Inferring the Total-Evidence Timescale of Marattialean Fern Evolution in the Face of Model Sensitivity

Bayesian Total-Evidence Dating Revisits Sloth Phylogeny and Biogeography: A Cautionary Tale on Morphological Clock Analyses

Combining morphological and molecular characters through Bayesian total-evidence dating allows inferring the phylogenetic and timescale framework of both extant and fossil taxa, while accounting for the stochasticity and incompleteness of the fossil record. Such an integrative approach is particularly needed when dealing with clades such as sloths (Mammalia: Folivora), for which developmental and biomechanical studies have shown high levels of morphological convergence whereas molecular data can only account for a limited percentage of their total species richness. Here, we propose an alternative hypothesis of sloth evolution that emphasizes the pervasiveness of morphological convergence and the importance of considering the fossil record and an adequate taxon sampling in both phylogenetic and biogeographic inferences. Regardless of different clock models and morphological datasets, the extant sloth Bradypus is consistently recovered as a megatherioid, and Choloepus as a mylodontoid, in agreement with molecular-only analyses. The recently extinct Caribbean sloths (Megalocnoidea) are found to be a monophyletic sister-clade of Megatherioidea, in contrast to previous phylogenetic hypotheses. Our results contradict previous morphological analyses and further support the polyphyly of "Megalonychidae," whose members were found in five different clades. Regardless of taxon sampling and clock models, the Caribbean colonization of sloths is compatible with the exhumation of islands along Aves Ridge and its geological time frame. Overall, our total-evidence analysis illustrates the difficulty of positioning highly incomplete fossils, although a robust phylogenetic framework was recovered by an a posteriori removal of taxa with high percentages of missing characters. Elimination of these taxa improved topological resolution by reducing polytomies and increasing node support. However, it introduced a systematic and geographic bias because most of these incomplete specimens are from northern South America. This is evident in biogeographic reconstructions, which suggest Patagonia as the area of origin of many clades when taxa are underrepresented, but Amazonia and/or Central and Southern Andes when all taxa are included. More generally, our analyses demonstrate the instability of topology and divergence time estimates when using different morphological datasets and clock models and thus caution against making macroevolutionary inferences when node support is weak or when uncertainties in the fossil record are not considered.

Skyline Fossilized Birth-Death Model is Robust to Violations of Sampling Assumptions in Total-Evidence Dating

Several total-evidence dating studies under the fossilized birth-death (FBD) model have produced very old age estimates, which are not supported by the fossil record. This phenomenon has been termed "deep root attraction (DRA)." For two specific data sets, involving divergence time estimation for the early radiations of ants, bees, and wasps (Hymenoptera) and of placental mammals (Eutheria), it has been shown that the DRA effect can be greatly reduced by accommodating the fact that extant species in these trees have been sampled to maximize diversity, so-called diversified sampling. Unfortunately, current methods to accommodate diversified sampling only consider the extreme case where it is possible to identify a cut-off time such that all splits occurring before this time are represented in the sampled tree but none of the younger splits. In reality, the sampling bias is rarely this extreme and may be difficult to model properly. Similar modeling challenges apply to the sampling of the fossil record. This raises the question of whether it is possible to find dating methods that are more robust to sampling biases. Here, we show that the skyline FBD (SFBD) process, where the diversification and fossil-sampling rates can vary over time in a piecewise fashion, provides age estimates that are more robust to inadequacies in the modeling of the sampling process and less sensitive to DRA effects. In the SFBD model we consider, rates in different time intervals are either considered to be independent and identically distributed or assumed to be autocorrelated following an Ornstein-Uhlenbeck (OU) process. Through simulations and reanalyses of Hymenoptera and Eutheria data, we show that both variants of the SFBD model unify age estimates under random and diversified sampling assumptions. The SFBD model can resolve DRA by absorbing the deviations from the sampling assumptions into the inferred dynamics of the diversification process over time. Although this means that the inferred diversification dynamics must be interpreted with caution, taking sampling biases into account, we conclude that the SFBD model represents the most robust approach currently available for addressing DRA in total-evidence dating.

