Lack of resolution in the animal phylogeny: closely spaced cladogeneses or undetected systematic errors?

A recent phylogenomic study reported that the animal phylogeny was unresolved despite the use of 50 genes. This lack of resolution was interpreted as "a positive signature of closely spaced cladogenetic events." Here, we propose that this lack of resolution is rather due to the mutual cancellation of the phylogenetic signal (historical) and the nonphylogenetic signal (due to systematic errors) that results from inadequate taxon sampling and/or model of sequence evolution. Starting with a data set of comparable size, we use 3 different strategies to reduce the nonphylogenetic signal: 1) increasing the number of species; 2) replacing a fast-evolving species by a slowly evolving one; and 3) using a better model of sequence evolution. In all cases, the phylogenetic resolution is markedly improved, in agreement with our hypothesis that the originally reported lack of resolution was artifactual.     

An empirical assessment of long-branch attraction artefacts in deep eukaryotic phylogenomics

In the context of exponential growing molecular databases, it becomes increasingly easy to assemble large multigene data sets for phylogenomic studies. The expected increase of resolution due to the reduction of the sampling (stochastic) error is becoming a reality. However, the impact of systematic biases will also become more apparent or even dominant. We have chosen to study the case of the long-branch attraction artefact (LBA) using real instead of simulated sequences. Two fast-evolving eukaryotic lineages, whose evolutionary positions are well established, microsporidia and the nucleomorph of cryptophytes, were chosen as model species. A large data set was assembled (44 species, 133 genes, and 24,294 amino acid positions) and the resulting rooted eukaryotic phylogeny (using a distant archaeal outgroup) is positively misled by an LBA artefact despite the use of a maximum likelihood-based tree reconstruction method with a complex model of sequence evolution. When the fastest evolving proteins from the fast lineages are progressively removed (up to 90%), the bootstrap support for the apparently artefactual basal placement decreases to virtually 0%, and conversely only the expected placement, among all the possible locations of the fast-evolving species, receives increasing support that eventually converges to 100%. The percentage of removal of the fastest evolving proteins constitutes a reliable estimate of the sensitivity of phylogenetic inference to LBA. This protocol confirms that both a rich species sampling (especially the presence of a species that is closely related to the fast-evolving lineage) and a probabilistic method with a complex model are important to overcome the LBA artefact. Finally, we observed that phylogenetic inference methods perform strikingly better with simulated as opposed to real data, and suggest that testing the reliability of phylogenetic inference methods with simulated data leads to overconfidence in their performance. Although phylogenomic studies can be affected by systematic biases, the possibility of discarding a large amount of data containing most of the nonphylogenetic signal allows recovering a phylogeny that is less affected by systematic biases, while maintaining a high statistical support.     

Automated Masking of AFLP Markers Improves Reliability of Phylogenetic Analyses

The amplified fragment length polymorphisms (AFLP) method has become an attractive tool in phylogenetics due to the ease with which large numbers of characters can be generated. In contrast to sequence-based phylogenetic approaches, AFLP data consist of anonymous multilocus markers. However, potential artificial amplifications or amplification failures of fragments contained in the AFLP data set will reduce AFLP reliability especially in phylogenetic inferences. In the present study, we introduce a new automated scoring approach, called “AMARE” (AFLP MAtrix REduction). The approach is based on replicates and makes marker selection dependent on marker reproducibility to control for scoring errors. To demonstrate the effectiveness of our approach we record error rate estimations, resolution scores, PCoA and stemminess calculations. As in general the true tree (i.e. the species phylogeny) is not known, we tested AMARE with empirical, already published AFLP data sets, and compared tree topologies of different AMARE generated character matrices to existing phylogenetic trees and/or other independent sources such as morphological and geographical data. It turns out that the selection of masked character matrices with highest resolution scores gave similar or even better phylogenetic results than the original AFLP data sets.     

