Navigating the unknown: model selection in phylogeography. Models of population structure: tools for thinkers

Despite the widespread use and obvious strengths of model-based methods for phylogeographic study, a persistent concern for such analyses is related to the definition of the model itself. The study by Peter et al. (2010) in this issue of Molecular Ecology demonstrates an approach for overcoming such hurdles. The authors were motivated by a deceptively simple goal; they sought to infer whether a population has remained at a low and stable size or has undergone a decline, and certainly there is no shortage of software packages for such a task (e.g., see list of programs in Excoffier & Heckel 2006). However, each of these software packages makes basic assumptions about the underling population (e.g., is the population subdivided or panmictic); these assumptions are explicit to any model-based approach but can bias parameter estimates and produce misleading inferences if the model does not approximate the actual demographic history in a reasonable manner. Rather than guessing which model might be best for analyzing the data (microsatellite data from samples of chimpanzees), Peter et al. (2010) quantify the relative fit of competing models for estimating the population genetic parameters of interest. Complemented by a revealing simulation study, the authors highlight the peril inherent to model-based inferences that lack a statistical evaluation of the fit of a model to the data, while also demonstrating an approach for model selection with broad applicability to phylogeographic analysis.     

Complex population dynamics and the coalescent under neutrality

Estimates of the coalescent effective population size N(e) can be poorly correlated with the true population size. The relationship between N(e) and the population size is sensitive to the way in which birth and death rates vary over time. The problem of inference is exacerbated when the mechanisms underlying population dynamics are complex and depend on many parameters. In instances where nonparametric estimators of N(e) such as the skyline struggle to reproduce the correct demographic history, model-based estimators that can draw on prior information about population size and growth rates may be more efficient. A coalescent model is developed for a large class of populations such that the demographic history is described by a deterministic nonlinear dynamical system of arbitrary dimension. This class of demographic model differs from those typically used in population genetics. Birth and death rates are not fixed, and no assumptions are made regarding the fraction of the population sampled. Furthermore, the population may be structured in such a way that gene copies reproduce both within and across demes. For this large class of models, it is shown how to derive the rate of coalescence, as well as the likelihood of a gene genealogy with heterochronous sampling and labeled taxa, and how to simulate a coalescent tree conditional on a complex demographic history. This theoretical framework encapsulates many of the models used by ecologists and epidemiologists and should facilitate the integration of population genetics with the study of mathematical population dynamics.     

Incorporating experimental design and error into coalescent/mutation models of population history

Coalescent theory provides a powerful framework for estimating the evolutionary, demographic, and genetic parameters of a population from a small sample of individuals. Current coalescent models have largely focused on population genetic factors (e.g., mutation, population growth, and migration) rather than on the effects of experimental design and error. This study develops a new coalescent/mutation model that accounts for unobserved polymorphisms due to missing data, sequence errors, and multiple reads for diploid individuals. The importance of accommodating these effects of experimental design and error is illustrated with evolutionary simulations and a real data set from a population of the California sea hare. In particular, a failure to account for sequence errors can lead to overestimated mutation rates, inflated coalescent times, and inappropriate conclusions about the population. This current model can now serve as a starting point for the development of newer models with additional experimental and population genetic factors. It is currently implemented as a maximum-likelihood method, but this model may also serve as the basis for the development of Bayesian approaches that incorporate experimental design and error.     

Amplicon sequencing of 42 nuclear loci supports directional gene flow between South Pacific populations of a hydrothermal vent limpet

In the past few decades, population genetics and phylogeographic studies have improved our knowledge of connectivity and population demography in marine environments. Studies of deep‐sea hydrothermal vent populations have identified barriers to gene flow, hybrid zones, and demographic events, such as historical population expansions and contractions. These deep‐sea studies, however, used few loci, which limit the amount of information they provided for coalescent analysis and thus our ability to confidently test complex population dynamics scenarios. In this study, we investigated population structure, demographic history, and gene flow directionality among four Western Pacific hydrothermal vent populations of the vent limpet Lepetodrilus aff. schrolli. These vent sites are located in the Manus and Lau back‐arc basins, currently of great interest for deep‐sea mineral extraction. A total of 42 loci were sequenced from each individual using high‐throughput amplicon sequencing. Amplicon sequences were analyzed using both genetic variant clustering methods and evolutionary coalescent approaches. Like most previously investigated vent species in the South Pacific, L. aff. schrolli showed no genetic structure within basins but significant differentiation between basins. We inferred significant directional gene flow from Manus Basin to Lau Basin, with low to no gene flow in the opposite direction. This study is one of the very few marine population studies using >10 loci for coalescent analysis and serves as a guide for future marine population studies.     

