Davis et al. BMC Biology 2014, 12:11 http://www.biomedcentral.com/1741-7007/12/11 COMMENTARY Open Access Plastid phylogenomics and green plant phylogeny: almost full circle but not quite there Charles C Davis*, Zhenxiang Xi and Sarah Mathews Congruence and conflict in plastid phylogenomics Abstract Ruhfel et al. present a phylogeny that is well resolved at A study in BMC Evolutionary Biology represents the most nodes, and largely in agreement with previous most comprehensive effort to clarify the phylogeny studies, including at nodes that have been difficult to of green plants using sequences from the plastid resolve (Figure 1). These include the splits between genome. This study highlights the strengths and land plants and their algal sister clade [3,4], and be- limitations of plastome data for resolving the green tween vascular plants and their non-vascular sister plant phylogeny, and points toward an exciting future clade [5]. Here, Zygnematophyceae, a large clade of for plant phylogenetics, during which the vast and mostly freshwater algal species, is identified as sister to largely untapped territory of nuclear genomes will land plants. This suggests that shared components of be explored. auxin signaling and chloroplast movement likely were present in their common ancestor [3]. Their analyses See research article: also support the non-monophyly of bryophytes, or http://www.biomedcentral.com/1471-2148/14/23 liverworts, mosses and hornworts. These land plants lack a well developed vascular system and have simi- lar ecologies. Hornworts are sister to vascular plants in Commentary the plastid tree, consistent with evidence that their spo- The plastid genome, or plastome, has so far been the rophytes may be at least partially free-living, unlike most important source of data for plant phylogenetics in thoseofliverwortsandmosses[5]. the era of comparative DNA sequencing. Its utility re- So how much closer does this new phylogeny bring us sults from its relatively small size (between 75 and 250 to a robust understanding of green plant evolution? This kilobases), largely uniparental inheritance, conservation study, like many others, has difficulty resolving key rela- of gene content and order, and its high copy number in tionships within green plants. This is most evident in green plant cells. From the early use of a single plastid the lack of resolution deep in the angiosperm phylogeny gene to infer the phylogeny of a broad sampling of among the mesangiosperm clades, Ceratophyllum, seed plants [1], to the now common use of around 80 Chloranthaceae, eudicots, magnoliids, and monocots. plastid genes to address finer-scale phylogenetic ques- Branching order among non-vascular plants, especially tions, this circular genome has been a mainstay for involving liverworts and mosses, remains contentious. evolutionary botanists. These problems persist due, in part, to the challenge of Efforts to understand green plant phylogeny from placing lineages that are species-poor and divergent in plastome data have now come full circle. In this issue, molecular trees, and due to the difficulty of assessing Ruhfel et al. [2] report results from their analyses of homologies among organisms with very diverse or 78 plastid genes from 360 species, from green algae to reduced morphologies. angiosperms. Their results provide insights into this Despite these few persistent problems, the Rhufel et ongoing effort, adding support for some relationships al. tree appears to address many if not all remaining and highlighting phylogenetic questions that require questions. In some cases, however, high support for rela- more data, especially from nuclear genomes. tionships should be interpreted cautiously because con- flicting topologies are supported by other data. Key * Correspondence: [email protected] examples include the previously mentioned sister groups Department of Organismic and Evolutionary Biology, Harvard University, 22 of land plants and vascular plants, but also relationships Divinity Ave, Cambridge, MA 02138, USA © 2014 Davis et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Davis et al. BMC Biology 2014, 12:11 Page 2 of 4 http://www.biomedcentral.com/1741-7007/12/11 1 2 3 4 5 1 4 2 5 3 chlorophytesMesostigmatophyceaeChlorokybophyceaeCharophyceaeColeochaetophyceaeZygnematophyceaemossesliverwortshornwortslycophytesferns cycadsGinkgoPinaceae bilobacupressophytesgnetophytesAmborellaNymphaeales trichopodaAustrobaileyalesCeratophyllumChloranthaceaeeudicotsmagnoliidsmonocots mesangiosperms living gymnosperms angiosperms