Detecting selection along environmental gradients: analysis of eight methods and their effectiveness for outbreeding and selfing populations
TL;DR: In this paper, the authors provide guidelines for the use of popular or recently developed statistical methods to detect footprints of selection, and investigate the power and robustness of eight methods to identify loci potentially under selection.
Abstract: Thanks to genome-scale diversity data, present-day studies can provide a detailed view of how natural and cultivated species adapt to their environment and particularly to environmental gradients. However, due to their sensitivity, up-to-date studies might be more sensitive to undocumented demographic effects such as the pattern of migration and the reproduction regime. In this study, we provide guidelines for the use of popular or recently developed statistical methods to detect footprints of selection. We simulated 100 populations along a selective gradient and explored different migration models, sampling schemes and rates of self-fertilization. We investigated the power and robustness of eight methods to detect loci potentially under selection: three designed to detect genotype–environment correlations and five designed to detect adaptive differentiation (based on FST or similar measures). We show that genotype– environment correlation methods have substantially more power to detect selection than differentiation-based methods but that they generally suffer from high rates of false positives. This effect is exacerbated whenever allele frequencies are correlated, either between populations or within populations. Our results suggest that, when the underlying genetic structure of the data is unknown, a number of robust methods are preferable. Moreover, in the simulated scenario we used, sampling many populations led to better results than sampling many individuals per population. Finally, care should be taken when using methods to identify genotype–environment correlations without correcting for allele frequency autocorrelation because of the risk of spurious signals due to allele frequency correlations between populations.
Citations
More filters
TL;DR: Genomic tools are now allowing genome-wide studies, and recent theoretical advances can help to design research strategies that combine genomics and field experiments to examine the genetics of local adaptation.
Abstract: It is increasingly important to improve our understanding of the genetic basis of local adaptation because of its relevance to climate change, crop and animal production, and conservation of genetic resources Phenotypic patterns that are generated by spatially varying selection have long been observed, and both genetic mapping and field experiments provided initial insights into the genetic architecture of adaptive traits Genomic tools are now allowing genome-wide studies, and recent theoretical advances can help to design research strategies that combine genomics and field experiments to examine the genetics of local adaptation These advances are also allowing research in non-model species, the adaptation patterns of which may differ from those of traditional model species
1,060 citations
TL;DR: Expected future directions in the field of landscape genomics are summarized, such as the extension of statistical approaches, environmental association analysis for ecological gene annotation, and the need for replication and post hoc validation studies.
Abstract: Landscape genomics is an emerging research field that aims to identify the environmental factors that shape adaptive genetic variation and the gene variants that drive local adaptation. Its development has been facilitated by next-generation sequencing, which allows for screening thousands to millions of single nucleotide polymorphisms in many individuals and populations at reasonable costs. In parallel, data sets describing environmental factors have greatly improved and increasingly become publicly accessible. Accordingly, numerous analytical methods for environmental association studies have been developed. Environmental association analysis identifies genetic variants associated with particular environmental factors and has the potential to uncover adaptive patterns that are not discovered by traditional tests for the detection of outlier loci based on population genetic differentiation. We review methods for conducting environmental association analysis including categorical tests, logistic regressions, matrix correlations, general linear models and mixed effects models. We discuss the advantages and disadvantages of different approaches, provide a list of dedicated software packages and their specific properties, and stress the importance of incorporating neutral genetic structure in the analysis. We also touch on additional important aspects such as sampling design, environmental data preparation, pooled and reduced-representation sequencing, candidate-gene approaches, linearity of allele-environment associations and the combination of environmental association analyses with traditional outlier detection tests. We conclude by summarizing expected future directions in the field, such as the extension of statistical approaches, environmental association analysis for ecological gene annotation, and the need for replication and post hoc validation studies.
533 citations
TL;DR: In this article, the authors developed a set of methods to robustly test for unusual allele frequency patterns and correlations between environmental variables and allele frequencies while accounting for these complications based on a Bayesian model previously implemented in the software Bayenv.
