Considering Genomic Scans for Selection as Coalescent Model Choice

Inferring demographic and selective histories from population genomic data using a 2-step approach in species with coding-sparse genomes: an application to human data

The demographic history of a population, and the distribution of fitness effects (DFE) of newly arising mutations in functional genomic regions, are fundamental factors dictating both genetic variation and evolutionary trajectories. Although both demographic and DFE inference has been performed extensively in humans, these approaches have generally either been limited to simple demographic models involving a single population, or, where a complex population history has been inferred, without accounting for the potentially confounding effects of selection at linked sites. Taking advantage of the coding-sparse nature of the genome, we propose a 2-step approach in which coalescent simulations are first used to infer a complex multi-population demographic model, utilizing large non-functional regions that are likely free from the effects of background selection. We then use forward-in-time simulations to perform DFE inference in functional regions, conditional on the complex demography inferred and utilizing expected background selection effects in the estimation procedure. Throughout, recombination and mutation rate maps were used to account for the underlying empirical rate heterogeneity across the human genome. Importantly, within this framework it is possible to utilize and fit multiple aspects of the data, and this inference scheme represents a generalized approach for such large-scale inference in species with coding-sparse genomes.

Inferring the Demographic History of Aye-Ayes (Daubentonia madagascariensis) from High-Quality, Whole-Genome, Population-Level Data

The nocturnal aye-aye, Daubentonia madagascariensis, is one of the most elusive lemurs on the island of Madagascar. The timing of its activity and arboreal lifestyle has generally made it difficult to obtain accurate assessments of population size using traditional census methods. Therefore, alternative estimates provided by population genetic inference are essential for yielding much needed information for conservation measures and for enabling ecological and evolutionary studies of this species. Here, we utilize genomic data from 17 individuals-including 5 newly sequenced, high-coverage genomes-to estimate this history. Essential to this estimation are recently published annotations of the aye-aye genome which allow for variation at putatively neutral genomic regions to be included in the estimation procedures, and regions subject to selective constraints, or in linkage to such sites, to be excluded owing to the biasing effects of selection on demographic inference. By comparing a variety of demographic estimation tools to develop a well-supported model of population history, we find strong support for two demes, separating northern Madagascar from the rest of the island. Additionally, we find that the aye-aye has experienced two severe reductions in population size. The first occurred rapidly, ∼3,000 to 5,000 years ago, and likely corresponded with the arrival of humans to Madagascar. The second occurred over the past few decades and is likely related to substantial habitat loss, suggesting that the species is still undergoing population decline and remains at great risk for extinction.

Inferring demographic and selective histories from population genomic data using a two-step approach in species with coding-sparse genomes: an application to human data

The demographic history of a population, and the distribution of fitness effects (DFE) of newly arising mutations in functional genomic regions, are fundamental factors dictating both genetic variation and evolutionary trajectories. Although both demographic and DFE inference has been performed extensively in humans, these approaches have generally either been limited to simple demographic models involving a single population, or, where a complex population history has been inferred, without accounting for the potentially confounding effects of selection at linked sites. Taking advantage of the coding-sparse nature of the genome, we propose a 2-step approach in which coalescent simulations are first used to infer a complex multi-population demographic model, utilizing large non-functional regions that are likely free from the effects of background selection. We then use forward-in-time simulations to perform DFE inference in functional regions, conditional on the complex demography inferred and utilizing expected background selection effects in the estimation procedure. Throughout, recombination and mutation rate maps were used to account for the underlying empirical rate heterogeneity across the human genome. Importantly, within this framework it is possible to utilize and fit multiple aspects of the data, and this inference scheme represents a generalized approach for such large-scale inference in species with coding-sparse genomes.

Inferring the demographic history of aye-ayes (Daubentonia madagascariensis) from high-quality, whole-genome, population-level data

The nocturnal aye-aye, Daubentonia madagascariensis, is one of the most elusive lemurs on the island of Madagascar. The timing of its activity and arboreal lifestyle has generally made it difficult to obtain accurate assessments of population size using traditional census methods. Therefore, alternative estimates provided by population genetic inference are essential for yielding much needed information for conservation measures and for enabling ecological and evolutionary studies of this species. Here, we utilize genomic data from 17 unrelated individuals - including 5 newly sequenced, high-coverage genomes - to estimate this history. Essential to this estimation are recently published annotations of the aye-aye genome which allow for variation at putatively neutral genomic regions to be included in the estimation procedures, and regions subject to selective constraints, or in linkage to such sites, to be excluded owing to the biasing effects of selection on demographic inference. By comparing a variety of demographic estimation tools to develop a well-supported model of population history, we find strong support for the species to consist of two demes, separating northern Madagascar from the rest of the island. Additionally, we find that the aye-aye has experienced two severe reductions in population size. The first occurred rapidly, approximately 3,000 to 5,000 years ago, and likely corresponded with the arrival of humans to Madagascar. The second occurred over the past few decades and is likely related to substantial habitat loss, suggesting that the species is still undergoing population decline and remains at great risk for extinction.