On the importance of homology in the age of phylogenomics

Homology is perhaps the most central concept of phylogenetic biology. Molecular systematists have traditionally paid due attention to the homology statements that are implied by their alignments of orthologous sequences, but some authors have suggested that manual gene-by-gene curation is not sustainable in the phylogenomics era. Here, we show that there are multiple ways to efficiently screen for and detect homology errors in phylogenomic data sets. Application of these screening approaches to two phylogenomic data sets, one for birds and another for mammals, shows that these data are replete with homology errors including alignments of different exons to each other, alignments of exons to introns, and alignments of paralogues to each other. The extent of these homology errors weakens the conclusions of studies based on these data sets. Despite advances in automated phylogenomic pipelines, we contend that much of the long, difficult, and sometimes tedious work of systematics is still required to guard against pervasive homology errors. This conclusion is underscored by recent studies that show that just a few outlier genes can impact phylogenetic results at short, tightly spaced internodes that are deep in the Tree of Life. The view that widespread DNA sequence alignment errors are not a major concern for rigorous systematic research is not tenable. If a primary goal of phylogenomics is to resolve the most challenging phylogenetic problems with the abundant data that are now available, researchers must employ effective procedures to screen for and correct homology errors prior to performing downstream phylogenetic analyses.     

Reconstructing the phylogeny of "Buarremon" brush-finches and near relatives (Aves, Emberizidae) from individual gene trees

Gene trees are often assumed to be equivalent to species trees, but processes such as incomplete lineage sorting can generate incongruence among gene topologies and analyzing multilocus data in concatenated matrices can be prone to systematic errors. Accordingly, a variety of new methods have been developed to estimate species trees using multilocus data sets. Here, we apply some of these methods to reconstruct the phylogeny of Buarremon and near relatives, a group in which phylogenetic analyses of mitochondrial DNA sequences produced results that were inconsistent with relationships implied by a taxonomy based on variation in external phenotype. Gene genealogies obtained for seven loci (one mitochondrial, six nuclear) were varied, with some supporting and some rejecting the monophyly of Buarremon. Overall, our species-tree analyses tended to support a monophyletic Buarremon, but due to lack of congruence between methodologies, resolution of the phylogeny of this group remains uncertain. More generally, our study indicates that the number of individuals sampled can have an important effect on phylogenetic reconstruction, that the use of seven markers does not guarantee obtaining a strongly-supported species tree, and that methods for species-tree reconstruction can produce different results using the same data; these are important considerations for researchers using these new phylogenetic approaches in other systems.     

Reproductive allometry and the size-number trade-off for lizards

Fundamental to life-history theory is the assumed inverse proportionality between the number of offspring and the resource allocation per offspring. Lizards have been model organisms for empirical tests of this theory for decades; however, the expected negative relationship between clutch size and offspring size is often not detected. Here we use the approach developed by Charnov and Ernest to demonstrate that this often concealed trade-off can be made apparent in an interspecific comparison by correcting for size-dependent resource allocation. Our data set also shows a tight allometry for annual production that is consistent with life-history models for indeterminate growers. To account for nonindependence of species data we also compare the fit of nonphylogenetic and phylogenetic regression models to test for phylogenetic signal in these allometry and trade-off patterns. When combined, these results demonstrate that the offspring size/clutch size trade-off is not isolated to a single clutch but is shaped by the resource investment made over an entire year. We conclude that, across diverse lizard species, there is strong evidence for the predicted trade-off between offspring size and the annual number of eggs produced.     

Treeness triangles: visualizing the loss of phylogenetic signal

It is well known that molecular data "saturates" with increasing sequence divergence (thereby losing phylogenetic information) and that in addition the accumulation of misleading information due to chance similarities or to systematic bias may accompany saturation as well. Exploratory data analysis methods that can quantify the extent of signal loss or convergence for a given data set are scarce. Such methods are needed because genomics delivers very long sequence alignments spanning substantial phylogenetic depth, where site saturation may be compounded by systematic biases or other alternative signals. Here we introduce the Treeness Triangle (TT) graph, in which signals detectable by Hadamard (spectral) analysis are summed into 3 categories--those supporting 1) external and 2) internal branches in the optimal tree, in addition to 3) the residuals (potential internal branches not present in the optimal tree). These 3 values are plotted in a standard ternary coordinate system. The approach is illustrated with simulated and real data sets, the latter from complete chloroplast genomes, where potential problems of paralogy or lateral gene acquisition can be excluded. The TT uncovers the divergence-dependent loss of phylogenetic signal as subsets of chloroplast genomes are investigated that span increasingly deeper evolutionary timescales. The rate of signal loss (or signal retention) varies with the gene and/or the method of analysis.     