seed plants vascular plants land plants green plants Figure 1. Simplified green plant phylogeny inferred by Ruhfel et al. [2]. Dashed line branches indicate phylogenetic placements that remain unresolved by the plastome. Relationships that are well resolved in the plastome tree but contentious are highlighted in red. The green algal lineage Klebsormidiophyceae was not included in the plastome study, and thus was not included in the figure. among major seed plant clades. The latter involves the biloba, where their plastid trees strongly unite the two, position of gnetophytes, and the close relationship of cy- but other studies place cycads alone as sister to extant cads and Ginkgo biloba. The gnetophytes especially rep- gymnosperms. resent a vexing problem in seed plant phylogenetics [6]. They form a small clade of approximately 90 species that Where do we go next? are highly divergent from other seed plants in both It is well known that biases within molecular data may morphology and molecules. Although plastid phylo- be exacerbated in large phylogenomic data sets, leading genomic studies are converging on the ‘Gnecup’ topology, to erroneous but well supported results, especially when in which gnetophytes are united with cupressophyte coni- trying to resolve ancient splits. Biases may result from fers, recent nuclear phylogenomic analyses yield the alter- phenomena such as pattern heterogeneity and uneven native ‘Gnepine’ topology in which gnetophytes are united base frequencies, and their exploration will help us to with Pinaceae conifers [7]. Even within individual plastid understand cases of incongruent relationships. For ex- loci, different nucleotide sites have been shown to favor ample, when Ruhfel et al. accounted for biased GC con- rival gnetophyte placements [6]. A similarly strong conflict tent, their data placed lycophytes as sister to ferns and in seed plants concerns the positions of cycads and Ginkgo seed plants, rather than as the lone sister to seed plants, Davis et al. BMC Biology 2014, 12:11 Page 3 of 4 http://www.biomedcentral.com/1741-7007/12/11 as in their total evidence tree. Additional approaches for lines, recent analyses already indicate that potential mitigating biases in molecular data sets include increas- plastid-nuclear genome conflicts involve the gneto- ing taxon sampling and better modeling of nucleotide phytes, early diverging flowering plants, and the evolution. Traditional data partitioning schemes try to large flowering plant orders Lamiales, Malpighiales, account for variation in evolutionary rates by partition- and Myrtales. Evaluating the extent to which these ing nucleotide sites based on their rate of evolution, incongruent placements demonstrate divergent ge- most commonly, by gene or codon position. The ap- nome histories requires further exploration, for proach developed by Xi et al. [8], in contrast, requires which the nuclear genome will be a particularly no a priori assumptions about evolutionary rates. In- valuable resource. stead, the optimal number of partitions, and their con- In addition to providing a wealth of new data for clari- tents, are identified using a Bayesian mixture model fying species trees, the nuclear genome will greatly im- analysis, which is not influenced by preconceptions prove our understanding of important innovations across about nucleotide evolution. This approach improved green plants. Whole genome duplications (WGDs), for resolution over analyses using traditional partitioning example, potentially enhance an organism’s success. Con- strategies and also reduced model complexity because sistent with this, recent analyses of transcriptomes from the optimal number of partitions identified in the seed plants indicate that at least three major WGDs search was smaller than in commonly used schemes. occurred very near to the origin of clades characte- Ruhfel et al. found this scheme to be computationally rized by putative key innovations [12,13]. These in- difficult to implement with their data, but improve- clude the origins of seeds, flowers, and pentamorous ments in the efficiency of Bayesian mixture model floral symmetry - the last of which characterizes more searches will help. Finally, despite the promise of these than approximately 70% of all angiosperms (the eudi- improvements to molecular phylogenetic studies, the cots) and may be related to their coevolution with evolution of green plants cannot be understood from bees [14]. molecular data alone. For example, 70% of seed plant Nuclear genomic data also more directly