The relative performance of Bayesian and parsimony approaches when sampling characters evolving under homogeneous and heterogeneous sets of parameters

We tested whether it is beneficial for the accuracy of phylogenetic inference to sample characters that are evolving under different sets of parameters, using both Bayesian MCMC (Markov chain Monte Carlo) and parsimony approaches. We examined differential rates of evolution among characters, differential character-state frequencies and character-state space, and differential relative branch lengths among characters. We also compared the relative performance of parsimony and Bayesian analyses by progressively incorporating more of these heterogeneous parameters and progressively increasing the severity of this heterogeneity. Bayesian analyses performed better than parsimony when heterogeneous simulation parameters were incorporated into the substitution model. However, parsimony outperformed Bayesian MCMC when heterogeneous simulation parameters were not incorporated into the Bayesian substitution model. The higher the rate of evolution simulated, the better parsimony performed relative to Bayesian analyses. Bayesian and parsimony analyses converged in their performance as the number of simulated heterogeneous model parameters increased. Up to a point, rate heterogeneity among sites was generally advantageous for phylogenetic inference using both approaches. In contrast, branch-length heterogeneity was generally disadvantageous for phylogenetic inference using both parsimony and Bayesian approaches. Parsimony was found to be more conservative than Bayesian analyses, in that it resolved fewer incorrect clades.
© The Willi Hennig Society 2006.     

Phylogenetic analysis and intraspecific variation: performance of parsimony,likelihood, and distance methods

Intraspecific variation is abundant in all types of systematic characters but is rarely addressed in simulation studies of phylogenetic method performance. We compared the accuracy of 15 phylogenetic methods using simulations to (1) determine the most accurate method(s) for analyzing polymorphic data (under simplified conditions) and (2) test if generalizations about the performance of phylogenetic methods based on previous simulations of fixed (nonpolymorphic) characters are robust to a very different evolutionary model that explicitly includes intraspecific variation. Simulated data sets consisted of allele frequencies that evolved by genetic drift. The phylogenetic methods included eight parsimony coding methods, continuous maximum likelihood, and three distance methods (UPGMA, neighbor joining, and Fitch-Margoliash) applied to two genetic distance measures (Nei's and the modified Cavalli-Sforza and Edwards chord distance). Two sets of simulations were performed. The first examined the effects of different branch lengths, sample sizes (individuals sampled per species), numbers of characters, and numbers of alleles per locus in the eight-taxon case. The second examined more extensively the effects of branch length in the four-taxon, two-allele case. Overall, the most accurate methods were likelihood, the additive distance methods (neighbor joining and Fitch-Margoliash), and the frequency parsimony method. Despite the use of a very different evolutionary model in the present article, many of the results are similar to those from simulations of fixed characters. Similarities include the presence of the "Felsenstein zone," where methods often fail, which suggests that long-branch attraction may occur among closely related species through genetic drift. Differences between the results of fixed and polymorphic data simulations include the following: (1) UPGMA is as accurate or more accurate than nonfrequency parsimony methods across nearly all combinations of branch lengths, and (2) likelihood and the additive distance methods are not positively misled under any combination of branch lengths tested (even when the assumptions of the methods are violated and few characters are sampled). We found that sample size is an important determinant of accuracy and affects the relative success of methods (i.e., distance and likelihood methods outperform parsimony at small sample sizes). Attempts to generalize about the behavior of phylogenetic methods should consider the extreme examples offered by fixed-mutation models of DNA sequence data and genetic-drift models of allele frequencies.     

Phylogenetic inference using non‐redundant coding of dependent characters versus alternative approaches for protein‐coding genes

Contemporary molecular phylogenetic analyses often encompass a broad range of taxonomic diversity while maintaining high levels of sampling within each major taxon. To help maximize phylogenetic signal in such studies, one may analyse multiple levels of characters simultaneously. We test the performance of both the original and the modified versions of non‐redundant coding of dependent characters (NRCDC) relative to commonly applied alternative character‐sampling strategies using codon‐based simulations under a range of conditions. Both original and modified NRCDC generally outperformed other character‐sampling strategies that only sampled characters at one level (nucleotides or amino acids) over a broader range of simulation parameters than any of the alternative character‐sampling strategies with respect to both overall success of resolution and averaged overall success of resolution in the parsimony‐based analyses. Based on theoretical considerations and the results of our simulations, we encourage application and further testing of modified NRCDC in parsimony‐based molecular phylogenetic analyses that sample exons of protein‐coding genes. We expect that modified NRCDC will generally increase both accuracy and branch‐support over commonly applied alternative character‐sampling strategies when analysed using the same phylogenetic inference method, particularly in studies that sample both closely and distantly related taxa with clades representing both ancient and recent divergences. © The Willi Hennig Society 2010.     