Phylogenetic analysis, structural evolution and functional divergence of the 12-oxo-phytodienoate acid reductase gene family in plants

Background

Whenever different data sets arrive at conflicting phylogenetic hypotheses, only testable causal explanations of sources of errors in at least one of the data sets allow us to critically choose among the conflicting hypotheses of relationships. The large (28S) and small (18S) subunit rRNAs are among the most popular markers for studies of deep phylogenies. However, some nodes supported by this data are suspected of being artifacts caused by peculiarities of the evolution of these molecules. Arthropod phylogeny is an especially controversial subject dotted with conflicting hypotheses which are dependent on data set and method of reconstruction. We assume that phylogenetic analyses based on these genes can be improved further i) by enlarging the taxon sample and ii) employing more realistic models of sequence evolution incorporating non-stationary substitution processes and iii) considering covariation and pairing of sites in rRNA-genes.

Results

We analyzed a large set of arthropod sequences, applied new tools for quality control of data prior to tree reconstruction, and increased the biological realism of substitution models. Although the split-decomposition network indicated a high noise content in the data set, our measures were able to both improve the analyses and give causal explanations for some incongruities mentioned from analyses of rRNA sequences. However, misleading effects did not completely disappear.

Conclusion

Analyses of data sets that result in ambiguous phylogenetic hypotheses demand for methods, which do not only filter stochastic noise, but likewise allow to differentiate phylogenetic signal from systematic biases. Such methods can only rely on our findings regarding the evolution of the analyzed data. Analyses on independent data sets then are crucial to test the plausibility of the results. Our approach can easily be extended to genomic data, as well, whereby layers of quality assessment are set up applicable to phylogenetic reconstructions in general.     

Confirming the phylogeny of mammals by use of large comparative sequence data sets

The ongoing generation of prodigious amounts of genomic sequence data from myriad vertebrates is providing unparalleled opportunities for establishing definitive phylogenetic relationships among species. The size and complexities of such comparative sequence data sets not only allow smaller and more difficult branches to be resolved but also present unique challenges, including large computational requirements and the negative consequences of systematic biases. To explore these issues and to clarify the phylogenetic relationships among mammals, we have analyzed a large data set of over 60 megabase pairs (Mb) of high-quality genomic sequence, which we generated from 41 mammals and 3 other vertebrates. All sequences are orthologous to a 1.9-Mb region of the human genome that encompasses the cystic fibrosis transmembrane conductance regulator gene (CFTR). To understand the characteristics and challenges associated with phylogenetic analyses of such a large data set, we partitioned the sequence data in several ways and utilized maximum likelihood, maximum parsimony, and Neighbor-Joining algorithms, implemented in parallel on Linux clusters. These studies yielded well-supported phylogenetic trees, largely confirming other recent molecular phylogenetic analyses. Our results provide support for rooting the placental mammal tree between Atlantogenata (Xenarthra and Afrotheria) and Boreoeutheria (Euarchontoglires and Laurasiatheria), illustrate the difficulty in resolving some branches even with large amounts of data (e.g., in the case of Laurasiatheria), and demonstrate the valuable role that very large comparative sequence data sets can play in refining our understanding of the evolutionary relationships of vertebrates.     

Inferring and testing hypotheses of cladistic character dependence by using character compatibility

The notion that two characters evolve independently is of interest for two reasons. First, theories of biological integration often predict that change in one character requires complementary change in another. Second, character independence is a basic assumption of most phylogenetic inference methods, and dependent characters might confound attempts at phylogenetic inference. Previously proposed tests of correlated character evolution require a model phylogeny and therefore assume that nonphylogenetic correlation has a negligible effect on initial tree construction. This paper develops "tree-free" methods for testing the independence of cladistic characters. These methods can test the character independence model as a hypothesis before phylogeny reconstruction, or can be used simply to test for correlated evolution. We first develop an approach for visualizing suites of correlated characters by using character compatibility. Two characters are compatible if they can be used to construct a tree without homoplasy. The approach is based on the examination of mutual compatibilities between characters. The number of times two characters i and j share compatibility with a third character is calculated, and a pairwise shared compatibility matrix is constructed. From this matrix, an association matrix analogous to a dissimilarity matrix is derived. Eigenvector analyses of this association matrix reveal suites of characters with similar compatibility patterns. A priori character subsets can be tested for significant correlation on these axes. Monte Carlo tests are performed to determine the expected distribution of mutual compatibilities, given various criteria from the original data set. These simulated distributions are then used to test whether the observed amounts of nonphylogenetic correlation in character suites can be attributed to chance alone. We have applied these methods to published morphological data for caecilian amphibians. The analyses corroborate instances of dependent evolution hypothesized by previous workers and also identify novel partitions. Phylogenetic analysis is performed after reducing correlated suites to single characters. The resulting cladogram has greater topological resolution and implies appreciably less change among the remaining characters than does a tree derived from the raw data matrix.     