Molecular evidence for a recent demographic expansion in the puma (Puma concolor) (Mammalia,Felidae)

The puma is an iconic predator that ranges throughout the Americas, occupying diverse habitats. Previous phylogeographic analyses have revealed that it exhibits moderate levels of genetic structure across its range, with few of the classically recognized subspecies being supported as distinct demographic units. Moreover, most of the species’ molecular diversity was found to be in South America. To further investigate the phylogeographic structure and demographic history of pumas we analyzed mtDNA sequences from 186 individuals sampled throughout their range, with emphasis on South America. Our objectives were to refine the phylogeographic assessment within South America and to investigate the demographic history of pumas using a coalescent approach. Our results extend previous phylogeographic findings, reassessing the delimitation of historical population units in South America and demonstrating that this species experienced a considerable demographic expansion in the Holocene, ca. 8,000 years ago. Our analyses indicate that this expansion occurred in South America, prior to the hypothesized re-colonization of North America, which was therefore inferred to be even more recent. The estimated demographic history supports the interpretation that pumas suffered a severe demographic decline in the Late Pleistocene throughout their distribution, followed by population expansion and re-colonization of the range, initiating from South America.     

It's about time: divergence, demography, and the evolution of developmental modes in marine invertebrates

Differences in larval developmental mode are predicted to affect ecological and evolutionary processes ranging from gene flow and population bottlenecks to rates of population recovery from anthropogenic disturbance and capacity for local adaptation. The most powerful tests of these predictions use comparisons among species to ask how phylogeographic patterns are correlated with the evolution and loss of prolonged planktonic larval development. An important and largely untested assumption of these studies is that interspecific differences in population genetic structure are mainly caused by differences in dispersal and gene flow (rather than by differences in divergence times among populations or changes in effective population sizes), and that species with similar patterns of spatial genetic variation have similar underlying temporal demographic histories. Teasing apart these temporal and spatial patterns is important for understanding the causes and consequences of evolutionary changes in larval developmental mode. New analytical methods that use the coalescent history of allelic diversity can reveal these temporal patterns, test the strength of traditional population-genetic explanations for variation in spatial structure based on differences in dispersal, and identify strongly supported alternative explanations for spatial structure based on demographic history rather than on gene flow alone. We briefly review some of these recent analytical developments, and show their potential for refining ideas about the correspondence between the evolution of larval developmental mode, population demographic history, and spatial genetic variation.     

Mid‐Holocene decline in African buffalos inferred from Bayesian coalescent‐based analyses of microsatellites and mitochondrial DNA

Genetic studies concerned with the demographic history of wildlife species can help elucidate the role of climate change and other forces such as human activity in shaping patterns of divergence and distribution. The African buffalo (Syncerus caffer) declined dramatically during the rinderpest pandemic in the late 1800s, but little is known about the earlier demographic history of the species. We analysed genetic variation at 17 microsatellite loci and a 302‐bp fragment of the mitochondrial DNA control region to infer past demographic changes in buffalo populations from East Africa. Two Bayesian coalescent‐based methods as well as traditional bottleneck tests were applied to infer detailed dynamics in buffalo demographic history. No clear genetic signature of population declines related to the rinderpest pandemic could be detected. However, Bayesian coalescent modelling detected a strong signal of African buffalo population declines in the order of 75–98%, starting in the mid‐Holocene (approximately 3–7000 years ago). The signature of decline was remarkably consistent using two different coalescent‐based methods and two types of molecular markers. Exploratory analyses involving various prior assumptions did not seriously affect the magnitude or timing of the inferred population decline. Climate data show that tropical Africa experienced a pronounced transition to a drier climate approximately 4500 years ago, concurrent with the buffalo decline. We therefore propose that the mid‐Holocene aridification of East Africa caused a major decline in the effective population size of the buffalo, a species reliant on moist savannah habitat for its existence.     