Dating the evolution of the complex thalloid liverworts (Marchantiopsida): total-evidence dating analysis supports a Late Silurian-Early Devonian origin and post-Mesozoic morphological stasis

Divergence times based on molecular clock analyses often differ from those derived from total-evidence dating (TED) approaches. For bryophytes, fossils have been excluded from previous assessments of divergence times, and thus, their utility in dating analyses remains unexplored. Here, we conduct the first TED analyses of the complex thalloid liverworts (Marchantiopsida) that include fossils and evaluate macroevolutionary trends in morphological 'diversity' (disparity) and rates. Phylogenetic analyses were performed on a combined dataset of 130 discrete characters and 11 molecular markers (sampled from nuclear, plastid and mitochondrial genomes). Taxon sampling spanned 56 extant species - representing all the orders within Marchantiophyta and extant genera within Marchantiales - and eight fossil taxa. Total-evidence dating analyses support the radiation of Marchantiopsida during Late Silurian-Early Devonian (or Middle Ordovician when the outgroup is excluded) and that of Ricciaceae in the Middle Jurassic. Morphological change rate was high early in the history of the group, but it barely increased after Late Cretaceous. Disparity-through-time analyses support a fast increase in diversity until the Middle Triassic (c. 250 Ma), after which phenotypic evolution slows down considerably. Incorporating fossils in analyses challenges previous assumptions on the affinities of extinct taxa and indicates that complex thalloid liverworts radiated c. 125 Ma earlier than previously inferred.

Reconciling fossils with phylogenies reveals the origin and macroevolutionary processes explaining the global cycad biodiversity

The determinants of biodiversity patterns can be understood using macroevolutionary analyses. The integration of fossils into phylogenies offers a deeper understanding of processes underlying biodiversity patterns in deep time. Cycadales are considered a relict of a once more diverse and globally distributed group but are restricted to low latitudes today. We still know little about their origin and geographic range evolution. Combining molecular data for extant species and leaf morphological data for extant and fossil species, we study the origin of cycad global biodiversity patterns through Bayesian total-evidence dating analyses. We assess the ancestral geographic origin and trace the historical biogeography of cycads with a time-stratified process-based model. Cycads originated in the Carboniferous on the Laurasian landmass and expanded in Gondwana in the Jurassic. Through now-vanished continental connections, Antarctica and Greenland were crucial biogeographic crossroads for cycad biogeography. Vicariance is an essential speciation mode in the deep and recent past. Their latitudinal span increased in the Jurassic and restrained toward subtropical latitudes in the Neogene in line with biogeographic inferences of high-latitude extirpations. We show the benefits of integrating fossils into phylogenies to estimate ancestral areas of origin and to study evolutionary processes explaining the global distribution of present-day relict groups.

Integrating Fossil Flowers into the Angiosperm Phylogeny Using Molecular and Morphological Evidence

Fossils are essential to infer past evolutionary processes. The assignment of fossils to extant clades has traditionally relied on morphological similarity and on apomorphies shared with extant taxa. The use of explicit phylogenetic analyses to establish fossil affinities has so far remained limited. In this study, we built a comprehensive framework to investigate the phylogenetic placement of 24 exceptionally preserved fossil flowers. For this, we assembled a new species-level data set of 30 floral traits for 1201 extant species that were sampled to capture the stem and crown nodes of all angiosperm families. We explored multiple analytical approaches to integrate the fossils into the phylogeny, including different phylogenetic estimation methods, topological-constrained analyses, and combining molecular and morphological data of extant and fossil species. Our results were widely consistent across approaches and showed minor differences in the support of fossils at different phylogenetic positions. The placement of some fossils agrees with previously suggested relationships, but for others, a new placement is inferred. We also identified fossils that are well supported within particular extant families, whereas others showed high phylogenetic uncertainty. Finally, we present recommendations for future analyses combining molecular and morphological evidence, regarding the selection of fossils and appropriate methodologies, and provide some perspectives on how to integrate fossils into the investigation of divergence times and the temporal evolution of morphological traits. [Angiosperms; fossil flowers; phylogenetic uncertainty; RoguePlots.].