Directional biases in phylogenetic structure quantification: a Mediterranean case study

Recent years have seen an increasing effort to incorporate phylogenetic hypotheses to the study of community assembly processes. The incorporation of such evolutionary information has been eased by the emergence of specialized software for the automatic estimation of partially resolved supertrees based on published phylogenies. Despite this growing interest in the use of phylogenies in ecological research, very few studies have attempted to quantify the potential biases related to the use of partially resolved phylogenies and to branch length accuracy, and no work has examined how tree shape may affect inference of community phylogenetic metrics. In this study, we tested the influence of phylogenetic resolution and branch length information on the quantification of phylogenetic structure, and also explored the impact of tree shape (stemminess) on the loss of accuracy in phylogenetic structure quantification due to phylogenetic resolution. For this purpose, we used 9 sets of phylogenetic hypotheses of varying resolution and branch lengths to calculate three indices of phylogenetic structure: the mean phylogenetic distance (NRI), the mean nearest taxon distance (NTI) and phylogenetic diversity (stdPD) metrics. The NRI metric was the less sensitive to phylogenetic resolution, stdPD showed an intermediate sensitivity, and NTI was the most sensitive one; NRI was also less sensitive to branch length accuracy than NTI and stdPD, the degree of sensitivity being strongly dependent on the dating method and the sample size. Directional biases were generally towards type II errors. Interestingly, we detected that tree shape influenced the accuracy loss derived from the lack of phylogenetic resolution, particularly for NRI and stdPD. We conclude that well‐resolved molecular phylogenies with accurate branch length information are needed to identify the underlying phylogenetic structure of communities, and also that sensitivity of phylogenetic structure measures to low phylogenetic resolution can strongly vary depending on phylogenetic tree shape.     

Phylogenetic signal common to three data sets: Combining data which initially appear heterogeneous

This study makes use of three sources of data, morphology and two chloroplast DNA sequences,ndhF andrbcL, to resolve relationships in Gesneriaceae. Cladograms from each of the three data sets separately are not topologically congruent. Statistical indices suggest that each data set is congruent with thendhF data althoughrbcL and morphology are themselves incongruent. Consensus methods provide no resolution of taxonomic relationships when trees from the different data sets are combined. Combining data sets generally results in cladograms that are more fully resolved than each of the data sets analyzed separately and support for the clades increases based on higher decay index and bootstrap values. These results indicate that there is a phylogenetic signal common to each of the data sets, however, the noise (errors due to homoplasy, mis-scoring, etc.) unique to each data source masks this signal. In combining the data, the evidence for the common evolutionary history in each data set overcomes the noise and is apparent in the resulting trees.     

Phylogenetic relationships within the Opisthokonta based on phylogenomic analyses of conserved single-copy protein domains

