Taxonomic names and phylogenetic trees

This paper addresses the issue of philosophy of names within the context of biological taxonomy, more specifically how names refer. By contrasting two philosophies of names, one that is based on the idea that names can be defined and one that they cannot be defined, I point out some advantages of the latter within phylogenetic systematics. Due to the changing nature of phylogenetic hypotheses, the former approach tends to rob taxonomy from its unique communicative value since a name that is defined refers to whatever fits the definition. This is particularly troublesome should the hypothesis of phylogenetic relationship change. I argue that, should we decide to accept a new phylogenetic hypothesis, it is also likely that our view of what to name may change. A system where names only refer acknowledge this, and accordingly leaves it open whether to keep a name (and accept the way it refers in the new hypothesis) or discard a name and introduce new names for the parts of the tree that we find scientifically interesting. One of the main differences between a phylogenetic system of definition (PSD) and a phylogenetic system of reference (PSR) is that the former is governed by laws of language while the latter by communicative needs of taxonomists. Thus, a PSR tends to give primacy to phylogenetic trees rather than phylogenetic definitions of names should our views of which phylogenetic hypothesis to accept change. © 1998 The Norwegian Academy of Sciences and Letters     

A fast method for approximating maximum likelihoods of phylogenetic trees from nucleotide sequences

We have developed a rapid parsimony method for reconstructing ancestral nucleotide states that allows calculation of initial branch lengths that are good approximations to optimal maximum-likelihood estimates under several commonly used substitution models. Use of these approximate branch lengths (rather than fixed arbitrary values) as starting points significantly reduces the time required for iteration to a solution that maximizes the likelihood of a tree. These branch lengths are close enough to the optimal values that they can be used without further iteration to calculate approximate maximum-likelihood scores that are very close to the "exact" scores found by iteration. Several strategies are described for using these approximate scores to substantially reduce times needed for maximum-likelihood tree searches.     

MAC5: Bayesian inference of phylogenetic trees from DNA sequences incorporating gaps

SUMMARY: MAC5 implements MCMC sampling of the posterior distribution of tree topologies from DNA sequences containing gaps by using a five state model of evolution (the four nucleotides and the gap character).     

QuickTree: building huge Neighbour-Joining trees of protein sequences

We have written a fast implementation of the popular Neighbor-Joining tree building algorithm. QuickTree allows the reconstruction of phylogenies for very large protein families (including the largest Pfam alignment containing 27000 HIV GP120 glycoprotein sequences) that would be infeasible using other popular methods.     

Phylogenetic trees constructed from hydrophobicity values of protein sequences

Information about conformational properties of a protein is contained in the hydrophobicity values of the amino acids in its primary sequence. We have investigated the possibility of extracting meaningful evolutionary information from the comparison of the hydrophobicity values of the corresponding amino acids in the sequences of homologous proteins. Distance matrices for six families of homologous proteins were made on the basis of the differences in hydrophobicity values of the amino acids. The phylogenetic trees constructed from such matrices were at least as good (as judged from their faithful reflection of evolutionary relationships), as trees constructed from the usual minimum mutation distance matrix.     

Taxonomic relationships among Turkish water frogs as revealed by phylogenetic analyses using mtDNA gene sequences

We assessed taxonomic relationships among Turkish water frogs through estimation of phylogenetic relationships among 62 adult specimens from 44 distinct populations inhabiting seven main geographical regions of Turkey using 2897 bp sequences of the mitochondrial Cytb, 12S rRNA and 16S rRNA genes with equally-weighted parsimony, likelihood, and Bayesian methods of inference. Monophyletic clade (Clade A) of the northwesternmost (Thrace) samples is identified as Pelophylax ridibundus. The other clade (Clade B) consisted of two monophyletic subclades. One of these contains specimens from southernmost populations that are regarded as an unnamed species. The other subclade consists of two lineages, of which one corresponds to P. caralitanus and another to P. bedriagae. Taxonomic relationships of these two species are discussed and recognition of P. caralitanus as a subspecies of P. bedriagae is proposed.     

Statistics for phylogenetic trees

This paper poses the problem of estimating and validating phylogenetic trees in statistical terms. The problem is hard enough to warrant several tacks: we reason by analogy to rounding real numbers, and dealing with ranking data. These are both cases where, as in phylogeny the parameters of interest are not real numbers. Then we pose the problem in geometrical terms, using distances and measures on a natural space of trees. We do not solve the problems of inference on tree space, but suggest some coherent ways of tackling them.     

fastDNAml: a tool for construction of phylogenetic trees of DNA sequences using maximum likelihood

We have developed a new tool, called fastDNAml, for constructingphylogenetic trees from DNA sequences. The program can be runon a wide variety of computers ranging from Unix workstationsto massively parallel systems, and is available from the RibosomalDatabase Project (RDP) by anonymous FTP. Our program uses amaximum likelihood approach and is based on version 3.3 of Felsenstein'sdnaml program. Several enhancements, including algorithmic changes,significantly improve performance and reduce memory usage, makingit feasible to construct even very large trees. Trees containing40–100 taxa have been easily generated, and phylogeneticestimates are possible even when hundreds of sequences exist.We are currently using the tool to construct a phylogenetictree based on 473 small subunit rRNA sequences from prokaryotes.     