Molecular timetrees using relaxed clocks and uncertain phylogenies

A common practice in molecular systematics is to infer phylogeny and then scale it to time by using a relaxed clock method and calibrations. This sequential analysis practice ignores the effect of phylogenetic uncertainty on divergence time estimates and their confidence/credibility intervals. An alternative is to infer phylogeny and times jointly to incorporate phylogenetic errors into molecular dating. We compared the performance of these two alternatives in reconstructing evolutionary timetrees using computer-simulated and empirical datasets. We found sequential and joint analyses to produce similar divergence times and phylogenetic relationships, except for some nodes in particular cases. The joint inference performed better when the phylogeny was not well resolved, situations in which the joint inference should be preferred. However, joint inference can be infeasible for large datasets because available Bayesian methods are computationally burdensome. We present an alternative approach for joint inference that combines the bag of little bootstraps, maximum likelihood, and RelTime approaches for simultaneously inferring evolutionary relationships, divergence times, and confidence intervals, incorporating phylogeny uncertainty. The new method alleviates the high computational burden imposed by Bayesian methods while achieving a similar result.

Diversification Models Conflate Likelihood and Prior, and Cannot be Compared Using Conventional Model-Comparison Tools

Time-calibrated phylogenetic trees are a tremendously powerful tool for studying evolutionary, ecological, and epidemiological phenomena. Such trees are predominantly inferred in a Bayesian framework, with the phylogeny itself treated as a parameter with a prior distribution (a "tree prior"). However, we show that the tree "parameter" consists, in part, of data, in the form of taxon samples. Treating the tree as a parameter fails to account for these data and compromises our ability to compare among models using standard techniques (e.g., marginal likelihoods estimated using path-sampling and stepping-stone sampling algorithms). Since accuracy of the inferred phylogeny strongly depends on how well the tree prior approximates the true diversification process that gave rise to the tree, the inability to accurately compare competing tree priors has broad implications for applications based on time-calibrated trees. We outline potential remedies to this problem, and provide guidance for researchers interested in assessing the fit of tree models. [Bayes factors; Bayesian model comparison; birth-death models; divergence-time estimation; lineage diversification].

Transcriptome-Based Study on the Phylogeny and Hybridization of Marattialean Ferns (Marattiaceae)

Marattiaceae is a phylogenetically isolated family of tropical eusporangiate ferns including six genera with more than one-hundred species. In Marattiaceae, monophyly of genera has been well-supported phylogenetically. However, the phylogenetic relationships among them were elusive and controversial. Here, a dataset of 26 transcriptomes (including 11 newly generated) were used to assess single-copy nuclear genes and to obtain the organelle gene sequences. Through phylotranscriptomic analysis, the phylogeny and hybridization events of Marattiaceae were explored and a robust phylogenomic framework for the evolution of Marattiaceae was provided. Using both concatenation- and coalescent-based phylogenies, the gene-tree discordance, incomplete lineage sorting (ILS) simulations, and network inference were examined. Except the low support with mitochondrial genes of Marattiaceae, nuclear genes and chloroplast genes strongly supported a sister relationship between Marattiaceae and leptosporangiate ferns. At the genus level, all phylogenetic analysis based on nuclear genes datasets recovered five genera in Marattiaceae as monophyletic with strong support. Danaea and Ptisana were the first two diverged clades in turn. Christensenia was a sister clade to the clade Marattia + Angiopteris s.l. In Angiopteris s.l., three clades (Angiopteris s.s., the Archangiopteris group, and An. sparsisora) were well identified with maximum support. The Archangiopteris group was derived from Angiopteris s.s. at ca. 18 Ma. The putative hybrid species An. sparsisora between Angiopteris s.s. and the Archangiopteris group was verified by the species network analyses and the maternal plastid genes. This study will improve our understanding for using the phylotranscriptomic method to explore phylogeny and investigate hybridization events for difficult taxa in ferns.

Fossil berries reveal global radiation of the nightshade family by the early Cenozoic

Fossil discoveries can transform our understanding of plant diversification over time and space. Recently described fossils in many plant families have pushed their known records farther back in time, pointing to alternative scenarios for their origin and spread. Here, we describe two new Eocene fossil berries of the nightshade family (Solanaceae) from the Esmeraldas Formation in Colombia and the Green River Formation in Colorado (USA). The placement of the fossils was assessed using clustering and parsimony analyses based on 10 discrete and five continuous characters, which were also scored in 291 extant taxa. The Colombian fossil grouped with members of the tomatillo subtribe, and the Coloradan fossil aligned with the chili pepper tribe. Along with two previously reported early Eocene fossils from the tomatillo genus, these findings indicate that Solanaceae were distributed at least from southern South America to northwestern North America by the early Eocene. Together with two other recently discovered Eocene berries, these fossils demonstrate that the diverse berry clade and, in turn, the entire nightshade family, is much older and was much more widespread in the past than previously thought.