Developmental characters in phylogenetic inference and their absolute timing information

Despite the recent surge of interest in studying the evolution of development, surprisingly little work has been done to investigate the phylogenetic signal in developmental characters. Yet, both the potential usefulness of developmental characters in phylogenetic reconstruction and the validity of inferences on the evolution of developmental characters depend on the presence of such a phylogenetic signal and on the ability of our coding scheme to capture it. In a recent study, we showed, using simulations, that a new method (called the continuous analysis) using standardized time or ontogenetic sequence data and squared-change parsimony outperformed event pairing and event cracking in analyzing developmental data on a reference phylogeny. Using the same simulated data, we demonstrate that all these coding methods (event pairing and standardized time or ontogenetic sequence data) can be used to produce phylogenetically informative data. Despite some dependence between characters (the position of an event in an ontogenetic sequence is not independent of the position of other events in the same sequence), parsimony analysis of such characters converges on the correct phylogeny as the amount of data increases. In this context, the new coding method (developed for the continuous analysis) outperforms event pairing; it recovers a lower proportion of incorrect clades. This study thus validates the use of ontogenetic data in phylogenetic inference and presents a simple coding scheme that can extract a reliable phylogenetic signal from these data.     

Reconstructing ancestral ecologies: challenges and possible solutions

There are several ways to extract information about the evolutionary ecology of clades from their phylogenies. Of these, character state optimization and 'ancestor reconstruction' are perhaps the most widely used despite their being fraught with assumptions and potential pitfalls. Requirements for robust inferences of ancestral traits in general (i.e. those applicable to all types of characters) include accurate and robust phylogenetic hypotheses, complete species-level sampling and the appropriate choice of optimality criterion. Ecological characters, however, also require careful consideration of methods for accounting for intraspecific variability. Such methods include 'Presence Coding' and 'Polymorphism Coding' for discrete ecological characters, and 'Range Coding' and 'MaxMin Coding' for continuously variable characters. Ultimately, however, historical inferences such as these are, as with phylogenetic inference itself, associated with a degree of uncertainty. Statistically based uncertainty estimates are available within the context of model-based inference (e.g. maximum likelihood and Bayesian); however, these measures are only as reliable as the chosen model is appropriate. Although generally thought to preclude the possibility of measuring relative uncertainty or support for alternative possible reconstructions, certain useful non-statistical support measures (i.e. 'Sharkey support' and 'Parsimony support') are applicable to parsimony reconstructions.     

Is there a direct ontogenetic criterion in systematics?

Much recent literature focuses on whether ontogenetic information can be used as a direct criterion for determining the polarity of character trasformations in systematic analysis. This paper reviews the relevant literature and concludes that the ontogenetic criterion is dependent on parsimony rather than the sequence observed during ontogeny. It is not, therefore, based on the discredited arguments of recapitulation. From the perspective of phyologenetic systematics the ontogenetic criterion is a valid means of polarizing character transformations that represents a special case of a broader methodology involving parsimony. The alternative perspective perspective of patttern cladistics holds that polarity should be contained within the data and not imposed upon it. Thus, ontogeny is not required to polarize characters, but ontogenetic information can generate unequivocal character interpretations in terms of the relative generality of related attributes, and in the sense that absence precedes presence. Furthermore, ontogeny is central to systematics, providing empirical evidence of character transformation, information on the whole life cycle and an escape from systematics being teleologically related to phylogenetic inference and the theory of evolution.     

