Explore the vast range of titles available.

In order to have confidence in model-based phylogenetic analysis, the model of nucleotide substitution adopted must be selected in a statistically rigorous manner. Several model-selection methods are applicable to maximum likelihood (ML) analysis, including the hierarchical likelihood-ratio test (hLRT), Akaike information criterion (AIC), Bayesian information criterion (BIC), and decision theory (DT), but their performance relative to empirical data has not been investigated thoroughly. In this study, we use 250 phylogenetic data sets obtained from TreeBASE to examine the effects that choice in model selection has on ML estimation of phylogeny, with an emphasis on optimal topology, bootstrap support, and hypothesis testing. We show that the use of different methods leads to the selection of two or more models for ∼ 80% of the data sets and that the AIC typically selects more complex models than alternative approaches. Although ML estimation with different best-fit models results in incongruent tree topologies ∼50% of the time, these differences are primarily attributable to alternative resolutions of poorly supported nodes. Furthermore, topologies and bootstrap values estimated with ML using alternative statistically supported models are more similar to each other than to topologies and bootstrap values estimated with ML under the Kimura two-parameter (K2P) model or maximum parsimony (MP). In addition, Swofford-Olsen-Waddell-Hillis (SOWH) tests indicate that ML trees estimated with alternative best-fit models are usually not significantly different from each other when evaluated with the same model. However, ML trees estimated with statistically supported models are often significantly suboptimal to ML trees made with the K2P model when both are evaluated with K2P, indicating that not all models perform in an equivalent manner. Nevertheless, the use of alternative statistically supported models generally does not affect tests of monophyletic relationships under either the Shimodaira-Hasegawa (S-H) or SOWH methods. Our results suggest that although choice in model selection has a strong impact on optimal tree topology, it rarely affects evolutionary inferences drawn from the data because differences are mainly confined to poorly supported nodes. Moreover, since ML with alternative best-fit models tends to produce more similar estimates of phylogeny than ML under the K2P model or MP, the use of any statistically based model-selection method is vastly preferable to forgoing the model-selection process altogether.

In order to have confidence in model-based phylogenetic methods, such as maximum likelihood (ML) and Bayesian analyses, one must use an appropriate model of molecular evolution identified using statistically rigorous criteria. Although model selection methods such as the likelihood ratio test and Akaike information criterion are widely used in the phylogenetic literature, model selection methods lack the ability to reject all models if they provide an inadequate fit to the data. There are two methods, however, that assess absolute model adequacy, the frequentist Goldman–Cox (GC) test and Bayesian posterior predictive simulations (PPSs), which are commonly used in conjunction with the multinomial log likelihood test statistic. In this study, we use empirical and simulated data to evaluate the adequacy of common substitution models using both frequentist and Bayesian methods and compare the results with those obtained with model selection methods. In addition, we investigate the relationship between model adequacy and performance in ML and Bayesian analyses in terms of topology, branch lengths, and bipartition support. We show that tests of model adequacy based on the multinomial likelihood often fail to reject simple substitution models, especially when the models incorporate among-site rate variation (ASRV), and normally fail to reject less complex models than those chosen by model selection methods. In addition, we find that PPSs often fail to reject simpler models than the GC test. Use of the simplest substitution models not rejected based on fit normally results in similar but divergent estimates of tree topology and branch lengths. In addition, use of the simplest adequate substitution models can affect estimates of bipartition support, although these differences are often small with the largest differences confined to poorly supported nodes. We also find that alternative assumptions about ASRV can affect tree topology, tree length, and bipartition support. Our results suggest that using the simplest substitution models not rejected based on fit may be a valid alternative to implementing more complex models identified by model selection methods. However, all common substitution models may fail to recover the correct topology and assign appropriate bipartition support if the true tree shape is difficult to estimate regardless of model adequacy.

Due to theoretical advancements and a rapid increase in computing power, phylogenetic inference has become a primarily statistical endeavor. Methods such as maximum-likelihood (ML) and Bayesian Markov chain Monte Carlo (Bayesian MCMC) analysis utilize explicit models of molecular evolution. Because confidence in the accuracy of ML and Bayesian MCMC analysis depends on using an appropriate model of sequence evolution, models must adequately capture the evolutionary process and be selected in a statistically rigorous manner. The research conducted for this dissertation focuses on the statistical evaluation of common substitution models and model-based phylogenetic methods. In Chapter I, we investigate the effects of using four model-selection methods on ML analysis. We find that although model-selection methods normally choose different substitution models for the same data and that use of alternative best-fit models often leads to different ML trees, use of any model-selection method is preferable to using a simple default model that does not incorporate among-site rate variation (ASRV). Although model-selection methods are useful, they cannot reject all models if they provide an inadequate fit to the data. In Chapter II, we evaluate the adequacy of common substitution models using frequentist and Bayesian methods and compare the results with four model-selection methods. We find that tests of model adequacy normally fail to reject less complex substitution models than those chosen by model-selection methods, especially when the models incorporate ASRV. Bipartition support in ML analysis is normally assessed using the nonparametric bootstrap. Although bootstrap replicates should be analyzed in the same manner as the original data, model selection is virtually never conducted for bootstrap replicates and replicates are often analyzed using a less rigorous tree-search strategy than the original data. In Chapter III, we investigate the effects of forgoing model-selection and using reduced tree-search strategies on ML bootstrap analysis and find that while using a less-intensive analysis may lead to significant differences among bootstrap values, these differences are normally small and would not change biological inferences drawn from the data.

Phylogeography of northern populations of the Pacific treefrog, Pseudacris regilla, was investigated using mitochondrial cytochrome b sequence data (725 bp). Thirty-six haplo-types were detected among 59 samples collected from 20 populations. Two divergent coastal and inland clades were supported by several phylogenetic analyses including maximum parsimony, maximum likelihood, and Bayesian methods. Sequence differences among these clades ranged from 5.0 to 6.5%, suggesting they diverged during the Pliocene (approximately 3 MYA), coinciding with High Cascade orogeny and subsequent xerification of the Columbia Basin. Further, haplotype divergence within each clade was lower (0 to 1.8%), possibly as a consequence of population reduction during the Pleistocene. The overall pattern of divergence was not detected by previous morphological and protein analysis and is concordant with many other Northwest taxa. These results do not support previous intraspecific classification schemes, indicating the need for further examination of the taxonomic status of the coastal and inland clades.

The International Monitoring System includes a network of radionuclide detectors operated around the world monitoring for nuclear explosions. A key aspect of the International Monitoring System is the detection of radioxenon with a network of stations and laboratories. Beta-gamma detectors are utilized extensively for the detection of radioxenon, and the beta detection is primarily performed with a plastic scintillator cell. Two areas of improvement for plastic scintillator are the sample carry-over (“memory effect”) and energy resolution. While the scintillator can be coated to remove the memory effect, the energy resolution must be improved with a different detector material. Silicon is the current leading candidate for the future beta cell material due to the much-improved energy resolution compared to plastic scintillators (factor of ~ 3x). PNNL is developing a silicon beta cell for use as a potential modular replacement within Xenon International (a next generation radioxenon detection system currently undergoing acceptance testing for potential inclusion in the International Monitoring System). The beta cell utilizes four different silicon detectors to create an active volume for the radioxenon within an outer gas cell. Since there are four separate beta signals (compared to one for plastic scintillators), data acquisition modifications are required. In this paper, we detail the design, efficiency measurements, and long-term testing of the silicon beta cell and potential improvements in isotopic discrimination.