Complex macroevolution of pterosaurs

Pterosaurs, the earliest flying tetrapods, are the subject of some recent quantitative macroevolutionary analyses from different perspectives.^1-2 Here, we use an integrative approach involving newly assembled phylogenetic and body size datasets, net diversification rates, morphological rates, and morphological disparity to gain a holistic understanding of the pterosaur macroevolution. The first two parameters are important in quantitative analyses of macroevolution, but they have been rarely used in previous pterosaur studies.^{1,3,4,2,5,6,7,8,9,10,11,12} Our study reveals an ∼115-Ma period-from Early Triassic to Early Cretaceous-of multi-wave increasing net diversification rates and disparity, as well as high morphological rates, followed by an ∼65-Ma period-from Early Cretaceous to the end of the Cretaceous-of mostly negative net diversification rates, decreasing disparity, and relatively low morphological rates in pterosaur evolution. Our study demonstrates the following: (1) body size plays an important role in pterosaur lineage diversification during nearly their whole evolutionary history, and the evolution of locomotion, trophic, and ornamental structures also plays a role in different periods; (2) birds, the other major flying tetrapod group at the time, might have affected pterosaur macroevolution for ∼100 Ma; and (3) different mass extinction events might have affected pterosaur evolution differently. Particularly, the revealed decline in pterosaur biodiversity during the Middle and Late Cretaceous periods provides further support for the possible presence of a biodiversity decline of large-sized terrestrial amniotes starting in the mid-Cretaceous,^13,14 which may have been caused by multiple factors including a global land area decrease during these periods.

Illusion of flight? Absence, evidence and the age of winged insects

The earliest fossils of winged insects (Pterygota) are mid-Carboniferous (latest Mississippian, 328–324 Mya), but estimates of their age based on fossil-calibrated molecular phylogenetic studies place their origin at 440–370 Mya during the Silurian or Devonian. This discrepancy would require that winged insects evaded fossilization for at least the first ~50 Myr of their history. Here, we examine the plausibility of such a gap in the fossil record, and possible explanations for it, based on comparisons with the fossil records of other arthropod groups, the distribution of first occurrence dates of pterygote families, phylogenetically informed simulations of the fossilization of Palaeozoic insects, and re-analysis of data presented by Misof and colleagues using updated fossil calibrations under a variety of prior probability settings. We do not find support for the mechanisms previously suggested to account for such an extended gap in the pterygote fossil record, including sampling bias, preservation bias, and body size. We suggest that inference of an early origin of Pterygota long prior to their first appearance in the fossil record is probably an analytical artefact of taxon sampling and choice of fossil calibration points, possibly compounded by heterogeneity in rates of sequence evolution or speciation, including radiations or ‘bursts’ during their early history.

Cutting the long branches: Consilience as a path to unearth the evolutionary history of Gnetales

The Gnetales are one of the most fascinating groups within seed plants. Although the advent of molecular phylogenetics has generated some confidence in their phylogenetic placement of Gnetales within seed plants, their macroevolutionary history still presents many unknowns. Here, we review the reasons for such unknowns, and we focus the discussion on the presence of “long branches” both in their molecular and morphological history. The increased rate of molecular evolution and genome instability as well as the numerous unique traits (both reproductive and vegetative) in the Gnetales have been obstacles to a better understanding of their evolution. Moreover, the fossil record of the Gnetales, though relatively rich, has not yet been properly reviewed and investigated using a phylogenetic framework. Despite these apparent blocks to progress we identify new avenues to enable us to move forward. We suggest that a consilience approach, involving different disciplines such as developmental genetics, paleobotany, molecular phylogenetics, and traditional anatomy and morphology might help to “break” these long branches, leading to a deeper understanding of this mysterious group of plants.

What is the age of flowering plants?

