Angiosperms represent one of the most spectacular terrestrial radiations on the planet1, but their early diversification and phylogenetic relationships remain uncertain2–5. A key reason for this impasse is the paucity of complete genomes representing early-diverging angiosperms. Here, we present high-quality, chromosomal-level genome assemblies of two aquatic species—prickly waterlily (Euryale ferox; Nymphaeales) and the rigid hornwort (Ceratophyllum demersum; Ceratophyllales)—and expand the genomic representation for key sectors of the angiosperm tree of life. We identify multiple independent polyploidization events in each of the five major clades (that is, Nymphaeales, magnoliids, monocots, Ceratophyllales and eudicots). Furthermore, our phylogenomic analyses, which spanned multiple datasets and diverse methods, confirm that Amborella and Nymphaeales are successively sister to all other angiosperms. Furthermore, these genomes help to elucidate relationships among the major subclades within Mesangiospermae, which contain about 350,000 species. In particular, the species-poor lineage Ceratophyllales is supported as sister to eudicots, and monocots and magnoliids are placed as successively sister to Ceratophyllales and eudicots. Finally, our analyses indicate that incomplete lineage sorting may account for the incongruent phylogenetic placement of magnoliids between nuclear and plastid genomes.

An increasing number of species are thought to have originated by homoploid hybrid speciation (HHS), but in only a handful of cases are details of the process known. A previous study indicated that Picea purpurea, a conifer in the Qinghai–Tibet Plateau (QTP), originated through HHS from P. likiangensis and P. wilsonii. To investigate this origin in more detail, we analysed transcriptome data for 114 individuals collected from 34 populations of the three Picea species from their core distributions in the QTP. Phylogenetic, principal component and admixture analyses of nuclear SNPs showed the species to be delimited genetically and that P. purpurea was admixed with approximately 60% of its ancestry derived from P. wilsonii and 40% from P. likiangensis. Coalescent simulations revealed the best‐fitting model of origin involved formation of an intermediate hybrid lineage between P. likiangensis and P. wilsonii approximately 6 million years ago (mya), which backcrossed to P. wilsonii to form P. purpurea approximately one mya. The intermediate hybrid lineage no longer exists and is referred to as a “ghost” lineage. Our study emphasizes the power of population genomic analysis combined with coalescent analysis for reconstructing the stages involved in the origin of a homoploid hybrid species over an extended period. In contrast to other studies, we show that these stages can in some instances span a relatively long period of evolutionary time.

Having a comprehensive understanding of population structure, genetic differentiation and demographic history is important for the conservation and management of threatened species. High‐throughput sequencing (HTS) provides exciting opportunities to address a wide range of factors for conservation genetics. Here, we generated HTS data and identified 266,884 high‐quality single nucleotide polymorphisms from 82 individuals of Cupressus chengiana, to assess population genomics across the species' full range, comprising the Daduhe River (DDH), Minjiang River (MJR) and Bailongjiang River (BLJ) catchments in western China. admixture, principal components analysis and phylogenetic analyses indicated that each region contains a distinct lineage, with high levels of differentiation between them (DDH, MJR and BLJ lineages). MJR was newly distinguished compared to previous surveys, and evidence including coalescent simulations supported a hybrid origin of MJR during the Quaternary. Each of these three lineages should be recognized as an evolutionarily significant unit (ESU), due to isolation, differing genetic adaptations and different demographic history. Currently, each ESU faces distinct threats, and will require different conservation strategies. Our work shows that population genomic approaches using HTS can reconstruct the complex evolutionary history of threatened species in mountainous regions, and hence inform conservation efforts, and contribute to the understanding of high biodiversity in mountains.

Hybridization and introgression are believed to play important roles in plant evolution. However, few empirical studies have been designed to clarify the ways in which these processes complicate taxonomic delimitation. Recent phylogenetic studies based on a number of different DNA fragments have indicated that Picea brachytyla in the eastern Qinghai-Tibet Plateau is polyphyletic, a finding that contrasts with traditional taxonomy based on morphological traits. We aimed to test this conflict using transcriptomic data from 26 trees collected from multiple localities for this and related species. Our phylogenomic analyses suggest that the sampled trees of P. brachytyla cluster into two distinct lineages corresponding to the two taxonomically recognized intraspecific varieties: var. brachytyla and var. complanata. However, var. complanata nested within Picea likiangensis and was sister to one of its three varieties, while var. brachytyla comprised an isolated lineage. The polyphyletic origin hypothesis was further supported by likelihood tree comparisons using Akaike's information criterion (AIC) and by coalescent analyses under the snapp model. However, our abba-baba and ∂a∂i analyses suggest that gene flow between these two independently evolved lineages has been extensive and bidirectional. Introgression, as well as parallel evolution in the arid habitats common to both lineages, may have given rise to their morphological similarity. Our study highlights the importance of genomic evidence and the use of newly developed coalescent analysis methods for clarifying the evolutionary complexity of certain plant taxa.

Nonbifurcating divergence caused by introgressive hybridization is continuously reported for groups of closely related species. In this study, we aimed to reconstruct the genome-scale classification of deep lineages of the conifer genus Picea, establish their phylogenetic relationships and test the bifurcating hypothesis between deeply branching lineages based on genomic data.We sequenced the transcriptomes of 35 individuals of 27 taxa covering all main lineages of the genus. Four major lineages, comprising three to 12 taxa each, largely consistent with morphological evidence, were recovered across the coalescent and integrated nuclear phylogeny. However, many of the individual gene trees recovered contradict one another. Moreover, the well-supported coalescent tree inferred here differs from previous studies based on various DNA markers, with respect to topology and inter-lineage relationships.We identified the shared polymorphisms between four major lineages. ABBA-BABA tests confirmed the inter-lineage gene flow and thus violated the bifurcating divergence model. Gene flow occurred more frequently between lineages distributed in the same continent than those disjunct between continents.Our results indicate that introgression and nonbifurcating diversification apply, even between deeply branching lineages of the conifer genus Picea.