Fast statistical tests for detecting heterotachy in protein evolution

The w statistic introduced by Lockhart et al. (1998. A covariotide model explains apparent phylogenetic structure of oxygenic photosynthetic lineages. Mol Biol Evol. 15:1183-1188) is a simple and easily calculated statistic intended to detect heterotachy by comparing amino acid substitution patterns between two monophyletic groups of protein sequences. It is defined as the difference between the fraction of varied sites in both groups and the fraction of varied sites in each group. The w test has been used to distinguish a covarion process from equal rates and rates variation across sites processes. Using simulation we show that the w test is effective for small data sets and for data sets that have low substitution rates in the groups but can have difficulties when these conditions are not met. Using site entropy as a measure of variability of a sequence site, we modify the w statistic to a w' statistic by assigning as varied in one group those sites that are actually varied in both groups but have a large entropy difference. We show that the w' test has more power to detect two kinds of heterotachy processes (covarion and bivariate rate shifts) in large and variable data. We also show that a test of Pearson's correlation of the site entropies between two monophyletic groups can be used to detect heterotachy and has more power than the w' test. Furthermore, we demonstrate that there are settings where the correlation test as well as w and w' tests do not detect heterotachy signals in data simulated under a branch length mixture model. In such cases, it is sometimes possible to detect heterotachy through subselection of appropriate taxa. Finally, we discuss the abilities of the three statistical tests to detect a fourth mode of heterotachy: lineage-specific changes in proportion of variable sites.     

Modelling heterotachy in phylogenetic inference by reversible-jump Markov chain Monte Carlo

The rate at which a given site in a gene sequence alignment evolves over time may vary. This phenomenon--known as heterotachy--can bias or distort phylogenetic trees inferred from models of sequence evolution that assume rates of evolution are constant. Here, we describe a phylogenetic mixture model designed to accommodate heterotachy. The method sums the likelihood of the data at each site over more than one set of branch lengths on the same tree topology. A branch-length set that is best for one site may differ from the branch-length set that is best for some other site, thereby allowing different sites to have different rates of change throughout the tree. Because rate variation may not be present in all branches, we use a reversible-jump Markov chain Monte Carlo algorithm to identify those branches in which reliable amounts of heterotachy occur. We implement the method in combination with our 'pattern-heterogeneity' mixture model, applying it to simulated data and five published datasets. We find that complex evolutionary signals of heterotachy are routinely present over and above variation in the rate or pattern of evolution across sites, that the reversible-jump method requires far fewer parameters than conventional mixture models to describe it, and serves to identify the regions of the tree in which heterotachy is most pronounced. The reversible-jump procedure also removes the need for a posteriori tests of 'significance' such as the Akaike or Bayesian information criterion tests, or Bayes factors. Heterotachy has important consequences for the correct reconstruction of phylogenies as well as for tests of hypotheses that rely on accurate branch-length information. These include molecular clocks, analyses of tempo and mode of evolution, comparative studies and ancestral state reconstruction. The model is available from the authors' website, and can be used for the analysis of both nucleotide and morphological data.     

Maximum likelihood outperforms maximum parsimony even when evolutionary rates are heterotachous

Heterotachy occurs when the relative evolutionary rates among sites are not the same across lineages. Sequence alignments are likely to exhibit heterotachy with varying severity because the intensity of purifying selection and adaptive forces at a given amino acid or DNA sequence position is unlikely to be the same in different species. In a recent study, the influence of heterotachy on the performance of different phylogenetic methods was examined using computer simulation for a four-species phylogeny. Maximum parsimony (MP) was reported to generally outperform maximum likelihood (ML). However, our comparisons of MP and ML methods using the methods and evaluation criteria employed in that study, but considering the possible range of proportions of sites involved in heterotachy, contradict their findings and indicate that, in fact, ML is significantly superior to MP even under heterotachy.     

