Increased congruence does not necessarily indicate increased phylogenetic accuracy--the behavior of the incongruence length difference test in mixed-model analyses

Comprehensive phylogenetic analyses utilize data from distinct sources, including nuclear, mitochondrial, and plastid molecular sequences and morphology. Such heterogeneous datasets are likely to require distinct models of analysis, given the different histories of mutational biases operating on these characters. The incongruence length difference (ILD) test is increasingly being used to arbitrate between competing models of phylogenetic analysis in cases where multiple data partitions have been collected. Our work suggests that the ILD test is unlikely to be an effective measure of congruence when two datasets differ markedly in size. We show that models that increase the contribution of one data partition over another are likely to increase congruence, as measured by this test. More alarmingly, for many bipartition comparisons, character congruence increases bimodally - either increasing or decreasing the contribution of one data partition will increase congruence - making it impossible to arrive at a single optimally congruent model of analysis.     

Failure of the ILD to determine data combinability for slow loris phylogeny

Tests for incongruence as an indicator of among-data partition conflict have played an important role in conditional data combination. When such tests reveal significant incongruence, this has been interpreted as a rationale for not combining data into a single phylogenetic analysis. In this study of lorisiform phylogeny, we use the incongruence length difference (ILD) test to assess conflict among three independent data sets. A large morphological data set and two unlinked molecular data sets--the mitochondrial cytochrome b gene and the nuclear interphotoreceptor retinoid binding protein (exon 1)--are analyzed with various optimality criteria and weighting mechanisms to determine the phylogenetic relationships among slow lorises (Primates, Loridae). When analyzed separately, the morphological data show impressive statistical support for a monophyletic Loridae. Both molecular data sets resolve the Loridae as paraphyletic, though with different branching orders depending on the optimality criterion or character weighting used. When the three data partitions are analyzed in various combinations, an inverse relationship between congruence and phylogenetic accuracy is observed. Nearly all combined analyses that recover monophyly indicate strong data partition incongruence (P = 0.00005 in the most extreme case), whereas all analyses that recover paraphyly indicate lack of significant incongruence. Numerous lines of evidence verify that monophyly is the accurate phylogenetic result. Therefore, this study contributes to a growing body of information affirming that measures of incongruence should not be used as indicators of data set combinability.     

Let them fall where they may: congruence analysis in massive phylogenetically messy data sets

Interest in congruence in phylogenetic data has largely focused on issues affecting multicellular organisms, and animals in particular, in which the level of incongruence is expected to be relatively low. In addition, assessment methods developed in the past have been designed for reasonably small numbers of loci and scale poorly for larger data sets. However, there are currently over a thousand complete genome sequences available and of interest to evolutionary biologists, and these sequences are predominantly from microbial organisms, whose molecular evolution is much less frequently tree-like than that of multicellular life forms. As such, the level of incongruence in these data is expected to be high. We present a congruence method that accommodates both very large numbers of genes and high degrees of incongruence. Our method uses clustering algorithms to identify subsets of genes based on similarity of phylogenetic signal. It involves only a single phylogenetic analysis per gene, and therefore, computation time scales nearly linearly with the number of genes in the data set. We show that our method performs very well with sets of sequence alignments simulated under a wide variety of conditions. In addition, we present an analysis of core genes of prokaryotes, often assumed to have been largely vertically inherited, in which we identify two highly incongruent classes of genes. This result is consistent with the complexity hypothesis.     

On gaps.

Gaps result from the alignment of sequences of unequal length during primary homology assessment. Viewed as character states originating from particular biological events (mutation), gaps contain historical information suitable for phylogenetic analysis. The effect of gaps as a source of phylogenetic data is explored via sensitivity analysis and character congruence among different data partitions. Example data sets are provided to show that gaps contain important phylogenetic information not recovered by those methods that omit gaps in their calculations. However, gap cost schemes are arbitrary (although they must be explicit) and thus data exploration is a necessity of molecular analyses, while character congruence is necessary as an external criterion for hypothesis decision.     

Data congruence, paedomorphosis and salamanders

Background

The retention of ancestral juvenile characters by adult stages of descendants is called paedomorphosis. However, this process can mislead phylogenetic analyses based on morphological data, even in combination with molecular data, because the assessment if a character is primary absent or secondary lost is difficult. Thus, the detection of incongruence between morphological and molecular data is necessary to investigate the reliability of simultaneous analyses. Different methods have been proposed to detect data congruence or incongruence. Five of them (PABA, PBS, NDI, LILD, DRI) are used herein to assess incongruence between morphological and molecular data in a case study addressing salamander phylogeny, which comprises several supposedly paedomorphic taxa. Therefore, previously published data sets were compiled herein. Furthermore, two strategies ameliorating effects of paedomorphosis on phylogenetic studies were tested herein using a statistical rigor. Additionally, efficiency of the different methods to assess incongruence was analyzed using this empirical data set. Finally, a test statistic is presented for all these methods except DRI.