Divergence genetics analysis reveals historical population genetic processes leading to contrasting phylogeographic patterns in co-distributed species

Coalescent samplers are computational time machines for inferring the historical demographic genetic processes that have given rise to observable patterns of spatial genetic variation among contemporary populations. We have used traditional characterizations of population structure and coalescent‐based inferences about demographic processes to reconstruct the population histories of two co‐distributed marine species, the frilled dog whelk, Nucella lamellosa, and the bat star, Patiria miniata. Analyses of population structure were consistent with previous work in both species except that additional samples of N. lamellosa showed a larger regional genetic break on Vancouver Island (VI) rather than between the southern Alexander Archipelago as in P. miniata. Our understanding of the causes, rather than just the patterns, of spatial genetic variation was dramatically improved by coalescent analyses that emphasized variation in population divergence times. Overall, gene flow was greater in bat stars (planktonic development) than snails (benthic development) but spatially homogeneous within species. In both species, these large phylogeographic breaks corresponded to relatively ancient divergence times between populations rather than regionally restricted gene flow. Although only N. lamellosa shows a large break on VI, population separation times on VI are congruent between species, suggesting a similar response to late Pleistocene ice sheet expansion. The absence of a phylogeographic break in P. miniata on VI can be attributed to greater gene flow and larger effective population size in this species. Such insights put the relative significance of gene flow into a more comprehensive historical biogeographic context and have important implications for conservation and landscape genetic studies that emphasize the role of contemporary gene flow and connectivity in shaping patterns of population differentiation.     

On the independent gene trees assumption in phylogenomic studies

Multilocus coalescent methods for inferring species trees or historical demographic parameters typically require the assumption that gene trees for sampled SNPs or DNA sequence loci are conditionally independent given their species tree. In practice, researchers have used different criteria to delimit “independent loci.” One criterion identifies sampled loci as being independent of each other if they undergo Mendelian independent assortment (IA criterion). O'Neill et al. (2013, Molecular Ecology, 22, 111–129) used this approach in their phylogeographic study of North American tiger salamander species complex. In two other studies, researchers developed a pair of related methods that employ an independent genealogies criterion (IG criterion), which considers the effects of population‐level recombination on correlations between the gene trees of intrachromosomal loci. Here, I explain these three methods, illustrate their use with example data, and evaluate their efficacies. I show that the IA approach is more conservative, is simpler to use and requires fewer assumptions than the IG approaches. However, IG approaches can identify much larger numbers of independent loci than the IA method, which, in turn, allows researchers to obtain more precise and accurate estimates of species trees and historical demographic parameters. A disadvantage of the IG methods is that they require an estimate of the population recombination rate. Despite their drawbacks, IA and IG approaches provide molecular ecologists with promising a priori methods for selecting SNPs or DNA sequence loci that likely meet the independence assumption in coalescent‐based phylogenomic studies.     

Backward simulation of ancestors of sampled individuals

If the population is large and the sampling mechanism is random, the coalescent is commonly used to model the haplotypes in the sample. Ordered genotypes can then be formed by random matching of the derived haplotypes. However, this approach is not realistic when (1) there is departure from random mating (e.g., dominant individuals in breeding populations or monogamy in humans), or (2) the population is small and/or the individuals in the sample are ascertained by applying some particular non-random sampling scheme, as is usually the case when considering the statistical modeling and analysis of pedigree data. For such situations, we present here a data generation method where an ancestral graph with non-overlapping generations is first generated backwards in time, using ideas from coalescent theory. Alleles are randomly assigned to the founders, and subsequently the gene flow over the entire genome is simulated forwards in time by dropping alleles down the graph according to recombination model without interference. The parameters controlling the mating behavior of generated individuals in the graph (degree of monogamy) can be tuned in order to match a particular demographic situation, without restriction to simple random mating.The performance of the approach is illustrated with a simulation example. The software (written in C-language) is freely available for research purposes at http://www.rni.helsinki.fi/∼dag/.     