A mixed branch length model of heterotachy improves phylogenetic accuracy

Evolutionary relationships are typically inferred from molecular sequence data using a statistical model of the evolutionary process. When the model accurately reflects the underlying process, probabilistic phylogenetic methods recover the correct relationships with high accuracy. There is ample evidence, however, that models commonly used today do not adequately reflect real-world evolutionary dynamics. Virtually all contemporary models assume that relatively fast-evolving sites are fast across the entire tree, whereas slower sites always evolve at relatively slower rates. Many molecular sequences, however, exhibit site-specific changes in evolutionary rates, called "heterotachy." Here we examine the accuracy of 2 phylogenetic methods for incorporating heterotachy, the mixed branch length model--which incorporates site-specific rate changes by summing likelihoods over multiple sets of branch lengths on the same tree--and the covarion model, which uses a hidden Markov process to allow sites to switch between variable and invariable as they evolve. Under a variety of simple heterogeneous simulation conditions, the mixed model was dramatically more accurate than homotachous models, which were subject to topological biases as well as biases in branch length estimates. When data were simulated with strong versions of the types of heterotachy observed in real molecular sequences, the mixed branch length model was more accurate than homotachous techniques. Analyses of empirical data sets confirmed that the mixed branch length model can improve phylogenetic accuracy under conditions that cause homotachous models to fail. In contrast, the covarion model did not improve phylogenetic accuracy compared with homotachous models and was sometimes substantially less accurate. We conclude that a mixed branch length approach, although not the solution to all phylogenetic errors, is a valuable strategy for improving the accuracy of inferred trees.     

Estimation of rates-across-sites distributions in phylogenetic substitution models

Previous work has shown that it is often essential to account for the variation in rates at different sites in phylogenetic models in order to avoid phylogenetic artifacts such as long branch attraction. In most current models, the gamma distribution is used for the rates-across-sites distributions and is implemented as an equal-probability discrete gamma. In this article, we introduce discrete distribution estimates with large numbers of equally spaced rate categories allowing us to investigate the appropriateness of the gamma model. With large numbers of rate categories, these discrete estimates are flexible enough to approximate the shape of almost any distribution. Likelihood ratio statistical tests and a nonparametric bootstrap confidence-bound estimation procedure based on the discrete estimates are presented that can be used to test the fit of a parametric family. We applied the methodology to several different protein data sets, and found that although the gamma model often provides a good parametric model for this type of data, rate estimates from an equal-probability discrete gamma model with a small number of categories will tend to underestimate the largest rates. In cases when the gamma model assumption is in doubt, rate estimates coming from the discrete rate distribution estimate with a large number of rate categories provide a robust alternative to gamma estimates. An alternative implementation of the gamma distribution is proposed that, for equal numbers of rate categories, is computationally more efficient during optimization than the standard gamma implementation and can provide more accurate estimates of site rates.     

On inconsistency of the neighbor-joining, least squares, and minimum evolution estimation when substitution processes are incorrectly modeled

Using analytical methods, we show that under a variety of model misspecifications, Neighbor-Joining, minimum evolution, and least squares estimation procedures are statistically inconsistent. Failure to correctly account for differing rates-across-sites processes, failure to correctly model rate matrix parameters, and failure to adjust for parallel rates-across-sites changes (a rates-across-subtrees process) are all shown to lead to a "long branch attraction" form of inconsistency. In addition, failure to account for rates-across-sites processes is also shown to result in underestimation of evolutionary distances for a wide variety of substitution models, generalizing an earlier analytical result for the Jukes-Cantor model reported in Golding and a similar bias result for the GTR or REV model in Kelly and Rice (1996). Although standard rates-across-sites models can be employed in many of these cases to restore consistency, current models cannot account for other kinds of misspecification. We examine an idealized but biologically relevant case, where parallel changes in rates at sites across subtrees is shown to give rise to inconsistency. This changing rates-across-subtrees type model misspecification cannot be adjusted for with conventional methods or without carefully considering the rate variation in the larger tree. The results are presented for four-taxon trees, but the expectation is that they have implications for larger trees as well. To illustrate this, a simulated 42-taxon example is given in which the microsporidia, an enigmatic group of eukaryotes, are incorrectly placed at the archaebacteria-eukaryotes split because of incorrectly specified pairwise distances. The analytical nature of the results lend insight into the reasons that long branch attraction tends to be a common form of inconsistency and reasons that other forms of inconsistency like "long branches repel" can arise in some settings. In many of the cases of inconsistency presented, a particular incorrect topology is estimated with probability converging to one, the implication being that measures of uncertainty like bootstrap support will be unable to detect that there is a problem with the estimation. The focus is on distance methods, but previous simulation results suggest that the zones of inconsistency for distance methods contain the zones of inconsistency for maximum likelihood methods as well.     

