Thomas R. Buckley

Model selection and model averaging in phylogenetics : Advantages of Akaike Information Criterion and Bayesian approaches over likelihood ratio tests

Model selection is a topic of special relevance in molecular phylogenetics that affects many, if not all, stages of phylogenetic inference. Here we discuss some fundamental concepts and techniques of model selection in the context of phylogenetics. We start by reviewing different aspects of the selection of substitution models in phylogenetics from a theoretical, philosophical and practical point of view, and summarize this comparison in table format. We argue that the most commonly implemented model selection approach, the hierarchical likelihood ratio test, is not the optimal strategy for model selection in phylogenetics, and that approaches like the Akaike Information Criterion (AIC) and Bayesian methods offer important advantages. In particular, the latter two methods are able to simultaneously compare multiple nested or nonnested models, assess model selection uncertainty, and allow for the estimation of phylogenies and model parameters using all available models (model-averaged inference or multimodel inference). We also describe how the relative importance of the different parameters included in substitution models can be depicted. To illustrate some of these points, we have applied AIC-based model averaging to 37 mitochondrial DNA sequences from the subgenus Ohomopterus(genus Carabus) ground beetles described by Sota and Vogler (2001).

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.

Combined Data, Bayesian Phylogenetics, and the Origin of the New Zealand Cicada Genera

We have applied Bayesian and maximum likelihood methods of phylogenetic estimation to data from four mitochondrial genes (COI, COII, 12S, and 16S) and a single nuclear gene (EF1alpha) from several genera of New Zealand, Australian, and New Caledonian cicada taxa. We specifically focused on the heterogeneity of phylogenetic signal among the different data partitions and the biogeographic origins of the New Zealand cicada fauna. The Bayesian analyses circumvent many of the problems associated with other statistical tests for comparing data partitions. We took an information-theoretic approach to model selection based on the Akaike Information Criterion (AIC). This approach indicated that there was considerable uncertainty in identifying the best-fit model for some of the partitions. Additionally, a large amount of uncertainty was associated with many parameter estimates from the substitution model. However, a sensitivity analysis on the combined dataset indicated that the model selection uncertainty had little effect on estimates of topology because these estimates were largely insensitive to changes in the assumed model. This outcome suggests strong signal in our data. Our analyses support a New Caledonian affiliation of the New Zealand cicada genera Maoricicada, Kikihia, and Rhodopsalta and Australian affinities for the genera Amphipsalta and Notopsalta. This result was surprising, given that previous cicada biologists suspected a close relationship between Amphipsalta, Notopsalta, and Rhodopsalta based on genitalic characters. Relationships among the closely related genera Maoricicada, Kikihia, and Rhodopsalta were poorly resolved, the mitochondrial data and the EF1alpha data favoring different arrangements within this clade.

Differentiating between Hypotheses of Lineage Sorting and Introgression in New Zealand Alpine Cicadas (Maoricicada Dugdale)

Lineage sorting and introgression can lead to incongruence among gene phylogenies, complicating the inference of species trees for large groups of taxa that have recently and rapidly radiated. In addition, it can be difficult to determine which of these processes is responsible for this incongruence. We explore these issues with the radiation of New Zealand alpine cicadas of the genus Maoricicada Dugdale. Gene trees were estimated from four putative independent loci: mitochondrial DNA (2274 nucleotides), elongation factor 1-alpha (1275 nucleotides), period (1709 nucleotides), and calmodulin (678 nucleotides). We reconstructed phylogenies using maximum likelihood and Bayesian methods from 44 individuals representing the 19 species and subspecies of Maoricicada and two outgroups. Species-level relationships were reconstructed using a novel extension of gene tree parsimony, whereby gene trees were weighted by their Bayesian posterior probabilities. The inferred gene trees show marked incongruence in the placement of some taxa, especially the enigmatic forest and scrub dwelling species, M. iolanthe. Using the species tree estimated by gene tree parsimony, we simulated coalescent gene trees in order to test the null hypothesis that the nonrandom placement of M. iolanthe among gene trees has arisen by chance. Under the assumptions of constant population size, known generation time, and panmixia, we were able to reject this null hypothesis. Furthermore, because the two alternative placements of M. iolanthe are in each case with species that share a similar song structure, we conclude that it is more likely that an ancient introgression event rather than lineage sorting has caused this incongruence.