Comparative Phylogeography Highlights the Double-Edged Sword of Climate Change Faced by Arctic- and Alpine-Adapted Mammals

Recent studies suggest that alpine and arctic organisms may have distinctly different phylogeographic histories from temperate or tropical taxa, with recent range contraction into interglacial refugia as opposed to post-glacial expansion out of refugia. We use a combination of phylogeographic inference, demographic reconstructions, and hierarchical Approximate Bayesian Computation to test for phylodemographic concordance among five species of alpine-adapted small mammals in eastern Beringia. These species (Collared Pikas, Hoary Marmots, Brown Lemmings, Arctic Ground Squirrels, and Singing Voles) vary in specificity to alpine and boreal-tundra habitat but share commonalities (e.g., cold tolerance and nunatak survival) that might result in concordant responses to Pleistocene glaciations. All five species contain a similar phylogeographic disjunction separating eastern and Beringian lineages, which we show to be the result of simultaneous divergence. Genetic diversity is similar within each haplogroup for each species, and there is no support for a post-Pleistocene population expansion in eastern lineages relative to those from Beringia. Bayesian skyline plots for four of the five species do not support Pleistocene population contraction. Brown Lemmings show evidence of late Quaternary demographic expansion without subsequent population decline. The Wrangell-St. Elias region of eastern Alaska appears to be an important zone of recent secondary contact for nearctic alpine mammals. Despite differences in natural history and ecology, similar phylogeographic histories are supported for all species, suggesting that these, and likely other, alpine- and arctic-adapted taxa are already experiencing population and/or range declines that are likely to synergistically accelerate in the face of rapid climate change. Climate change may therefore be acting as a double-edged sword that erodes genetic diversity within populations but promotes divergence and the generation of biodiversity.     

DnaSAM: Software to perform neutrality testing for large datasets with complex null models

Patterns of DNA sequence polymorphisms can be used to understand the processes of demography and adaptation within natural populations. High-throughput generation of DNA sequence data has historically been the bottleneck with respect to data processing and experimental inference. Advances in marker technologies have largely solved this problem. Currently, the limiting step is computational, with most molecular population genetic software allowing a gene-by-gene analysis through a graphical user interface. An easy-to-use analysis program that allows both high-throughput processing of multiple sequence alignments along with the flexibility to simulate data under complex demographic scenarios is currently lacking. We introduce a new program, named DnaSAM, which allows high-throughput estimation of DNA sequence diversity and neutrality statistics from experimental data along with the ability to test those statistics via Monte Carlo coalescent simulations. These simulations are conducted using the ms program, which is able to incorporate several genetic parameters (e.g. recombination) and demographic scenarios (e.g. population bottlenecks). The output is a set of diversity and neutrality statistics with associated probability values under a user-specified null model that are stored in easy to manipulate text file.     

New Routes to Phylogeography: A Bayesian Structured Coalescent Approximation

Phylogeographic methods aim to infer migration trends and the history of sampled lineages from genetic data. Applications of phylogeography are broad, and in the context of pathogens include the reconstruction of transmission histories and the origin and emergence of outbreaks. Phylogeographic inference based on bottom-up population genetics models is computationally expensive, and as a result faster alternatives based on the evolution of discrete traits have become popular. In this paper, we show that inference of migration rates and root locations based on discrete trait models is extremely unreliable and sensitive to biased sampling. To address this problem, we introduce BASTA (BAyesian STructured coalescent Approximation), a new approach implemented in BEAST2 that combines the accuracy of methods based on the structured coalescent with the computational efficiency required to handle more than just few populations. We illustrate the potentially severe implications of poor model choice for phylogeographic analyses by investigating the zoonotic transmission of Ebola virus. Whereas the structured coalescent analysis correctly infers that successive human Ebola outbreaks have been seeded by a large unsampled non-human reservoir population, the discrete trait analysis implausibly concludes that undetected human-to-human transmission has allowed the virus to persist over the past four decades. As genomics takes on an increasingly prominent role informing the control and prevention of infectious diseases, it will be vital that phylogeographic inference provides robust insights into transmission history.     