Abstract: Comparing allele frequencies among populations that differ in environment has long been a tool for detecting loci involved in local adaptation. However, such analyses are complicated by an imperfect knowledge of population allele frequencies and neutral correlations of allele frequencies among populations due to shared population history and gene flow. Here we develop a set of methods to robustly test for unusual allele frequency patterns and correlations between environmental variables and allele frequencies while accounting for these complications based on a Bayesian model previously implemented in the software Bayenv. Using this model, we calculate a set of “standardized allele frequencies” that allows investigators to apply tests of their choice to multiple populations while accounting for sampling and covariance due to population history. We illustrate this first by showing that these standardized frequencies can be used to detect nonparametric correlations with environmental variables; these correlations are also less prone to spurious results due to outlier populations. We then demonstrate how these standardized allele frequencies can be used to construct a test to detect SNPs that deviate strongly from neutral population structure. This test is conceptually related to FST and is shown to be more powerful, as we account for population history. We also extend the model to next-generation sequencing of population pools—a cost-efficient way to estimate population allele frequencies, but one that introduces an additional level of sampling noise. The utility of these methods is demonstrated in simulations and by reanalyzing human SNP data from the Human Genome Diversity Panel populations and pooled next-generation sequencing data from Atlantic herring. An implementation of our method is available from http://gcbias.org.
523 citations
TL;DR: It is concluded that in species that exhibit IBD or have undergone range expansion, many of the published FST outliers based on FDIST2 and BayeScan are probably false positives, but FLK and Bayenv2 show great promise for accurately identifying loci under spatially divergent selection.
Abstract: FST outlier tests are a potentially powerful way to detect genetic loci under spatially divergent selection. Unfortunately, the extent to which these tests are robust to nonequilibrium demographic histories has been understudied. We developed a landscape genetics simulator to test the effects of isolation by distance (IBD) and range expansion on FST outlier methods. We evaluated the two most commonly used methods for the identification of FST outliers (FDIST2 and BayeScan, which assume samples are evolutionarily independent) and two recent methods (FLK and Bayenv2, which estimate and account for evolutionary nonindependence). Parameterization with a set of neutral loci (‘neutral parameterization’) always improved the performance of FLK and Bayenv2, while neutral parameterization caused FDIST2 to actually perform worse in the cases of IBD or range expansion. BayeScan was improved when the prior odds on neutrality was increased, regardless of the true odds in the data. On their best performance, however, the widely used methods had high false-positive rates for IBD and range expansion and were outperformed by methods that accounted for evolutionary nonindependence. In addition, default settings in FDIST2 and BayeScan resulted in many false positives suggesting balancing selection. However, all methods did very well if a large set of neutral loci is available to create empirical P-values. We conclude that in species that exhibit IBD or have undergone range expansion, many of the published FST outliers based on FDIST2 and BayeScan are probably false positives, but FLK and Bayenv2 show great promise for accurately identifying loci under spatially divergent selection.
493 citations
TL;DR: This work compares FST outlier and genetic–environment association (GEA) methods for each of two approaches that control for population structure: with a covariance matrix or with latent factors and shows that while the relative power of two methods in the same category depended largely on the number of individuals sampled, overall GEA tests had higher power in the island model and FST had higherPower under isolation by distance.
Abstract: Although genome scans have become a popular approach towards understanding the genetic basis of local adaptation, the field still does not have a firm grasp on how sampling design and demographic history affect the performance of genome scans on complex landscapes. To explore these issues, we compared 20 different sampling designs in equilibrium (i.e. island model and isolation by distance) and nonequilibrium (i.e. range expansion from one or two refugia) demographic histories in spatially heterogeneous environments. We simulated spatially complex landscapes, which allowed us to exploit local maxima and minima in the environment in 'pair' and 'transect' sampling strategies. We compared F(ST) outlier and genetic-environment association (GEA) methods for each of two approaches that control for population structure: with a covariance matrix or with latent factors. We show that while the relative power of two methods in the same category (F(ST) or GEA) depended largely on the number of individuals sampled, overall GEA tests had higher power in the island model and F(ST) had higher power under isolation by distance. In the refugia models, however, these methods varied in their power to detect local adaptation at weakly selected loci. At weakly selected loci, paired sampling designs had equal or higher power than transect or random designs to detect local adaptation. Our results can inform sampling designs for studies of local adaptation and have important implications for the interpretation of genome scans based on landscape data.
438 citations
References
More filters
TL;DR: In this paper, a different approach to problems of multiple significance testing is presented, which calls for controlling the expected proportion of falsely rejected hypotheses -the false discovery rate, which is equivalent to the FWER when all hypotheses are true but is smaller otherwise.