Improved resolution on the phylogenetic relationships among Pseudomonas by the combined analysis of atpD, carA, recA and 16S rDNA

A study of representatives of the bacterial genus Pseudomonas, analysing a combined data set of four molecular sequences with completely different properties and evolutionary constraints, is reported. The best evolutionary model was obtained with a hierarchical hypothesis testing program to describe each data set and the combined data set is presented and analysed under the likelihood criterion. The resolution among Pseudomonas taxa based on the combined data set analysis of the different lineages increased due to a synergistic effect of the individual data sets. The unresolved fluorescens lineage, as well as other weakly supported lineages in the single data set trees, should be revised in detail at the biochemical and molecular level. The taxonomic status of biovars of P. putida is discussed.

Extreme convergence in stick insect evolution: phylogenetic placement of the Lord Howe Island tree lobster

The ‘tree lobsters’ are an enigmatic group of robust, ground-dwelling stick insects (order Phasmatodea) from the subfamily Eurycanthinae, distributed in New Guinea, New Caledonia and associated islands. Its most famous member is the Lord Howe Island stick insect Dryococelus australis (Montrouzier), which was believed to have become extinct but was rediscovered in 2001 and is considered to be one of the rarest insects in the world. To resolve the evolutionary position of Dryococelus, we constructed a phylogeny from approximately 2.4 kb of mitochondrial and nuclear sequence data from representatives of all major phasmatodean lineages. Our data placed Dryococelus and the New Caledonian tree lobsters outside the New Guinean Eurycanthinae as members of an unrelated Australasian stick insect clade, the Lanceocercata. These results suggest a convergent origin of the ‘tree lobster’ body form. Our reanalysis of tree lobster characters provides additional support for our hypothesis of convergent evolution. We conclude that the phenotypic traits leading to the traditional classification are convergent adaptations to ground-living behaviour. Our molecular dating analyses indicate an ancient divergence (more than 22 Myr ago) between Dryococelus and its Australian relatives. Hence, Dryococelus represents a long-standing separate evolutionary lineage within the stick insects and must be regarded as a key taxon to protect with respect to phasmatodean diversity.

Determining the origin and age of the Westland beech (Nothofagus) gap, New Zealand, using fungus beetle genetics

The formation and maintenance of the Nothofagus beech gap in the South Island, New Zealand, has been the focus of biogeographical debate since the 1920s. We examine the historical process of gap formation by investigating the population genetics of fungus beetles: Brachynopus scutellaris (Staphylinidae) inhabits logs and is absent from the beech gap, and Hisparonia hystrix (Nitidulidae) is contiguous through the gap and is found commonly on sooty mould growing on several plant species. Both species show distinctive northern and southern haplotype distributions while H. hystrix recolonized the gap as shown by definitive mixing. B. scutellaris shows two major haplotype clades with strong geographical concordance, and unlike H. hystrix, has clearly defined lineages that can be partitioned for molecular dating. Based on coalescence dating methods, disjunct lineages of B. scutellaris indicate that the gap was formed less than 200 000 years ago. Phylogenetic imprints from both species reveal similar patterns of population divergence corresponding to recent glacial cycles, favouring a glacial explanation for the origin of the gap. Post‐gap colonization by H. hystrix may have been facilitated by the spread of Leptospermum scoparium host trees to the area, and they may be better at dispersing than B. scutellaris which may be constrained by fungal host and/or microhabitat. The gap‐excluded species B. scutellaris is found in both beech and podocarp‐broadleaf forests flanking the Westland gap and its absence in the gap may be related to incomplete recolonization following glacial retreat. We also discuss species status and an ancient polymorphism within B. scutellaris.