The origin of flowering plants (angiosperms) was one of the most transformative events in the history of our planet. Despite considerable interest from multiple research fields, numerous questions remain, including the age of the group as a whole. Recent studies have reported a perplexing range of estimates for the crown-group age of angiosperms, from ~140 million years (Ma; Early Cretaceous) to 270 Ma (Permian). Both ends of the spectrum are now supported by both macroevolutionary analyses of the fossil record and fossil-calibrated molecular dating analyses. Here, we first clarify and distinguish among the three ages of angiosperms: the age of their divergence with acrogymnosperms (stem age); the age(s) of emergence of their unique, distinctive features including flowers (morphological age); and the age of the most recent common ancestor of all their living species (crown age). We then demonstrate, based on recent studies, that fossil-calibrated molecular dating estimates of the crown-group age of angiosperms have little to do with either the amount of molecular data or the number of internal fossil calibrations included. Instead, we argue that this age is almost entirely conditioned by its own prior distribution (typically a calibration density set by the user in Bayesian analyses). Lastly, we discuss which future discoveries or novel types of analyses are most likely to bring more definitive answers. In the meantime, we propose that the age of angiosperms is best described as largely unknown (140-270 Ma) and that contrasting age estimates in the literature mostly reflect conflicting prior distributions. We also suggest that future work that depends on the time scale of flowering plant diversification be designed to integrate over this vexing uncertainty.

Studying speciation and extinction dynamics from phylogenies: addressing identifiability issues

A lot of what we know about past speciation and extinction dynamics is based on statistically fitting birth-death processes to phylogenies of extant species. Despite their wide use, the reliability of these tools is regularly questioned. It was recently demonstrated that vast 'congruent' sets of alternative diversification histories cannot be distinguished (i.e., are not identifiable) using extant phylogenies alone, reanimating the debate about the limits of phylogenetic diversification analysis. Here, we summarize what we know about the identifiability of the birth-death process and how identifiability issues can be addressed. We conclude that extant phylogenies, when combined with appropriate prior hypotheses and regularization techniques, can still tell us a lot about past diversification dynamics.

Early cephalopod evolution clarified through Bayesian phylogenetic inference

BACKGROUND

Despite the excellent fossil record of cephalopods, their early evolution is poorly understood. Different, partly incompatible phylogenetic hypotheses have been proposed in the past, which reflected individual author's opinions on the importance of certain characters but were not based on thorough cladistic analyses. At the same time, methods of phylogenetic inference have undergone substantial improvements. For fossil datasets, which typically only include morphological data, Bayesian inference and in particular the introduction of the fossilized birth-death model have opened new possibilities. Nevertheless, many tree topologies recovered from these new methods reflect large uncertainties, which have led to discussions on how to best summarize the information contained in the posterior set of trees.

RESULTS

We present a large, newly compiled morphological character matrix of Cambrian and Ordovician cephalopods to conduct a comprehensive phylogenetic analysis and resolve existing controversies. Our results recover three major monophyletic groups, which correspond to the previously recognized Endoceratoidea, Multiceratoidea, and Orthoceratoidea, though comprising slightly different taxa. In addition, many Cambrian and Early Ordovician representatives of the Ellesmerocerida and Plectronocerida were recovered near the root. The Ellesmerocerida is para- and polyphyletic, with some of its members recovered among the Multiceratoidea and early Endoceratoidea. These relationships are robust against modifications of the dataset. While our trees initially seem to reflect large uncertainties, these are mainly a consequence of the way clade support is measured. We show that clade posterior probabilities and tree similarity metrics often underestimate congruence between trees, especially if wildcard taxa are involved.

CONCLUSIONS

Our results provide important insights into the earliest evolution of cephalopods and clarify evolutionary pathways. We provide a classification scheme that is based on a robust phylogenetic analysis. Moreover, we provide some general insights on the application of Bayesian phylogenetic inference on morphological datasets. We support earlier findings that quartet similarity metrics should be preferred over the Robinson-Foulds distance when higher-level phylogenetic relationships are of interest and propose that using a posteriori pruned maximum clade credibility trees help in assessing support for phylogenetic relationships among a set of relevant taxa, because they provide clade support values that better reflect the phylogenetic signal.