Abstract: SUMMARY The common approach to the multiplicity problem calls for controlling the familywise error rate (FWER). This approach, though, has faults, and we point out a few. A different approach to problems of multiple significance testing is presented. It calls for controlling the expected proportion of falsely rejected hypotheses -the false discovery rate. This error rate is equivalent to the FWER when all hypotheses are true but is smaller otherwise. Therefore, in problems where the control of the false discovery rate rather than that of the FWER is desired, there is potential for a gain in power. A simple sequential Bonferronitype procedure is proved to control the false discovery rate for independent test statistics, and a simulation study shows that the gain in power is substantial. The use of the new procedure and the appropriateness of the criterion are illustrated with examples.
83,420 citations
Book•
01 Jan 1983TL;DR: In this paper, a generalization of the analysis of variance is given for these models using log- likelihoods, illustrated by examples relating to four distributions; the Normal, Binomial (probit analysis, etc.), Poisson (contingency tables), and gamma (variance components).
Abstract: The technique of iterative weighted linear regression can be used to obtain maximum likelihood estimates of the parameters with observations distributed according to some exponential family and systematic effects that can be made linear by a suitable transformation. A generalization of the analysis of variance is given for these models using log- likelihoods. These generalized linear models are illustrated by examples relating to four distributions; the Normal, Binomial (probit analysis, etc.), Poisson (contingency tables) and gamma (variance components).
23,215 citations
TL;DR: The purpose of this discussion is to offer some unity to various estimation formulae and to point out that correlations of genes in structured populations, with which F-statistics are concerned, are expressed very conveniently with a set of parameters treated by Cockerham (1 969, 1973).
Abstract: This journal frequently contains papers that report values of F-statistics estimated from genetic data collected from several populations. These parameters, FST, FIT, and FIS, were introduced by Wright (1951), and offer a convenient means of summarizing population structure. While there is some disagreement about the interpretation of the quantities, there is considerably more disagreement on the method of evaluating them. Different authors make different assumptions about sample sizes or numbers of populations and handle the difficulties of multiple alleles and unequal sample sizes in different ways. Wright himself, for example, did not consider the effects of finite sample size. The purpose of this discussion is to offer some unity to various estimation formulae and to point out that correlations of genes in structured populations, with which F-statistics are concerned, are expressed very conveniently with a set of parameters treated by Cockerham (1 969, 1973). We start with the parameters and construct appropriate estimators for them, rather than beginning the discussion with various data functions. The extension of Cockerham's work to multiple alleles and loci will be made explicit, and the use of jackknife procedures for estimating variances will be advocated. All of this may be regarded as an extension of a recent treatment of estimating the coancestry coefficient to serve as a mea-
17,890 citations
TL;DR: UNLABELLED Analysis of Phylogenetics and Evolution (APE) is a package written in the R language for use in molecular evolution and phylogenetics that provides both utility functions for reading and writing data and manipulating phylogenetic trees.
Abstract: Summary: Analysis of Phylogenetics and Evolution (APE) is a package written in the R language for use in molecular evolution and phylogenetics. APE provides both utility functions for reading and writing data and manipulating phylogenetic trees, as well as several advanced methods for phylogenetic and evolutionary analysis (e.g. comparative and population genetic methods). APE takes advantage of the many R functions for statistics and graphics, and also provides a flexible framework for developing and implementing further statistical methods for the analysis of evolutionary processes.
Availability: The program is free and available from the official R package archive at http://cran.r-project.org/src/contrib/PACKAGES.html#ape. APE is licensed under the GNU General Public License.
10,818 citations
Book•
01 Jan 1986
TL;DR: It is argued that the common assumption that selection is usually weak in natural populations is no longer tenable, but that natural selection is only one component of the process of evolution; natural selection can explain the change of frequencies of variants, but not their origins.
Abstract: Natural selection is an immense and important subject, yet there have been few attempts to summarize its effects on natural populations, and fewer still which discuss the problems of working with natural selection in the wild. These are the purposes of John Endler's book. In it, he discusses the methods and problems involved in the demonstration and measurement of natural selection, presents the critical evidence for its existence, and places it in an evolutionary perspective. Professor Endler finds that there are a remarkable number of direct demonstrations of selection in a wide variety of animals and plants. The distribution of observed magnitudes of selection in natural populations is surprisingly broad, and it overlaps extensively the range of values found in artificial selection. He argues that the common assumption that selection is usually weak in natural populations is no longer tenable, but that natural selection is only one component of the process of evolution; natural selection can explain the change of frequencies of variants, but not their origins.
3,161 citations