A living fossil tale of Pangaean biogeography

The current distributions of widespread groups of terrestrial animals and plants are supposedly the result of a mixture of either vicariance owing to continental split or more recent trans-oceanic dispersal. For organisms exhibiting a vicariant biogeographic pattern—achieving their current distribution by riding on the plates of former supercontinents—this view is largely inspired by the belief that Pangaea lacked geographical or ecological barriers, or that extinctions and dispersal would have erased any biogeographic signal since the early Mesozoic. We here present a time-calibrated molecular phylogeny of Onychophora (velvet worms), an ancient and exclusively terrestrial panarthropod group distributed throughout former Pangaean landmasses. Our data not only demonstrate that trans-oceanic dispersal does not need be invoked to explain contemporary distributions, but also reveal that the early diversification of the group pre-dates the break-up of Pangaea, maintaining regionalization even in landmasses that have remained contiguous throughout the history of the group. These results corroborate a growing body of evidence from palaeontology, palaeogeography and palaeoclimatic modelling depicting ancient biogeographic regionalization over the continuous landmass of Pangaea.

Identifying glacial refugia in a geographic parthenogen using palaeoclimate modelling and phylogeography: the New Zealand stick insect Argosarchus horridus (White)

We have used phylogeographic analysis of mitochondrial DNA (COI and COII genes) and ecological niche modelling (ENM) to reconstruct the population history of Argosarchus horridus (White), a widespread species of New Zealand stick insect. These data were used to address outstanding questions on the role of glacial refugia in determining the distribution and genetic structure of New Zealand species. Phylogeographic analysis shows a general pattern of high diversity in upper North Island and reduced diversity in lower North Island and South Island. The ENM indicates that during the last glacial maximum, A. horridus was largely restricted to refugia around coastal areas of North Island. The ENM also suggests refugia on the northeast coast of South Island and southeast coast of North Island and this prediction is verified by phylogeographic analysis, which shows a clade restricted to this region. Argosarchus horridus is also most likely a geographic parthenogen where males are much rarer at higher latitudes. The higher levels of genetic variation in northern, bisexual populations suggest southern and largely unisexual populations originated from southwardly expanding parthenogenetic lineages. Bayesian skyline analysis also provides support for a recent population size increase consistent with a large increase in geographic distribution in the late Pleistocene. These results exemplify the utility of integrating ENM and phylogeographic analysis in testing hypotheses on the origin of geographic parthenogenesis and effects of Pleistocene environmental change on biodiversity.

Phylogenetic analysis of New Zealand earthworms (Oligochaeta: Megascolecidae) reveals ancient clades and cryptic taxonomic diversity.

We have constructed the first ever phylogeny for the New Zealand earthworm fauna (Megascolecinae and Acanthodrilinae) including representatives from other major continental regions. Bayesian and maximum likelihood phylogenetic trees were constructed from 427 base pairs from the mitochondrial large subunit (16S) rRNA gene and 661 base pairs from the nuclear large subunit (28S) rRNA gene. Within the Acanthodrilinae we were able to identify a number of well-supported clades that were restricted to continental landmasses. Estimates of nodal support for these major clades were generally high, but relationships among clades were poorly resolved. The phylogenetic analyses revealed several independent lineages in New Zealand, some of which had a comparable phylogenetic depth to monophyletic groups sampled from Madagascar, Africa, North America and Australia. These results are consistent with at least some of these clades having inhabited New Zealand since rifting from Gondwana in the Late Cretaceous. Within the New Zealand Acanthodrilinae, major clades tended to be restricted to specific regions of New Zealand, with the central North Island and Cook Strait representing major biogeographic boundaries. Our field surveys of New Zealand and subsequent identification has also revealed extensive cryptic taxonomic diversity with approximately 48 new species sampled in addition to the 199 species recognized by previous authors. Our results indicate that further survey and taxonomic work is required to establish a foundation for future biogeographic and ecological research on this vitally important component of the New Zealand biota.