Evolutionary models of phylogenetic trees

The most widely used evolutionary model for phylogenetic trees is the equal-rates Markov (ERM) model. A problem is that the ERM model predicts less imbalance than observed for trees inferred from real data; in fact, the observed imbalance tends to fall between the values predicted by the ERM model and those predicted by the proportional-to-distinguishable-arrangements (PDA) model. Here, a continuous multi-rate (MR) family of evolutionary models is presented which contains entire subfamilies corresponding to both the PDA and ERM models. Furthermore, this MR family covers an entire range from 'completely balanced' to 'completely unbalanced' models. In particular, the MR family contains other known evolutionary models. The MR family is very versatile and virtually free of assumptions on the character of evolution; yet it is highly susceptible to rigorous analyses. In particular, such analyses help to uncover adaptability, quasi-stabilization and prolonged stasis as major possible causes of the imbalance. However, the MR model is functionally simple and requires only three parameters to reproduce the observed imbalance.     

Techniques for the verification of minimal phylogenetic trees illustrated with ten mammalian haemoglobin sequences.

We have recently reported a method to identify the shortest possible phylogenetic tree for a set of protein sequences [Foulds Hendy & Penny (1979) J. Mol. Evol. 13. 127--150; Foulds, Penny & Hendy (1979) J. Mol. Evol. 13, 151--166]. The present paper discusses issues that arise during the construction of minimal phylogenetic trees from protein-sequence data. The conversion of the data from amino acid sequences into nucleotide sequences is shown to be advantageous. A new variation of a method for constructing a minimal tree is presented. Our previous methods have involved first constructing a tree and then either proving that it is minimal or transforming it into a minimal tree. The approach presented in the present paper progressively builds up a tree, taxon by taxon. We illustrate this approach by using it to construct a minimal tree for ten mammalian haemoglobin alpha-chain sequences. Finally we define a measure of the complexity of the data and illustrate a method to derive a directed phylogenetic tree from the minimal tree.     

Secator: a program for inferring protein subfamilies from phylogenetic trees

With the huge increase of protein data, an important problem is to estimate, within a large protein family, the number of sensible subsets for subsequent in-depth structural, functional, and evolutionary analyses. To tackle this problem, we developed a new program, Secator, which implements the principle of an ascending hierarchical method using a distance matrix based on a multiple alignment of protein sequences. Dissimilarity values assigned to the nodes of a deduced phylogenetic tree are partitioned by a new stopping rule introduced to automatically determine the significant dissimilarity values. The quality of the clusters obtained by Secator is verified by a separate Jackknife study. The method is demonstrated on 24 large protein families covering a wide spectrum of structural and sequence conservation and its usefulness and accuracy with real biological data is illustrated on two well-studied protein families (the Sm proteins and the nuclear receptors).     

Multipolar consensus for phylogenetic trees

Collections of phylogenetic trees are usually summarized using consensus methods. These methods build a single tree, supposed to be representative of the collection. However, in the case of heterogeneous collections of trees, the resulting consensus may be poorly resolved (strict consensus, majority-rule consensus, ...), or may perform arbitrary choices among mutually incompatible clades, or splits (greedy consensus). Here, we propose an alternative method, which we call the multipolar consensus (MPC). Its aim is to display all the splits having a support above a predefined threshold, in a minimum number of consensus trees, or poles. We show that the problem is equivalent to a graph-coloring problem, and propose an implementation of the method. Finally, we apply the MPC to real data sets. Our results indicate that, typically, all the splits down to a weight of 10% can be displayed in no more than 4 trees. In addition, in some cases, biologically relevant secondary signals, which would not have been present in any of the classical consensus trees, are indeed captured by our method, indicating that the MPC provides a convenient exploratory method for phylogenetic analysis. The method was implemented in a package freely available at http://www.lirmm.fr/~cbonnard/MPC.html     

Bayesian phylogenetic analysis reveals two-domain topology of S-adenosylhomocysteine hydrolase protein sequences