Results

The addition of morphological data to molecular data results in both different positions of three of the four paedomorphic taxa and strong incongruence, but treating the morphological data using different strategies ameliorating the negative impact of paedomorphosis revokes these changes and minimizes the conflict. Of these strategies the strategy to just exclude paedomorphic character traits seem to be most beneficial. Of the three molecular partitions analyzed herein the RAG1 partition seems to be the most suitable to resolve deep salamander phylogeny. The rRNA and mtDNA partition are either too conserved or too variable, respectively. Of the different methods to detect incongruence, the NDI and PABA approaches are more conservative in the indication of incongruence than LILD and PBS.

Conclusion

Paedomorphosis induces strong conflicts and can mislead the phylogenetic analyses even in combined analyses. However, different strategies are efficiently minimizing these problems. Though the exploration of different methods to detect incongruence is preferable NDI and PABA are more conservative than the others and NDI is computational less extensive than PABA.     

Measuring Topological Congruence by Extending Character Techniques

A measure of topological congruence which is an extension of the Mickevich–Farris character incongruence metric ( i.e. , ILD; Mickevich and Farris, 1981) is proposed. Group inclusion characters (1 = member of a clade; 0 = not a member) are constructed for each topology to be considered. The sets of characters derived from the topologies are then compared for character incongruence due to data set combination. Each homoplasy signifies a disagreement among topological statements. The value is normalized for potential maximum incongruence to adjust values for unresolved topologies. This measure is compared to other topological and character congruence techniques and explored in test data.     

Congruence of Morphological and Molecular Phylogenies

When phylogenetic trees constructed from morphological and molecular evidence disagree (i.e. are incongruent) it has been suggested that the differences are spurious or that the molecular results should be preferred a priori. Comparing trees can increase confidence (congruence), or demonstrate that at least one tree is incorrect (incongruence). Statistical analyses of 181 molecular and 49 morphological trees shows that incongruence is greater between than within the morphological and molecular partitions, and this difference is significant for the molecular partition. Because the level of incongruence between a pair of trees gives a minimum bound on how much error is present in the two trees, our results indicate that the level of error may be underestimated by congruence within partitions. Thus comparisons between morphological and molecular trees are particularly useful for detecting this incongruence (spurious or otherwise). Molecular trees have higher average congruence than morphological trees, but the difference is not significant, and both within- and between-partition incongruence is much lower than expected by chance alone. Our results suggest that both molecular and morphological trees are, in general, useful approximations of a common underlying phylogeny and thus, when molecules and morphology clash, molecular phylogenies should not be considered more reliable a priori.     

Data decisiveness,data quality,and incongruence in phylogenetic analysis: an example from the monocotyledons using mitochondrial atp A sequences

We examined three parallel data sets with respect to qualities relevant to phylogenetic analysis of 20 exemplar monocotyledons and related dicotyledons. The three data sets represent restriction-site variation in the inverted repeat region of the chloroplast genome, and nucleotide sequence variation in the chloroplast-encoded gene rbcL and in the mitochondrion-encoded gene atpA, the latter of which encodes the alpha-subunit of mitochondrial ATP synthase. The plant mitochondrial genome has been little used in plant systematics, in part because nucleotide sequence evolution in enzyme-encoding genes of this genome is relatively slow. The three data sets were examined in separate and combined analyses, with a focus on patterns of congruence, homoplasy, and data decisiveness. Data decisiveness (described by P. Goloboff) is a measure of robustness of support for most parsimonious trees by a data set in terms of the degree to which those trees are shorter than the average length of all possible trees. Because indecisive data sets require relatively fewer additional steps than decisive ones to be optimized on nonparsimonious trees, they will have a lesser tendency to be incongruent with other data sets. One consequence of this relationship between decisiveness and character incongruence is that if incongruence is used as a criterion of noncombinability, decisive data sets, which provide robust support for relationships, are more likely to be assessed as noncombinable with other data sets than are indecisive data sets, which provide weak support for relationships. For the sampling of taxa in this study, the atpA data set has about half as many cladistically informative nucleotides as the rbcL data set per site examined, and is less homoplastic and more decisive. The rbcL data set, which is the least decisive of the three, exhibits the lowest levels of character incongruence. Whatever the molecular evolutionary cause of this phenomenon, it seems likely that the poorer performance of rbcL than atpA, in terms of data decisiveness, is due to both its higher overall level of homoplasy and the fact that it is performing especially poorly at nonsynonymous sites.     