Heterozygote deficiencies caused by a Wahlund effect: Dispelling unfounded expectations

Population genetic tools can facilitate successful conservation and management of wildlife populations. However, the ability of such approaches to inform wildlife management and conservation programs depends upon assumptions linking genetic patterns to ecological processes, one implicit assumption usually being that genetic parameters (e.g., population genetic differentiation) estimated using a set of loci accurately reflect underlying demographic and microevolutionary forces affecting the population(s) under study. This is an important assumption because it also implies that we have acknowledged that genetic parameters estimated by a set of target loci inherently are associated with a sampling variance. Specifically, a perception exists that heterozygote deficits caused by biological mechanisms (e.g., a Wahlund effect) and null alleles can be differentiated by the expectation that the former leads to a concordant pattern across all loci, whereas the latter leads to locus-specific effects. We use Monte-Carlo simulation to demonstrate that these expectations do not always hold under biologically realistic conditions. Our analyses indicate that the conservative approach of discarding loci deviating from Hardy-Weinberg equilibrium expectations could rob us of our most informative markers, weakening our ability to interpret biological phenomena. © 2012 The Wildlife Society.     

Did Late Pleistocene climate change result in parallel genetic structure and demographic bottlenecks in sympatric Central African crocodiles,Mecistops and Osteolaemus?

The mid‐Holocene has had profound demographic impacts on wildlife on the African continent, although there is little known about the impacts on species from Central Africa. Understanding the impacts of climate change on codistributed species can enhance our understanding of ecosystem dynamics and for formulating restoration objectives. We took a multigenome comparative approach to examine the phylogeographic structure of two poorly known Central African crocodile species—Mecistops sp. aff. cataphractus and Osteolaemus tetraspis. In addition, we conducted coalescent‐based demographic reconstructions to test the hypothesis that population decline was driven by climate change since the Last Glacial Maximum, vs. more recent anthropogenic pressures. Using a hierarchical Bayesian model to reconstruct demographic history, we show that both species had dramatic declines (>97%) in effective population size in the 'period following the Last Glacial Maximum 1,500–18,000 YBP. Identification of genetic structuring showed both species have similar regional structure corresponding to major geological features (i.e., hydrologic basin) and that small observed differences between them are best explained by the differences in their ecology and the likely impact that climate change had on their habitat needs. Our results support our hypothesis that climatic effects, presumably on forest and wetland habitat, had a congruent negative impact on both species.     

Detecting bottlenecks and selective sweeps from DNA sequence polymorphism

A coalescence-based maximum-likelihood method is presented that aims to (i) detect diversity-reducing events in the recent history of a population and (ii) distinguish between demographic (e.g., bottlenecks) and selective causes (selective sweep) of a recent reduction of genetic variability. The former goal is achieved by taking account of the distortion in the shape of gene genealogies generated by diversity-reducing events: gene trees tend to be more star-like than under the standard coalescent. The latter issue is addressed by comparing patterns between loci: demographic events apply to the whole genome whereas selective events affect distinct regions of the genome to a varying extent. The maximum-likelihood approach allows one to estimate the time and strength of diversity-reducing events and to choose among competing hypotheses. An application to sequence data from an African population of Drosophila melanogaster shows that the bottleneck hypothesis is unlikely and that one or several selective sweeps probably occurred in the recent history of this population.     

Retrospective coalescent methods and the reconstruction of metapopulation histories in the sea

Phylogeographic analyses are a key interface between ecological and evolutionary ways of knowing because such analyses integrate the cumulative effects of demographic (ecological) processes over geological (evolutionary) time scales. Newly developed coalescent methods allow evolutionary ecologists to overcome some limitations associated with inferring population history from classic methods such as Wright’s F _ST. Here we briefly contrast classic and coalescent methods for looking backward in time through a population genetic lens, focusing on the key advantages of the isolation-with-migration (IM) class of coalescent methods for distinguishing ancient connectedness from actual recurrent contemporary gene flow as causes of genetic similarity or differentiation among populations. Making this critical distinction can lead to the discovery of otherwise obscured histories underlying conventional patterns of spatial variation. We illustrate the importance of these insights using analyses of Pacific fishes, snails, and sea stars in which population sizes and divergence times are more important than rates of contemporary gene flow as determinants of population genetic differentiation. We then extend the IM method to genetic data from two model metapopulation species (California abalone, Australian damselfish). The analyses show the potential use of non-equilibrium IM methods for differentiating among metapopulation models that make different predictions about population parameters and have different implications for the design of marine protected areas and other conservation goals. At face value, the results largely rule out classic metapopulation dynamics (dominated by extinction and colonization rather than connectivity via ongoing recurrent gene flow) but, at the same time, do not strongly support a modern marine metapopulation dynamic (ecologically significant connectivity between demes). However, the results also highlight the need for much more data (i.e., loci) sampled on different spatial scales in order to determine whether metapopulation dynamics might exist on smaller scales than are typically sampled by most phylogeographers and landscape geneticists.     