Heterotachy, an important process of protein evolution.

Because of functional constraints, substitution rates vary among the positions of a protein but are usually assumed to be constant at a given site during evolution. The distribution of the rates across the sequence positions generally fits a Gamma distribution. Models of sequence evolution were accordingly designed and led to improved phylogenetic reconstruction. However, it has been convincingly demonstrated that the evolutionary rate of a given position is not always constant throughout time. We called such within-site rate variations heterotachy (for "different speed" in Greek). Yet, heterotachy was found among homologous sequences of distantly related organisms, often with different functions. In such cases, the functional constraints are likely different, which would explain the different distribution of variable sites. To evaluate the importance of heterotachy, we focused on amino acid sequences of mitochondrial cytochrome b, for which the function is likely the same in all vertebrates. Using 2,038 sequences, we demonstrate that 95% of the variable positions are heterotachous, i.e., underwent dramatic variations of substitution rate among vertebrate lineages. Heterotachy even occurs at small evolutionary scale, and in these cases it is very unlikely to be related to functional changes. Since a large number of sequences are required to efficiently detect heterotachy, the extent of this phenomenon could not be estimated for all proteins yet. It could be as large as for cytochrome b, since this protein is not a peculiar case. The observations made here open several new avenues of research, such as the understanding of the evolution of functional constraints or the improvement of phylogenetic reconstruction methods.     

Rates of molecular evolution in tree ferns are associated with body size,environmental temperature,and biological productivity

Variation in rates of molecular evolution (heterotachy) is a common phenomenon among plants. Although multiple theoretical models have been proposed, fundamental questions remain regarding the combined effects of ecological and morphological traits on rate heterogeneity. Here, we used tree ferns to explore the correlation between rates of molecular evolution in chloroplast DNA sequences and several morphological and environmental factors within a Bayesian framework. We revealed direct and indirect effects of body size, biological productivity, and temperature on substitution rates, where smaller tree ferns living in warmer and less productive environments tend to have faster rates of molecular evolution. In addition, we found that variation in the ratio of nonsynonymous to synonymous substitution rates (dN/dS) in the chloroplast rbcL gene was significantly correlated with ecological and morphological variables. Heterotachy in tree ferns may be influenced by effective population size associated with variation in body size and productivity. Macroevolutionary hypotheses should go beyond explaining heterotachy in terms of mutation rates and instead, should integrate population‐level factors to better understand the processes affecting the tempo of evolution at the molecular level.     

Application of two tests of multivariate discordancy to fisheries data sets

The generalized (Mahalanobis) distance and multivariate kurtosis are two powerful tests of multivariate discordancies (outliers). Unlike the generalized distance test, the multivariate kurtosis test has not been applied as a test of discordancy to fisheries data heretofore. We applied both tests, along with published algorithms for identifying suspected causal variable(s) of discordant observations, to two fisheries data sets from Lake Erie: total length, mass, and age from 1,234 burbot, Lota lota; and 22 combinations of unique subsets of 10 morphometrics taken from 119 yellow perch, Perca flavescens. For the burbot data set, the generalized distance test identified six discordant observations and the multivariate kurtosis test identified 24 discordant observations. In contrast with the multivariate tests, the univariate generalized distance test identified no discordancies when applied separately to each variable. Removing discordancies had a substantial effect on length-versus-mass regression equations. For 500-mm burbot, the percent difference in estimated mass after removing discordancies in our study was greater than the percent difference in masses estimated for burbot of the same length in lakes that differed substantially in productivity. The number of discordant yellow perch detected ranged from 0 to 2 with the multivariate generalized distance test and from 6 to 11 with the multivariate kurtosis test. With the kurtosis test, 108 yellow perch (90.7%) were identified as discordant in zero to two combinations, and five (4.2%) were identified as discordant in either all or 21 of the 22 combinations. The relationship among the variables included in each combination determined which variables were identified as causal. The generalized distance test identified between zero and six discordancies when applied separately to each variable. Removing the discordancies found in at least one-half of the combinations (k = 5) had a marked effect on a principal components analysis. In particular, the percent of the total variation explained by second and third principal components, which explain shape, increased by 52 and 44% respectively when the discordancies were removed. Multivariate applications of the tests have numerous ecological advantages over univariate applications, including improved management of fish stocks and interpretation of multivariate morphometric data.     