A priori assumptions about characters as a cause of incongruence between molecular and morphological hypotheses of primate interrelationships

When molecules and morphology produce incongruent hypotheses of primate interrelationships, the data are typically viewed as incompatible, and molecular hypotheses are often considered to be better indicators of phylogenetic history. However, it has been demonstrated that the choice of which taxa to include in cladistic analysis as well as assumptions about character weighting, character state transformation order, and outgroup choice all influence hypotheses of relationships and may positively influence tree topology, so that relationships between extant taxa are consistent with those found using molecular data. Thus, the source of incongruence between morphological and molecular trees may lie not in the morphological data themselves but in assumptions surrounding the ways characters evolve and their impact on cladistic analysis. In this study, we investigate the role that assumptions about character polarity and transformation order play in creating incongruence between primate phylogenies based on morphological data and those supported by multiple lines of molecular data. By releasing constraints imposed on published morphological analyses of primates from disparate clades and subjecting those data to parsimony analysis, we test the hypothesis that incongruence between morphology and molecules results from inherent flaws in morphological data. To quantify the difference between incongruent trees, we introduce a new method called branch slide distance (BSD). BSD mitigates many of the limitations attributed to other tree comparison methods, thus allowing for a more accurate measure of topological similarity. We find that releasing a priori constraints on character behavior often produces trees that are consistent with molecular trees. Case studies are presented that illustrate how congruence between molecules and unconstrained morphological data may provide insight into issues of polarity, transformation order, homology, and homoplasy.     

Phylogenetic analysis of Themira (Sepsidae: Diptera): sensitivity analysis, alignment, and indel treatment in a multigene study

In this study we use sensitivity analysis sensu Wheeler (1995 ) for a matrix entirely composed of DNA sequences. We propose that not only congruence but also phylogenetic structure, as measured by character resampling, should be used to choose among competing weighting regimes. An extensive analysis of a five‐gene data set for Themira (Sepsidae: Diptera) reveals that even with different ways of partitioning the data, measures of topological congruence, character incongruence, and phylogenetic structure favor similar weighting regimes involving the down‐weighting of transitions. We furthermore use sensitivity analysis for obtaining empirical evidence that allows us to select weights for third positions, deciding between treating indels as fifth character states or missing values, and choosing between manual and computational alignments. For our data, sensitivity analysis favors manual alignment over a Clustal‐generated numerical alignment, the treatment of indels as fifth character states over considering them missing values, and equal weights for all positions in protein‐encoding genes over the down‐weighting of third positions. Among the topological congruence measures compared, symmetric tree distance performed best. Partitioned Bremer Support analysis reveals that COI contributes the largest amount of support for our phylogenetic tree for Themira. © The Willi Hennig Society 2005.     

Sources of incongruence among mammalian mitochondrial sequences: COII, COIII, and ND6 genes are main contributors

To investigate the origins of incongruence among mammalian mitochondrial protein-coding genes, we compiled a matrix that included 13 protein-coding-genes for 41 mammals from 14 different orders. This matrix was examined for congruence using different partitioning strategies. The incongruence length difference test showed significant incongruence among the 13 gene partitions used simultaneously, and the result was not affected by third codon or transversion weighting. In the pair-wise comparisons, significant incongruence was detected between NADH:ubiquinone oxidoreductase subunit 6 gene (ND6), cytochrome oxidase subunit II (COII), or cytochrome oxidase subunit III (COIII) gene partitioned individually against the rest of the genes. Omission of any of the 14 mammalian orders alone or in combinations from the matrix did not result in a statistically significant improvement of congruence, suggesting that taxonomic sampling will not improve congruence among the data sets. However, omission of the ND6, COII, and COIII significantly improved congruence in our data matrix. Possible origins of unusual phylogenetic properties of the three genes are discussed.     

Homology and the Optimization of DNA Sequence Data

Three methods of nucleotide character analysis are discussed. Their implications for molecular sequence homology and phylogenetic analysis are compared. The criterion of inter-data set congruence, both character based and topological, are applied to two data sets to elucidate and potentially discriminate among these parsimony-based ideas.     

mILD: a tool for constructing and analyzing matrices of pairwise phylogenetic character incongruence tests

SUMMARY: Pairwise comparisons of disagreement in phylogenetic datasets offer a powerful tool for isolating historical incongruence for closer analysis. Statistically significant phylogenetic character incongruence may reflect important differences in evolutionary history, such as horizontal gene transfer. Such testing can also be used to specify possible combinations of datasets for further phylogenetic analysis. The process of comparing multiple datasets can be very time consuming, and it is sometimes unclear how to combine data partitions given the observed patterns of incongruence. Here we present an application that automates the process of making pairwise comparisons between large numbers of phylogenetic datasets using the Incongruence Length Difference (ILD) test. The application also implements strategies for data combination based on the patterns of incongruence observed in pairwise comparisons.     