Many of the eukaryotic phylogenomic analyses published to date were based on alignments of hundreds to thousands of genes. Frequently, in such analyses, the most realistic evolutionary models currently available are often used to minimize the impact of systematic error. However, controversy remains over whether or not idiosyncratic gene family dynamics (i.e., gene duplications and losses) and incorrect orthology assignments are always appropriately taken into account. In this paper, we present an innovative strategy for overcoming orthology assignment problems. Rather than identifying and eliminating genes with paralogy problems, we have constructed a data set comprised exclusively of conserved single-copy protein domains that, unlike most of the commonly used phylogenomic data sets, should be less confounded by orthology miss-assignments. To evaluate the power of this approach, we performed maximum likelihood and Bayesian analyses to infer the evolutionary relationships within the opisthokonts (which includes Metazoa, Fungi, and related unicellular lineages). We used this approach to test 1) whether Filasterea and Ichthyosporea form a clade, 2) the interrelationships of early-branching metazoans, and 3) the relationships among early-branching fungi. We also assessed the impact of some methods that are known to minimize systematic error, including reducing the distance between the outgroup and ingroup taxa or using the CAT evolutionary model. Overall, our analyses support the Filozoa hypothesis in which Ichthyosporea are the first holozoan lineage to emerge followed by Filasterea, Choanoflagellata, and Metazoa. Blastocladiomycota appears as a lineage separate from Chytridiomycota, although this result is not strongly supported. These results represent independent tests of previous phylogenetic hypotheses, highlighting the importance of sophisticated approaches for orthology assignment in phylogenomic analyses.     

Phylogenetic networks: modeling, reconstructibility, and accuracy

Phylogenetic networks model the evolutionary history of sets of organisms when events such as hybrid speciation and horizontal gene transfer occur. In spite of their widely acknowledged importance in evolutionary biology, phylogenetic networks have so far been studied mostly for specific data sets. We present a general definition of phylogenetic networks in terms of directed acyclic graphs (DAGs) and a set of conditions. Further, we distinguish between model networks and reconstructible ones and characterize the effect of extinction and taxon sampling on the reconstructibility of the network. Simulation studies are a standard technique for assessing the performance of phylogenetic methods. A main step in such studies entails quantifying the topological error between the model and inferred phylogenies. While many measures of tree topological accuracy have been proposed, none exist for phylogenetic networks. Previously, we proposed the first such measure, which applied only to a restricted class of networks. In this paper, we extend that measure to apply to all networks, and prove that it is a metric on the space of phylogenetic networks. Our results allow for the systematic study of existing network methods, and for the design of new accurate ones.     

Phylogenetic signal and noise: predicting the power of a data set to resolve phylogeny

A principal objective for phylogenetic experimental design is to predict the power of a data set to resolve nodes in a phylogenetic tree. However, proactively assessing the potential for phylogenetic noise compared with signal in a candidate data set has been a formidable challenge. Understanding the impact of collection of additional sequence data to resolve recalcitrant internodes at diverse historical times will facilitate increasingly accurate and cost-effective phylogenetic research. Here, we derive theory based on the fundamental unit of the phylogenetic tree, the quartet, that applies estimates of the state space and the rates of evolution of characters in a data set to predict phylogenetic signal and phylogenetic noise and therefore to predict the power to resolve internodes. We develop and implement a Monte Carlo approach to estimating power to resolve as well as deriving a nearly equivalent faster deterministic calculation. These approaches are applied to describe the distribution of potential signal, polytomy, or noise for two example data sets, one recent (cytochrome c oxidase I and 28S ribosomal rRNA sequences from Diplazontinae parasitoid wasps) and one deep (eight nuclear genes and a phylogenomic sequence for diverse microbial eukaryotes including Stramenopiles, Alveolata, and Rhizaria). The predicted power of resolution for the loci analyzed is consistent with the historic use of the genes in phylogenetics.     

A phylogeny of the Platyhelminthes: towards a total-evidence solution

We advocate a total-evidence approach for the reconstruction of working phylogenies for the Turbellaria and the phylum Platyhelminthes. Few morphology-based character matrices are available in the systematic literature concerning flatworms, and molecular-based phylogenies are rapidly providing the only means by which we can estimate phylogenies cladistically. Character matrices based on gross morphology and ultrastructure are required and should be internally consistent, i.e. character coding should follow a set of a priori guidelines and character duplication and contradiction is avoided. In order to test our molecular phylogenies we need complementary data sets from morphology. To understand morphological homology we need phylogenetic evidence from independent (e.g. molecular) data. Fully complementary morphological and molecular data sets enable us to validate phylogenetic hypotheses and the combination of these sets in phylogenetic reconstruction utilises all statements of homology. Working phylogenies which include all phylogenetic information not only shed light on individual character evolution, but form a strong basis for comparative studies investigating the origin and evolutionary radiation of the taxonomic group under scrutiny. This revised version was published online in July 2006 with corrections to the Cover Date.     