Is a multivariate consensus representation of genetic relationships among populations always meaningful?

To determine the relationships among closely related populations or species, two methods are commonly used in the literature: phylogenetic reconstruction or multivariate analysis. The aim of this article is to assess the reliability of multivariate analysis. We describe a method that is based on principal component analysis and Mantel correlations, using a two-step process: The first step consists of a single-marker analysis and the second step tests if each marker reveals the same typology concerning population differentiation. We conclude that if single markers are not congruent, the compromise structure is not meaningful. Our model is not based on any particular mutation process and it can be applied to most of the commonly used genetic markers. This method is also useful to determine the contribution of each marker to the typology of populations. We test whether our method is efficient with two real data sets based on microsatellite markers. Our analysis suggests that for closely related populations, it is not always possible to accept the hypothesis that an increase in the number of markers will increase the reliability of the typology analysis.     

Model misspecification and probabilistic tests of topology: evidence from empirical data sets

Probabilistic tests of topology offer a powerful means of evaluating competing phylogenetic hypotheses. The performance of the nonparametric Shimodaira-Hasegawa (SH) test, the parametric Swofford-Olsen-Waddell-Hillis (SOWH) test, and Bayesian posterior probabilities were explored for five data sets for which all the phylogenetic relationships are known with a very high degree of certainty. These results are consistent with previous simulation studies that have indicated a tendency for the SOWH test to be prone to generating Type 1 errors because of model misspecification coupled with branch length heterogeneity. These results also suggest that the SOWH test may accord overconfidence in the true topology when the null hypothesis is in fact correct. In contrast, the SH test was observed to be much more conservative, even under high substitution rates and branch length heterogeneity. For some of those data sets where the SOWH test proved misleading, the Bayesian posterior probabilities were also misleading. The results of all tests were strongly influenced by the exact substitution model assumptions. Simple models, especially those that assume rate homogeneity among sites, had a higher Type 1 error rate and were more likely to generate misleading posterior probabilities. For some of these data sets, the commonly used substitution models appear to be inadequate for estimating appropriate levels of uncertainty with the SOWH test and Bayesian methods. Reasons for the differences in statistical power between the two maximum likelihood tests are discussed and are contrasted with the Bayesian approach.     

The effect of heterotachy in multigene analysis using the neighbor joining method

Sequence alignments of multiple genes are routinely used to infer phylogenetic relationships among species. The analysis of their concatenation is more likely to give correct results under an assumption of homotachy (i.e., the evolutionary rates within lineages in each of the concatenated genes are constant during evolution). Here, we examine how the violation of homotachy (i.e., presence of within-site rate variation, called heterotachy) distorts species phylogenies. A theoretical examination has been conducted using a four taxon case and the neighbor joining (NJ) method, concluding that NJ recovers the incorrect tree when concatenated genes exhibit heterotachy. The application of average and weighted-average distance approaches, where gene boundaries are kept intact, overcomes the detrimental effect of heterotachy in multigene analysis using the NJ method.     