Phylogeny of eriphioid crabs (Brachyura,Eriphioidea) inferred from molecular and morphological studies

The evolutionary relationships of the brachyuran crab superfamily Eriphioidea, commonly known as stone or rubble crabs, are examined. Analysis of three mitochondrial (12S, 16S and COI) and two nuclear loci (18S and Histone 3) was carried out for 51 taxa representing the Carpilioidea, Dairoidea, Eriphioidea, Goneplacoidea, Parthenopoidea, Pilumnoidea, Portunoidea, Pseudozioidea and Xanthoidea. Phylogenetic analyses of molecular data used three methods of inference that recovered similar topologies with minor differences. Maximum parsimony analysis of 20 morphological characters taken from first zoeas of 11 species yielded two equally parsimonious trees and generally supported the molecular analyses. None of the analyses recovered Eriphioidea as monophyletic, and each of the eriphioid families represented by two or more taxa was shown to be polyphyletic in both molecular and larval analyses. This study indicates that the present classification based on adult morphology is incongruent with phylogenetic relationships and that the diagnostic characters the result of convergence (particularly in feeding morphology) rather than shared ancestry.     

Bias and conflict in phylogenetic inference of myco-heterotrophic plants: a case study in Thismiaceae

Due to morphological reduction and absence of amplifiable plastid genes, the identification of photosynthetic relatives of heterotrophic plants is problematic. Although nuclear and mitochondrial gene sequences may offer a welcome alternative source of phylogenetic markers, the presence of rate heterogeneity in these genes may introduce bias/systematic error in phylogenetic analyses. We examine the phylogenetic position of Thismiaceae based on nuclear 18S rDNA and mitochondrial atpA DNA sequence data, as well as using parsimony, likelihood and Bayesian inference methods. Significant differences in evolutionary rates of these genes between closely related taxa lead to conflicting results: while parsimony analyses of 18S rDNA and combined data strongly support the monophyly of Thismiaceae, Bayesian inference, with and without a relaxed molecular clock, as well as the Swofford–Olsen–Waddell–Hillis (SOWH) test confidently reject this hypothesis. We show that rate heterogeneity in our data leads to long-branch attraction artifacts in parsimony analysis. However, using model-based inference methods the question of whether Thismiaceae are monophyletic remains elusive. On the one hand maximum likelihood nonparametric bootstrapping and parametric hypothesis tests fail to support a paraphyletic Thismiaceae, on the other hand Bayesian inference methods (both without and with a relaxed clock) significantly reject a monophyletic Thismiaceae. These results show that an adequate sampling, the use of rate homogeneous data, and the application of different inference methods are important factors for developing phylogenetic hypotheses of myco-heterotrophic plants. © The Willi Hennig Society 2009.     

The enigmatic monotypic crab plover Dromas ardeola is closely related to pratincoles and coursers (Aves, Charadriiformes, Glareolidae)

The phylogenetic placement of the monotypic crab plover Dromasardeola (Aves, Charadriiformes) remains controversial. Phylogenetic analysis of anatomical and behavioral traits using phenetic and cladistic methods of tree inference have resulted in conflicting tree topologies, suggesting a close association of Dromas to members of different suborders and lineages within Charadriiformes. Here, we revisited the issue by applying Bayesian and parsimony methods of tree inference to 2,012 anatomical and 5,183 molecular characters to a set of 22 shorebird genera (including Turnix). Our results suggest that Bayesian analysis of anatomical characters does not resolve the phylogenetic relationship of shorebirds with strong statistical support. In contrast, Bayesian and parsimony tree inference from molecular data provided much stronger support for the phylogenetic relationships within shorebirds, and support a sister relationship of Dromas to Glareolidae (pratincoles and coursers), in agreement with previously published DNA-DNA hybridization studies.     

Phylogeny of Australian species of flatheads (Teleostei, Platycephalidae) as determined by allozyme electrophoresis

Phylogenetic relationships of 24 taxa of Platycephalid fish were examined for up to 40 loci using starch and polyacrylarnide gel electrophoresis. Over 400 different enzyme mobilities were scored. The extent of homoplasy was estimated to be between 5 and 12%. Polymorphism at the 1% level was observed at 26 of the 40 loci. The importance of shared polymorphic alleles in cladistic analysis was endorsed by the different results obtained from multistate and binary coding of alleles. Many loci were shown to contain phylogenetically informative ancestral alleles which appear irregularly in extant species. The polymorphism parsimony method provided the most informative results. Genetic similarities between endemic Australian species from each family stem provide the basis for a major revision of subfamilies and genera within the family.     