S-Adenosylhomocysteine hydrolase (SahH) is involved in the degradation of the compound which inhibits methylation reactions. Using a Bayesian approach and other methods, we reconstructed a phylogenetic tree of amino acid sequences of this protein originating from all three major domains of living organisms. The SahH sequences formed two major branches: one composed mainly of Archaea and the other of eukaryotes and majority of bacteria, clearly contradicting the three-domain topology shown by small subunit rRNA gene. This topology suggests the occurrence of lateral transfer of this gene between the domains. Poor resolution of eukaryotes and bacteria excluded an ultimate conclusion in which out of the two domains this gene appeared first, however, the congruence of the secondary branches with SS rRNA and/or concatenated ribosomal protein datasets phylogenies suggested an "early" acquisition by some bacterial and eukaryotic phyla. Similarly, the branching pattern of Archaea reflected the phylogenies shown by SS rRNA and ribosomal proteins. SahH is widespread in Eucarya, albeit, due to reductive evolution, it is missing in the intracellular parasite Encephalitozoon cuniculi. On the other hand, the lack of affinity to the sequences from the alpha-Proteobacteria and cyanobacteria excludes a possibility of its acquisition in the course of mitochondrial or chloroplast endosymbioses. Unlike Archaea, most bacteria carry MTA/SAH nucleosidase, an enzyme involved also in metabolism of methylthioadenosine. However, the double function of MTA/SAH nucleosidase may be a barrier to ensure the efficient degradation of S-adenosylhomocysteine, specially when the intensity of methylation processes is high. This would explain the presence of S-adenosylhomocysteine hydrolase in the bacteria that have more complex metabolism. On the other hand, majority of obligate pathogenic bacteria due to simpler metabolism rely entirely on MTA/SAH nucleosidase. This could explain the observed phenetic pattern in which bacteria with larger (>6 Mb-million base pairs) genomes carry SAH hydrolase, whereas bacteria that have undergone reductive evolution usually carry MTA/SAH nucleosidase. This suggests that the presence or acquisition of S-adenosylhomocysteine hydrolase in bacteria may predispose towards higher metabolic, and in consequence, higher genomic complexity. The good examples are the phototrophic bacteria all of which carry this gene, however, the SahH phylogeny shows lack of congruence with SSU rRNA and photosyntethic genes, implying that the acquisition was independent and presumably preceded the acquisition of photosyntethic genes. The majority of cyanobacteria acquired this gene from Archaea, however, in some species the sahH gene was replaced by a copy from the beta- or gamma-Proteobacteria.     

Constraining prior probabilities of phylogenetic trees

Although Bayesian methods are widely used in phylogenetic systematics today, the foundations of this methodology are still debated among both biologists and philosophers. The Bayesian approach to phylogenetic inference requires the assignment of prior probabilities to phylogenetic trees. As in other applications of Bayesian epistemology, the question of whether there is an objective way to assign these prior probabilities is a contested issue. This paper discusses the strategy of constraining the prior probabilities of phylogenetic trees by means of the Principal Principle. In particular, I discuss a proposal due to Velasco (Biol Philos 23:455–473, 2008) of assigning prior probabilities to tree topologies based on the Yule process. By invoking the Principal Principle I argue that prior probabilities of tree topologies should rather be assigned a weighted mixture of probability distributions based on Pinelis’ (P Roy Soc Lond B Bio 270:1425–1431, 2003) multi-rate branching process including both the Yule distribution and the uniform distribution. However, I argue that this solves the problem of the priors of phylogenetic trees only in a weak form.     

Using genetic algorithms for the construction of phylogenetic trees: application to G-protein coupled receptor sequences

Many different phylogenetic clustering techniques are used currently. One approach is to first determine the topology with a common clustering method and then calculate the branch lengths of the tree. If the resulting tree is not optimal exchanging tree branches can make some local changes in the tree topology. The whole process can be iterated until a satisfactory result has been obtained. The efficiency of this method fully depends on the initially generated tree. Although local changes are made, the optimal tree will never be found if the initial tree is poorly chosen. In this article, genetic algorithms are applied such that the optimal tree can be found even with a bad initial tree topology. This tree generating method is tested by comparing its results with the results of the FITCH program in the PHYLIP software package. Two simulated data sets and a real data set are used.     

The prior probabilities of phylogenetic trees

Bayesian methods have become among the most popular methods in phylogenetics, but theoretical opposition to this methodology remains. After providing an introduction to Bayesian theory in this context, I attempt to tackle the problem mentioned most often in the literature: the “problem of the priors”—how to assign prior probabilities to tree hypotheses. I first argue that a recent objection—that an appropriate assignment of priors is impossible—is based on a misunderstanding of what ignorance and bias are. I then consider different methods of assigning prior probabilities to trees. I argue that priors need to be derived from an understanding of how distinct taxa have evolved and that the appropriate evolutionary model is captured by the Yule birth–death process. This process leads to a well-known statistical distribution over trees. Though further modifications may be necessary to model more complex aspects of the branching process, they must be modifications to parameters in an underlying Yule model. Ignoring these Yule priors commits a fallacy leading to mistaken inferences both about the trees themselves and about macroevolutionary processes more generally.     

The Shapley value of phylogenetic trees

Every weighted tree corresponds naturally to a cooperative game that we call a tree game; it assigns to each subset of leaves the sum of the weights of the minimal subtree spanned by those leaves. In the context of phylogenetic trees, the leaves are species and this assignment captures the diversity present in the coalition of species considered. We consider the Shapley value of tree games and suggest a biological interpretation. We determine the linear transformation M that shows the dependence of the Shapley value on the edge weights of the tree, and we also compute a null space basis of M. Both depend on the split counts of the tree. Finally, we characterize the Shapley value on tree games by four axioms, a counterpart to Shapley’s original theorem on the larger class of cooperative games. We also include a brief discussion of the core of tree games. Research of Francis Edward Su was partially supported by NSF Grants DMS-0301129 and DMS-0701308.