Likelihood, parsimony, and heterogeneous evolution

Evolutionary rates vary among sites and across the phylogenetic tree (heterotachy). A recent analysis suggested that parsimony can be better than standard likelihood at recovering the true tree given heterotachy. The authors recommended that results from parsimony, which they consider to be nonparametric, be reported alongside likelihood results. They also proposed a mixture model, which was inconsistent but better than either parsimony or standard likelihood under heterotachy. We show that their main conclusion is limited to a special case for the type of model they study. Their mixture model was inconsistent because it was incorrectly implemented. A useful nonparametric model should perform well over a wide range of possible evolutionary models, but parsimony does not have this property. Likelihood-based methods are therefore the best way to deal with heterotachy.     

Evaluation of the models handling heterotachy in phylogenetic inference

Background  

The evolutionary rate at a given homologous position varies across time. When sufficiently pronounced, this phenomenon – called heterotachy – may produce artefactual phylogenetic reconstructions under the commonly used models of sequence evolution. These observations have motivated the development of models that explicitly recognize heterotachy, with research directions proposed along two main axes: 1) the covarion approach, where sites switch from variable to invariable states; and 2) the mixture of branch lengths (MBL) approach, where alignment patterns are assumed to arise from one of several sets of branch lengths, under a given phylogeny.     

Topological Estimation Biases with Covarion Evolution

Covarion processes allow changes in evolutionary rates at sites along the branches of a phylogenetic tree. Covarion-like evolution is increasingly recognized as an important mode of protein evolution. Several recent reports suggest that maximum likelihood estimation employing covarion models may support different optimal topologies than estimation using standard rates-across-sites (RAS) models. However, it remains to be demonstrated that ignoring covarion evolution will generally result in topological misestimation. In this study we performed analytical and theoretical studies of limiting distances under the covarion model and four-taxon tree simulations to investigate the extent to which the covarion process impacts on phylogenetic estimation. In particular, we assessed the limits of an RAS model-based maximum likelihood method to recover the phylogenies when the sequence data were simulated under the covarion processes. We find that, when ignored, covarion processes can induce systematic errors in phylogeny reconstruction. Surprisingly, when sequences are evolved under a covarion process but an RAS model is used for estimation, we find that a long branch repel bias occurs.     

A simulation comparison of phylogeny algorithms under equal and unequal evolutionary rates [published erratum appears in Mol Biol Evol 1995 May;12(3):525]

Using simulated data, we compared five methods of phylogenetic tree estimation: parsimony, compatibility, maximum likelihood, Fitch- Margoliash, and neighbor joining. For each combination of substitution rates and sequence length, 100 data sets were generated for each of 50 trees, for a total of 5,000 replications per condition. Accuracy was measured by two measures of the distance between the true tree and the estimate of the tree, one measure sensitive to accuracy of branch lengths and the other not. The distance-matrix methods (Fitch- Margoliash and neighbor joining) performed best when they were constrained from estimating negative branch lengths; all comparisons with other methods used this constraint. Parsimony and compatibility had similar results, with compatibility generally inferior; Fitch- Margoliash and neighbor joining had similar results, with neighbor joining generally slightly inferior. Maximum likelihood was the most successful method overall, although for short sequences Fitch- Margoliash and neighbor joining were sometimes better. Bias of the estimates was inferred by measuring whether the independent estimates of a tree for different data sets were closer to the true tree than to each other. Parsimony and compatibility had particular difficulty with inaccuracy and bias when substitution rates varied among different branches. When rates of evolution varied among different sites, all methods showed signs of inaccuracy and bias.      

Combining multiple data sets in a likelihood analysis: which models are the best?