Combining molecular and morphological analyses of water strider phylogeny (Hemiptera–Heteroptera, Gerromorpha): effects of alignment and taxon sampling

Abstract. The semiaquatic bugs (Hemiptera–Heteroptera, infraorder Gerromorpha), comprising water striders and their allies (c. 1900 described species), are familiar inhabitants of water surfaces in all continents. Recent fossil evidence indicates that the evolutionary history of semiaquatic bugs spans more than 120 million years of geological time. At present, our insight into the phylogeny of higher taxa is based upon Andersen's manual cladistic analysis of a suite of morphological characters. The present work expands the phylogenetic insight with numerical cladistic analyses of morphological and molecular datasets (partial sequences of 16S and 28S rDNA) for forty species of Gerromorpha covering most higher taxa (families, subfamilies), estimates of branch support, character incongruence, and topological congruence (nodal stability). For the molecular data we apply different alignment options (manual vs numerical alignment; multiple alignment vs direct optimization), treat insertion–deletion events (indels) as either missing data or as a fifth character state, subject the data to a sensitivity analysis, and estimate topological congruence between different analysis trees. Relationships change considerably under different analysis conditions, which means that there is little node stability, and for selecting preferred analysis conditions there is conflicting evidence from rescaled incongruence length difference and the key node criterion. Based on the analysis of the combined morphological and molecular datasets, this study supports the close relationship between the families Gerridae, Hermatobatidae and Veliidae (superfamily Gerroidea), but not the monophyly of the family Veliidae. The results suggest that the genus Ocellovelia (Ocelloveliinae) should be excluded from this family and placed as a sister group to Gerridae + the remaining species of Veliidae. Our study also supports a close relationship between the subfamilies Halobatinae and Ptilomerinae (Gerridae), and that the subfamily Veliinae is probably nonmonophyletic.     

Phylogenetic frameworks: towards a firmer foundation for the comparative approach

The growing interest in using phylogenies to test evolutionary hypotheses has focused attention on the need for robust estimates of phylogenetic history. Whether specific branching structures are correct summaries of evolutionary history can be estimated only through the examination of congruence of many sets of characters. After consideration of practical and philosophical aspects of congruence, I conclude that taxonomic congruence (analysis of congruence of topologies produced from independent datasets) is preferable to character congruence (analysis of congruence between individual characters) for estimating accuracy of phylogenetic hypotheses. Existing methods for examining taxonomic congruence are discussed and the combinable components approach, when preceded by application of rigorous statistical manipulations (e.g. jackknifing or bootstrapping), found most appropriate. Implementation of the method of combinable components is described, and is demonstrated using published data for Menidia and Rana. The robust branching structure resulting from this analysis (a phylogenetic framework) contains those nodes (phylogenetic hypotheses) that are strongly supported by at least one dataset and are consistent with all others. This approach is the most appropriate/conservative for testing hypotheses about evolutionary history.     

Molecules and Morphology,Phylogenetics and Genetics

Various explanations can be offered for the incongruence between phylogenetic hypotheses resulting from morphological and molecular data sets. Of these, the possibility that incongruence may result from the mutation of major morphogenetic genes leading to dramatic morphological divergence unaccompanied by equivalent change of the phylogenetic marker molecule(s) used is discussed in detail. As evidence for this hypothesis, several examples for such incongruence are surveyed. It seems possible that in many cases the genetic basis of the morphological characters responsible for the incongruence found may be simple, and that the genes involved may be homologous to genes known from mutant systems. It is suggested that: 1. the systematic documentation of incongruence between molecular and morphological phylogenies may help to assess the frequency of evolutionary change through the mutation of major morphogenetic genes, and that 2. the identification of major morphological characters distinguishing closely related taxa with mutant phenotypes known from mutant systems eventually may allow an experimental approach to the problem of evolutionary change resulting from major genes. Natural taxa suspected to be the result of such processes could be changed morphologically through transformation with the relevant genes.     