叶绿体DNA标记在谱系地理学中的应用研究进展

谱系地理学研究旨在探究历史上发生的影响目前遗传谱系系统发育和空间分布格局关系的生态与进化过程。叶绿体DNA具有单亲遗传、低突变率、单倍体等特征, 其分子标记不同程度地保留着植物长期进化的历史遗传痕迹, 有助于深度解析谱系地理变异的形成机制。本文探讨了上述特征是怎样影响分子标记的选择、扩大或缩小群体遗传结构分化、延长或缩短空间基因溯祖时间、促进或阻碍种间基因渐渗及谱系分选(复系、并系和单系形成)进程, 重点阐述了这些影响过程的理论基础, 并结合实际例子阐述谱系地理研究进展。由于位点间在突变率、选择强度及它们与漂变互作等方面存在异质性, 今后一个研究重点就是基于叶绿体全基因组序列分析谱系地理变化格局, 包括分析DNA位点间的基因渐渗或基因流动程度差异分布及沿着叶绿体DNA序列上谱系分选差异分布。     

Coalescent Times and Patterns of Genetic Diversity in Species with Facultative Sex: Effects of Gene Conversion,Population Structure,and Heterogeneity

Many diploid organisms undergo facultative sexual reproduction. However, little is currently known concerning the distribution of neutral genetic variation among facultative sexual organisms except in very simple cases. Understanding this distribution is important when making inferences about rates of sexual reproduction, effective population size, and demographic history. Here we extend coalescent theory in diploids with facultative sex to consider gene conversion, selfing, population subdivision, and temporal and spatial heterogeneity in rates of sex. In addition to analytical results for two-sample coalescent times, we outline a coalescent algorithm that accommodates the complexities arising from partial sex; this algorithm can be used to generate multisample coalescent distributions. A key result is that when sex is rare, gene conversion becomes a significant force in reducing diversity within individuals. This can reduce genomic signatures of infrequent sex (i.e., elevated within-individual allelic sequence divergence) or entirely reverse the predicted patterns. These models offer improved methods for assessing null patterns of molecular variation in facultative sexual organisms.     

Influence of spatial and temporal heterogeneities on the estimation of demographic parameters in a continuous population using individual microsatellite data

Drift and migration disequilibrium are very common in animal and plant populations. Yet their impact on methods of estimation of demographic parameters was rarely evaluated especially in complex realistic population models. The effect of such disequilibria on the estimation of demographic parameters depends on the population model, the statistics, and the genetic markers used. Here we considered the estimation of the product Dsigma2 from individual microsatellite data, where D is the density of adults and sigma2 the average squared axial parent-offspring distance in a continuous population evolving under isolation by distance. A coalescence-based simulation algorithm was used to study the effect on Dsigma2 estimation of temporal and spatial fluctuations of demographic parameters. Estimation of present-time Dsigma2 values was found to be robust to temporal changes in dispersal, to density reduction, and to spatial expansions with constant density, even for relatively recent changes (i.e., a few tens of generations ago). By contrast, density increase in the recent past gave Dsigma2 estimations biased largely toward past demographic parameters values. The method was also robust to spatial heterogeneity in density and estimated local demographic parameters when the density is homogenous around the sampling area (e.g., on a surface that equals four times the sampling area). Hence, in the limit of the situations studied in this article, and with the exception of the case of density increase, temporal and spatial fluctuations of demographic parameters appear to have a limited influence on the estimation of local and present-time demographic parameters with the method studied.