A phylogenetic framework for root lesion nematodes of the genus Pratylenchus (Nematoda): Evidence from 18S and D2-D3 expansion segments of 28S ribosomal RNA genes and morphological characters

The root lesion nematodes of the genus Pratylenchus Filipjev, 1936 are migratory endoparasites of plant roots, considered among the most widespread and important nematode parasites in a variety of crops. We obtained gene sequences from the D2 and D3 expansion segments of 28S rRNA partial and 18S rRNA from 31 populations belonging to 11 valid and two unidentified species of root lesion nematodes and five outgroup taxa. These datasets were analyzed using maximum parsimony and Bayesian inference. The alignments were generated using the secondary structure models for these molecules and analyzed with Bayesian inference under the standard models and the complex model, considering helices under the doublet model and loops and bulges under the general time reversible model. The phylogenetic informativeness of morphological characters is tested by reconstruction of their histories on rRNA based trees using parallel parsimony and Bayesian approaches. Phylogenetic and sequence analyses of the 28S D2–D3 dataset with 145 accessions for 28 species and 18S dataset with 68 accessions for 15 species confirmed among large numbers of geographical diverse isolates that most classical morphospecies are monophyletic. Phylogenetic analyses revealed at least six distinct major clades of examined Pratylenchus species and these clades are generally congruent with those defined by characters derived from lip patterns, numbers of lip annules, and spermatheca shape. Morphological results suggest the need for sophisticated character discovery and analysis for morphology based phylogenetics in nematodes.     

The phylogeny of Iberian Aphodiini species (Coleoptera, Scarabaeoidea, Scarabaeidae, Aphodiinae) based on morphology

Abstract.  The phylogeny of Iberian Aphodiini species was reconstructed based on morphology. Wing venation, mouthparts, male and female genitalia, and external morphology provided ninety-four characters scored for ninety-three Aphodiini species. Phylogenetic analyses were based on maximum parsimony and Bayesian inference criteria. Maximum parsimony consensus trees recovered Acrossus species as a sister group of the remaining Aphodiini, followed by two other branches, one including Neagolius , Plagiogonus , Ahermodontus and Ammoecius species, and the other including Oxyomus , Nimbus , Heptaulacus and Euheptaulacus species. The remaining studied taxa clustered in an unresolved group. Bayesian inference trees recovered Acrossus as the sister group of the remaining Iberian Aphodiini, followed by Colobopterus erraticus and the rest of the Iberian Aphodiini, but this latter branch was unresolved. The general lack of statistical support for the inferred phylogenetic relationships at terminal nodes using both maximum parsimony and Bayesian inference suggests that variation in morphological characters useful for phylogenetic inference in the present study is small, perhaps as a consequence of a radiation event occurring at the origin of the tribe. A probable evolutionary pattern for Aphodiini is proposed which infers six groups, namely Acrossian, Ammoecian, Oxyomian, Aphodian s.str., Colobopteran and Aphodian s.l. clades.     

Relative benefits of amino‐acid,codon, degeneracy,DNA, and purine‐pyrimidine character coding for phylogenetic analyses of exons

Both traditional as well as 10 more recent methods of coding characters from exons of protein‐coding genes are reviewed. The more recent methods collectively blur the distinction between nucleotide and amino‐acid coding and enable investigators to carefully quantify the effects of different sources of phylogenetic signal as well as their potential biases. Codon models, which explicitly model silent and replacement substitutions, are a major advance and are expected to be broadly useful for simultaneously inferring recent and ancient divergences, unlike amino‐acid coding. Degeneracy coding, wherein ambiguity codes are used to eliminate silent substitutions at the individual‐nucleotide level, has clear advantages over scoring amino‐acid characters. Nucleotide, codon, and amino‐acid models are now directly comparable with easy‐to‐use programs, and widely used phylogenetics programs can analyze partitioned supermatrices that incorporate all three types of model. Therefore, it should become standard practice to test among these alternative model types before conducting parametric phylogenetic analyses. An earlier study of 78 protein‐coding genes from 360 green‐plant plastid genomes is used as an empirical example with which to quantify the relative performance of alternative character‐coding methods using five quantification measures. Codon models were selected as having the best fit to the data, yet were outperformed by nucleotide models for all five quantification measures. Third‐codon positions were found to be an important source of phylogenetic signal and even outperformed analyses of first and second positions for some measures. Degeneracy coding generally performed at least as well as amino‐acid coding and is an arguably more effective alternative.     