Detecting phylogenetic incongruence using BIONJ: an improvement of the ILD test

The problem of testing for congruence between phylogenetic data has long been debated among phylogeneticists, but reaches a critical point with the availability of large amount of biological sequences. Notably in prokaryotes, where the amount of lateral transfers is believed to be important, the inference of phylogenies using multiple genes requires testing for incongruence before concatenating the genes. On another scale, incongruence tests can be used to detect recombination points within single gene alignments. The incongruence length difference test (ILD), based on parsimony, has been proved to be useful for finding incongruent data sets, but its application remains limited to small data sets for computational time reasons. Here, we have adapted the principle of ILD to the BIONJ algorithm. This algorithm is based on a tree length minimisation criterion and is suitable to replace parsimony in this test when used with uncorrected distance (model-free approach). We show that this new test, ILD-BIONJ, while being much faster, is often more accurate than the ILD test, especially when the alignments compared are simulated under different evolutionary models.     

Phylogenetic signal in AFLP data sets

AFLP markers provide a potential source of phylogenetic information for molecular systematic studies. However, there are properties of restriction fragment data that limit phylogenetic interpretation of AFLPs. These are (a) possible nonindependence of fragments, (b) problems of homology assignment of fragments, (c) asymmetry in the probability of losing and gaining fragments, and (d) problems in distinguishing heterozygote from homozygote bands. In the present study, AFLP data sets of Lactuca s.l. were examined for the presence of phylogenetic signal. An indication of this signal was provided by carrying out tree length distribution skewness (g1) tests, permutation tail probability (PTP) tests, and relative apparent synapomorphy analysis (RASA). A measure of the support for internal branches in the optimal parsimony tree (MPT) was made using bootstrap, jackknife, and decay analysis. Finally, the extent of congruence in MPTs for AFLP and internal transcribed spacer (ITS)-1 data sets for the same taxa was made using the partition homogeneity test (PHT) and the Templeton test. These analytical studies suggested the presence of phylogenetic signal in the AFLP data sets, although some incongruence was found between AFLP and ITS MPTs. An extensive literature survey undertaken indicated that authors report a general congruence of AFLP and ITS tree topologies across a wide range of taxonomic groups, suggesting that the present results and conclusions have a general bearing. In these earlier studies and those for Lactuca s.l., AFLP markers have been found to be informative at somewhat lower taxonomic levels than ITS sequences. Tentative estimates are suggested for the levels of ITS sequence divergence over which AFLP profiles are likely to be phylogenetically informative.     

When is enough,enough in phylogenetics? A case in point from Hordeum (Poaceae)

Direct optimization was used to reconstruct the phylogeny of the 26 diploid taxa included in the genus Hordeum. The total data set was composed of 16 nucleotide sequence regions from the nuclear as well as the plastid genome. The nine nuclear regions were from single‐copy, protein coding genes located on six of the seven chromosome pairs in the diploid H. vulgare genome. The seven plastid regions comprise protein coding genes as well as intergenic regions. Studies of character congruence between data partitions showed no correlation between chromosomal location and congruence among the nuclear sequences and a level of congruence among the plastid sequences comparable with the level among the nuclear sequences. Combined analysis of all data resolved the phylogeny completely with most clades being robust and well supported. However, due to incongruence among data partitions some relationships are still and likely to remain ambiguously inferred. Rather than adding still more genes to the phylogenetic analyses, patterns of incongruence may be better explored by adding data from multiple specimens per taxon. For some species relationships the plastid data appear positively misleading, emphasizing the need for caution if plastid data are the only or dominant type of data used for phylogenetic reconstruction and subsequent re‐classification.
© The Willi Hennig Society 2011.     

Running WILD: the case for exploring mixed parameter sets in sensitivity analysis

The robustness of clades to parameter variation may be a desirable quality or even a goal in phylogenetic analyses. Sensitivity analyses used to assess clade stability have invoked the incongruence length difference (ILD or _WILD) metric, a measure of congruence among datasets, to compare a series of most‐parsimonious results from re‐running analyses under different analytical conditions. It is also common practice to select a single “optimal” parameter set that minimizes _WILD across all parameter sets. However, the divergent molecular evolution of ribosomal genes and protein‐encoding genes—specifically the bias against transversion events in coding genes of conserved function—suggests that deployment of multiple parameter sets could outperform the use of a single parameter set applied to all molecules. We explored congruence in five published datasets by including mixed parameter sets in our sensitivity analysis. In four cases, mixed parameter sets outperformed the previously reported, single optimal parameter set. Conversely, multiple parameter sets did not outperform a single optimal parameter set in a case in which actual strong topological conflict exists between data partitions. Exploration of mixed parameter sets may prove useful when combining ribosomal and protein‐encoding genes, due to the relatively higher frequency of single‐ and double‐base pair indel events in the former, and the relatively lower frequency of transversions in the latter.
© The Willi Hennig Society 2010.