SuperTRI: A new approach based on branch support analyses of multiple independent data sets for assessing reliability of phylogenetic inferences

Supermatrix and supertree are two methods for constructing a phylogenetic tree by using multiple data sets. However, these methods are not a panacea, as conflicting signals between data sets can lead to misinterpret the evolutionary history of taxa. In particular, the supermatrix approach is expected to be misleading if the species-tree signal is not dominant after the combination of the data sets. Moreover, most current supertree methods suffer from two limitations: (i) they ignore or misinterpret secondary (non-dominant) phylogenetic signals of the different data sets; and (ii) the logical basis of node robustness measures is unclear.To overcome these limitations, we propose a new approach, called SuperTRI, which is based on the branch support analyses of the independent data sets, and where the reliability of the nodes is assessed using three measures: the supertree Bootstrap percentage and two other values calculated from the separate analyses: the mean branch support (mean Bootstrap percentage or mean posterior probability) and the reproducibility index.The SuperTRI approach is tested on a data matrix including seven genes for 82 taxa of the family Bovidae (Mammalia, Ruminantia), and the results are compared to those found with the supermatrix approach. The phylogenetic analyses of the supermatrix and independent data sets were done using four methods of tree reconstruction: Bayesian inference, maximum likelihood, and unweighted and weighted maximum parsimony. The results indicate, firstly, that the SuperTRI approach shows less sensitivity to the four phylogenetic methods, secondly, that it is more accurate to interpret the relationships among taxa, and thirdly, that interesting conclusions on introgression and radiation can be drawn from the comparisons between SuperTRI and supermatrix analyses. To cite this article: A. Ropiquet et al., C. R. Biologies 332 (2009).     

Using Confidence Set Heuristics During Topology Search Improves the Robustness of Phylogenetic Inference

We examine the impact of likelihood surface characteristics on phylogenetic inference. Amino acid data sets simulated from topologies with branch length features chosen to represent varying degrees of difficulty for likelihood maximization are analyzed. We present situations where the tree found to achieve the global maximum in likelihood is often not equal to the true tree. We use the program covSEARCH to demonstrate how the use of adaptively sized pools of candidate trees that are updated using confidence tests results in solution sets that are highly likely to contain the true tree. This approach requires more computation than traditional maximum likelihood methods, hence covSEARCH is best suited to small to medium-sized alignments or large alignments with some constrained nodes. The majority rule consensus tree computed from the confidence sets also proves to be different from the generating topology. Although low phylogenetic signal in the input alignment can result in large confidence sets of trees, some biological information can still be obtained based on nodes that exhibit high support within the confidence set. Two real data examples are analyzed: mammal mitochondrial proteins and a small tubulin alignment. We conclude that the technique of confidence set optimization can significantly improve the robustness of phylogenetic inference at a reasonable computational cost. Additionally, when either very short internal branches or very long terminal branches are present, confident resolution of specific bipartitions or subtrees, rather than whole-tree phylogenies, may be the most realistic goal for phylogenetic methods. [Reviewing Editor: Dr. Nicolas Galtier]     

Time squared: repeated measures on phylogenies

Studies of gene expression profiles in response to external perturbation generate repeated measures data that generally follow nonlinear curves. To explore the evolution of such profiles across a gene family, we introduce phylogenetic repeated measures (PR) models. These models draw strength from 2 forms of correlation in the data. Through gene duplication, the family's evolutionary relatedness induces the first form. The second is the correlation across time points within taxonic units, individual genes in this example. We borrow a Brownian diffusion process along a given phylogenetic tree to account for the relatedness and co-opt a repeated measures framework to model the latter. Through simulation studies, we demonstrate that repeated measures models outperform the previously available approaches that consider the longitudinal observations or their differences as independent and identically distributed by using deviance information criteria as Bayesian model selection tools; PR models that borrow phylogenetic information also perform better than nonphylogenetic repeated measures models when appropriate. We then analyze the evolution of gene expression in the yeast kinase family using splines to estimate nonlinear behavior across 3 perturbation experiments. Again, the PR models outperform previous approaches and afford the prediction of ancestral expression profiles. To demonstrate PR model applicability more generally, we conclude with a short examination of variation in brain development across 4 primate species.