Comparison of heuristic approaches to the generalized tree alignment problem

Two commonly used heuristic approaches to the generalized tree alignment problem are compared in the context of phylogenetic analysis of DNA sequence data. These approaches, multiple sequence alignment + phylogenetic tree reconstruction (MSA+TR) and direct optimization (DO), are alternative heuristic procedures used to approach the nested NP‐Hard optimizations presented by the phylogenetic analysis of unaligned sequences under maximum parsimony. Multiple MSA+TR implementations and DO were compared in terms of optimality score (phylogenetic tree cost) over multiple empirical and simulated datasets with differing levels of heuristic intensity. In all cases examined, DO outperformed MSA+TR with average improvement in parsimony score of 14.78% (5.64–52.59%).     

Reconstructing phylogenies from allozyme data: comparing method performance with congruence

Allozyme data are widely used to infer the phylogenies of populations and closely-related species. Numerous parsimony, distance, and likelihood methods have been proposed for phylogenetic analysis of these data; the relative merits of these methods have been debated vigorously, but their accuracy has not been well explored. In this study, I compare the performance of 13 phylogenetic methods (six parsimony, six distance, and continuous maximum likelihood) by applying a congruence approach to eight allozyme data sets from the literature. Clades are identified that are supported by multiple data sets other than allozymes (e.g. morphology, DNA sequences), and the ability of different methods to recover these 'known' clades is compared. The results suggest that (1) distance and likelihood methods generally outperform parsimony methods, (2) methods that utilize frequency data tend to perform well, and (3) continuous maximum likelihood is among the most accurate methods, and appears to be robust to violations of its assumptions. These results are in agreement with those from recent simulation studies, and help provide a basis for empirical workers to choose among the many methods available for analysing allozyme characters.     

Incorporation of gap characters and lineage-specific regions into phylogenetic analyses of gene families from divergent clades: an example from the kinesin superfamily across eukaryotes

The kinesin superfamily across eukaryotes was used to examine how incorporation of gap characters scored from conserved regions shared by all members of a gene family and incorporation of amino acid and gap characters scored from lineage‐specific regions affect gene‐tree inference of the gene family as a whole. We addressed these two questions in the context of two different densities of sequence sampling, four alignment programs, and two methods of tree construction. Taken together, our findings suggest the following. First, gap characters should be incorporated into gene‐tree inference, even for divergent sequences. Second, gene regions that are not conserved among all or most sequences sampled should not be automatically discarded without evaluation of potential phylogenetic signal that may be contained in gap and/or sequence characters. Third, among the four alignment programs evaluated using their default alignment parameters, Clustal may be expected to output alignments that result in the greatest gene‐tree resolution and support. Yet, this high resolution and support should be regarded as optimistic, rather than conservative, estimates. Fourth, this same conclusion regarding resolution and support holds for Bayesian gene‐tree analyses relative to parsimony‐jackknife gene‐tree analyses. We suggest that a more conservative approach, such as aligning the sequences using DIALIGN‐T or MAFFT, analyzing the appropriate characters using parsimony, and assessing branch support using the jackknife, is more appropriate for inferring gene trees of divergent gene families. © The Willi Hennig Society 2007.     

Phylogeny of the genus Palmanura (Collembola: Neanuridae)

In order to assess the phylogenetic structure of the springtail genus Palmanura, as well as to test the monophyly of the tribe Sensillanurini (Neanuridae: Neanurinae), a data matrix of morphological (chaetotactic and other) characters of members of this group was assembled and analysed in the light of Wagner parsimony. The data matrix included all the known members of the Neotropical genus Palmanura, plus representatives of Sensillanura and Americanura. Although not all the clades obtained were highly supported by bootstrap resampling, some structures were relatively constant under different approaches. Alternative analyses (unordered and ordered character states, rescaled weighting procedure) were applied. While alternative solutions were obtained, a number of structures were shared by the results irrespective of the method used. On this basis, the results suggest that some further reassessment is required to confirm formally the monophyly of the tribe Sensillanurini. The genera Palmanura and Americanura are mutually poly/paraphyletic; we thus suggest that Palmanura should be considered as a synonym of Americanura, although some character reassessment and more varied outgroup species may be necessary before a formal generic redefinition can be proposed. Finally, a comparison of the performance of the characters under Wagner parsimony analysis indicated that differences in the characters’ retention indexes are due not to the topological (tagmal) position of the traits involved, but to character coding: the characters describing quantitative features (generally numbers of setae) generally performed worse than other types of characters under parsimony. An updated list of the known members of the Sensillanurini (Collembola: Neanuridae: Neanurinae) is presented.
© The Willi Hennig Society 2009.     

