KosakovskySD PondW. FrostSV Muse2005HyPhy: hypothesis testing using phylogenies.Bioinformatics21676679
- 135. Ho SYW (2007) Calibrating molecular estimates of substitution rates and divergence times in birds. J Avian Biol 38: 409–414.SYW Ho2007Calibrating molecular estimates of substitution rates and divergence times in birds.J Avian Biol38409414
- 136. Ronquist F (1997) Dispersal-vicariance analysis: a new approach to the quantification of historical biogeography. Syst Biol 46: 195–203.F. Ronquist1997Dispersal-vicariance analysis: a new approach to the quantification of historical biogeography.Syst Biol46195203
- 137. Ronquist F (1996) DIVA v. 1.1 and 1.2. Computer program for MacOS and Win32. F. Ronquist1996DIVA v. 1.1 and 1.2. Computer program for MacOS and Win32.Available from http://www.ebc.uu.se/systzoo/research/diva/diva.html. Available from http://www.ebc.uu.se/systzoo/research/diva/diva.html.
- 138. Nelson GJ, Platnick NI (1978) A method of analysis for historical biogeography. Syst Zool 27: 1–16.GJ NelsonNI Platnick1978A method of analysis for historical biogeography.Syst Zool27116
- 139. Brooks DR (1990) Parsimony analysis in historical biogeography and coevolution: methodological and theoretical update. Syst Zool 39: 14–30.DR Brooks1990Parsimony analysis in historical biogeography and coevolution: methodological and theoretical update.Syst Zool391430
- 140. Page RDM (1994) Maps between trees and cladistic analysis of relationships among genes, organisms, and areas. Syst Biol 43: 58–77.RDM Page1994Maps between trees and cladistic analysis of relationships among genes, organisms, and areas.Syst Biol435877
- 141. Meerow AW, Lehmiller DJ, Clayton JR (2003) Phylogeny and biogeography of Crinum L. (Amaryllidaceae) inferred from nuclear and limited plastid non-coding DNA sequences. Bot J Linn Soc 141: 349–363.AW MeerowDJ LehmillerJR Clayton2003Phylogeny and biogeography of Crinum L. (Amaryllidaceae) inferred from nuclear and limited plastid non-coding DNA sequences.Bot J Linn Soc141349363
- 142. Sanmartín I (2003) Dispersal vs. vicariance in the Mediterranean: historical biogeography of the Palearctic Pachydeminae (Coleoptera, Scarabaedoidea). J Biogeogr 30: 1883–1897.I. Sanmartín2003Dispersal vs. vicariance in the Mediterranean: historical biogeography of the Palearctic Pachydeminae (Coleoptera, Scarabaedoidea).J Biogeogr3018831897