Until recently, phylogenetic analyses have been routinely based on homologous sequences of a single gene. Given the vast number of gene sequences now available, phylogenetic studies are now based on the analysis of multiple genes. Thus, it has become necessary to devise statistical methods to combine multiple molecular data sets. Here, we compare several models for combining different genes for the purpose of evaluating the likelihood of tree topologies. Three methods of branch length estimation were studied: assuming all genes have the same branch lengths (concatenate model), assuming that branch lengths are proportional among genes (proportional model), or assuming that each gene has a separate set of branch lengths (separate model). We also compared three models of among-site rate variation: the homogenous model, a model that assumes one gamma parameter for all genes, and a model that assumes one gamma parameter for each gene. On the basis of two nuclear and one mitochondrial amino acid data sets, our results suggest that, depending on the data set chosen, either the separate model or the proportional model represents the most appropriate method for branch length analysis. For all the data sets examined, one gamma parameter for each gene represents the best model for among-site rate variation. Using these models we analyzed alternative mammalian tree topologies, and we describe the effect of the assumed model on the maximum likelihood tree. We show that the choice of the model has an impact on the best phylogeny obtained.     

JML: testing hybridization from species trees

I introduce the software JML that tests for the presence of hybridization in multispecies sequence data sets by posterior predictive checking following Joly, McLenachan and Lockhart (2009, American Naturalist 174, e54). Although their method could potentially be applied on any data set, the lack of appropriate software made its application difficult. The software JML thus fills a need for an easy application of the method but also includes improvements such as the possibility to incorporate uncertainty in the species tree topology. The JML software uses a posterior distribution of species trees, population sizes and branch lengths to simulate replicate sequence data sets using the coalescent with no migration. A test quantity, defined as the minimum pairwise sequence distance between sequences of two species, is then evaluated on the simulated data sets and compared to the one estimated from the original data. Because the test quantity is a good predictor of hybridization events, departure from the bifurcating species tree model could be interpreted as evidence of hybridization. Software performance in terms of computing time is evaluated for several parameters. I also show an application example of the software for detecting hybridization among native diploid North American roses.     

Distance-based tests for homogeneity of multivariate dispersions

Summary The traditional likelihood‐based test for differences in multivariate dispersions is known to be sensitive to nonnormality. It is also impossible to use when the number of variables exceeds the number of observations. Many biological and ecological data sets have many variables, are highly skewed, and are zero‐inflated. The traditional test and even some more robust alternatives are also unreasonable in many contexts where measures of dispersion based on a non‐Euclidean dissimilarity would be more appropriate. Distance‐based tests of homogeneity of multivariate dispersions, which can be based on any dissimilarity measure of choice, are proposed here. They rely on the rotational invariance of either the multivariate centroid or the spatial median to obtain measures of spread using principal coordinate axes. The tests are straightforward multivariate extensions of Levene's test, with P‐values obtained either using the traditional F‐distribution or using permutation of either least‐squares or LAD residuals. Examples illustrate the utility of the approach, including the analysis of stabilizing selection in sparrows, biodiversity of New Zealand fish assemblages, and the response of Indonesian reef corals to an El Niño. Monte Carlo simulations from the real data sets show that the distance‐based tests are robust and powerful for relevant alternative hypotheses of real differences in spread.     

Artifactual phylogenies caused by correlated distribution of substitution rates among sites and lineages: the good, the bad, and the ugly

Despite the advances in understanding molecular evolution, current phylogenetic methods barely take account of a fraction of the complexity of evolution. We are chiefly constrained by our incomplete knowledge of molecular evolutionary processes and the limits of computational power. These limitations lead to the establishment of either biologically simplistic models that rarely account for a fraction of the complexity involved or overfitting models that add little resolution to the problem. Such oversimplified models may lead us to assign high confidence to an incorrect tree (inconsistency). Rate-across-site (RAS) models are commonly used evolutionary models in phylogenetic studies. These account for heterogeneity in the evolutionary rates among sites but do not account for changing within-site rates across lineages (heterotachy). If heterotachy is common, using RAS models may lead to systematic errors in tree inference. In this work we show possible misleading effects in tree inference when the assumption of constant within-site rates across lineages is violated using maximum likelihood. Using a simulation study, we explore the ways in which gamma stationary models can lead to wrong topology or to deceptive bootstrap support values when the within-site rates change across lineages. More precisely, we show that different degrees of heterotachy mislead phylogenetic inference when the model assumed is stationary. Finally, we propose a geometry-based approach to visualize and to test for the possible existence of bias due to heterotachy.