Weighted parsimony outperforms other methods of phylogenetic inference under models appropriate for morphology

One of the lasting controversies in phylogenetic inference is the degree to which specific evolutionary models should influence the choice of methods. Model‐based approaches to phylogenetic inference (likelihood, Bayesian) are defended on the premise that without explicit statistical models there is no science, and parsimony is defended on the grounds that it provides the best rationalization of the data, while refraining from assigning specific probabilities to trees or character‐state reconstructions. Authors who favour model‐based approaches often focus on the statistical properties of the methods and models themselves, but this is of only limited use in deciding the best method for phylogenetic inference—such decision also requires considering the conditions of evolution that prevail in nature. Another approach is to compare the performance of parsimony and model‐based methods in simulations, which traditionally have been used to defend the use of models of evolution for DNA sequences. Some recent papers, however, have promoted the use of model‐based approaches to phylogenetic inference for discrete morphological data as well. These papers simulated data under models already known to be unfavourable to parsimony, and modelled morphological evolution as if it evolved just like DNA, with probabilities of change for all characters changing in concert along tree branches. The present paper discusses these issues, showing that under reasonable and less restrictive models of evolution for discrete characters, equally weighted parsimony performs as well or better than model‐based methods, and that parsimony under implied weights clearly outperforms all other methods.     

Systematics and morphological evolution within the moss family Bryaceae: a comparison between parsimony and Bayesian methods for reconstruction of ancestral character states

The Bryaceae are a large cosmopolitan moss family including genera of significant morphological and taxonomic complexity. Phylogenetic relationships within the Bryaceae were reconstructed based on DNA sequence data from all three genomic compartments. In addition, maximum parsimony and Bayesian inference were employed to reconstruct ancestral character states of 38 morphological plus four habitat characters and eight insertion/deletion events. The recovered phylogenetic patterns are generally in accord with previous phylogenies based on chloroplast DNA sequence data and three major clades are identified. The first clade comprises Bryum bornholmense, B. rubens, B. caespiticium, and Plagiobryum. This corroborates the hypothesis suggested by previous studies that several Bryum species are more closely related to Plagiobryum than to the core Bryum species. The second clade includes Acidodontium, Anomobryum, and Haplodontium, while the third clade contains the core Bryum species plus Imbribryum. Within the latter clade, B. subapiculatum and B. tenuisetum form the sister clade to Imbribryum. Reconstructions of ancestral character states under maximum parsimony and Bayesian inference suggest fourteen morphological synapomorphies for the ingroup and synapomorphies are detected for most clades within the ingroup. Maximum parsimony and Bayesian reconstructions of ancestral character states are mostly congruent although Bayesian inference shows that the posterior probability of ancestral character states may decrease dramatically when node support is taken into account. Bayesian inference also indicates that reconstructions may be ambiguous at internal nodes for highly polymorphic characters.     

Uninode coding vs gene tree parsimony for phylogenetic reconstruction using duplicate genes

Two different methods of using paralogous genes for phylogenetic inference have been proposed: reconciled trees (or gene tree parsimony) and uninode coding. Gene tree parsimony suffers from 10 serious problems, including differential weighting of nucleotide and gap characters, undersampling which can be misinterpreted as synapomorphy, all of the characters not being allowed to interact, and conflict between gene trees being given equal weight, regardless of branch support. These problems are largely avoided by using uninode coding. The uninode coding method is elaborated to address multiple gene duplications within a single gene tree family and handle problems caused by lack of gene tree resolution. An example of vertebrate phylogeny inferred from nine genes is reanalyzed using uninode coding. We suggest that uninode coding be used instead of gene tree parsimony for phylogenetic inference from paralogous genes.