A whole-genome scan for evidence of recent positive and balancing selection in aye-ayes (Daubentonia madagascariensis) utilizing a well-fit evolutionary baseline model

The aye-aye (Daubentonia madagascariensis) is one of the 25 most endangered primate species in the world, maintaining amongst the lowest genetic diversity of any primate measured to date. Characterizing patterns of genetic variation within aye-aye populations, and the relative influences of neutral and selective processes in shaping that variation, is thus important for future conservation efforts. In this study, we performed the first whole-genome scans for recent positive and balancing selection in the species, utilizing high-coverage population genomic data from newly sequenced individuals. We generated null thresholds for our genomic scans by creating an evolutionarily appropriate baseline model that incorporates the demographic history of this aye-aye population, and identified a small number of candidate genes. Most notably, a suite of genes involved in olfaction - a key trait in these nocturnal primates - were identified as experiencing long-term balancing selection. We also conducted analyses to quantify the expected statistical power to detect positive and balancing selection in this population using site frequency spectrum-based inference methods, once accounting for the potentially confounding contributions of population history, recombination and mutation rate variation, and purifying and background selection. This work, presenting the first high-quality, genome-wide polymorphism data across the functional regions of the aye-aye genome, thus provides important insights into the landscape of episodic selective forces in this highly endangered species.

Temporal challenges in detecting balancing selection from population genomic data

The role of balancing selection in maintaining genetic variation remains an open question in population genetics. Recent years have seen numerous studies identifying candidate loci potentially experiencing balancing selection, most predominantly in human populations. There are however numerous alternative evolutionary processes that may leave similar patterns of variation, thereby potentially confounding inference, and the expected signatures of balancing selection additionally change in a temporal fashion. Here we use forward-in-time simulations to quantify expected statistical power to detect balancing selection using both site frequency spectrum- and linkage disequilibrium-based methods under a variety of evolutionarily realistic null models. We find that whilst site frequency spectrum-based methods have little power immediately after a balanced mutation begins segregating, power increases with time since the introduction of the balanced allele. Conversely, linkage disequilibrium-based methods have considerable power whilst the allele is young, and power dissipates rapidly as the time since introduction increases. Taken together, this suggests that site frequency spectrum-based methods are most effective at detecting long-term balancing selection (>25N generations since the introduction of the balanced allele) whilst linkage disequilibrium-based methods are effective over much shorter timescales (<1N generations), thereby leaving a large time frame over which current methods have little power to detect the action of balancing selection. Finally, we investigate the extent to which alternative evolutionary processes may mimic these patterns, and demonstrate the need for caution in attempting to distinguish the signatures of balancing selection from those of both neutral processes (e.g. population structure and admixture) as well as of alternative selective processes (e.g. partial selective sweeps).

The Effects of Mutation and Recombination Rate Heterogeneity on the Inference of Demography and the Distribution of Fitness Effects

Disentangling the effects of demography and selection has remained a focal point of population genetic analysis. Knowledge about mutation and recombination is essential in this endeavor; however, despite clear evidence that both mutation and recombination rates vary across genomes, it is common practice to model both rates as fixed. In this study, we quantify how this unaccounted for rate heterogeneity may impact inference using common approaches for inferring selection (DFE-alpha, Grapes, and polyDFE) and/or demography (fastsimcoal2 and δaδi). We demonstrate that, if not properly modeled, this heterogeneity can increase uncertainty in the estimation of demographic and selective parameters and in some scenarios may result in mis-leading inference. These results highlight the importance of quantifying the fundamental evolutionary parameters of mutation and recombination before utilizing population genomic data to quantify the effects of genetic drift (i.e. as modulated by demographic history) and selection; or, at the least, that the effects of uncertainty in these parameters can and should be directly modeled in downstream inference.

Evaluating power to detect recurrent selective sweeps under increasingly realistic evolutionary null models

The detection of selective sweeps from population genomic data often relies on the premise that the beneficial mutations in question have fixed very near the sampling time. As it has been previously shown that the power to detect a selective sweep is strongly dependent on the time since fixation as well as the strength of selection, it is naturally the case that strong, recent sweeps leave the strongest signatures. However, the biological reality is that beneficial mutations enter populations at a rate, one that partially determines the mean wait time between sweep events and hence their age distribution. An important question thus remains about the power to detect recurrent selective sweeps when they are modeled by a realistic mutation rate and as part of a realistic distribution of fitness effects, as opposed to a single, recent, isolated event on a purely neutral background as is more commonly modeled. Here we use forward-in-time simulations to study the performance of commonly used sweep statistics, within the context of more realistic evolutionary baseline models incorporating purifying and background selection, population size change, and mutation and recombination rate heterogeneity. Results demonstrate the important interplay of these processes, necessitating caution when interpreting selection scans; specifically, false-positive rates are in excess of true-positive across much of the evaluated parameter space, and selective sweeps are often undetectable unless the strength of selection is exceptionally strong.

Evaluating power to detect recurrent selective sweeps under increasingly realistic evolutionary null models

The detection of selective sweeps from population genomic data often relies on the premise that the beneficial mutations in question have fixed very near the sampling time. As it has been previously shown that the power to detect a selective sweep is strongly dependent on the time since fixation as well as the strength of selection, it is naturally the case that strong, recent sweeps leave the strongest signatures. However, the biological reality is that beneficial mutations enter populations at a rate, one that partially determines the mean wait time between sweep events and hence their age distribution. An important question thus remains about the power to detect recurrent selective sweeps when they are modelled by a realistic mutation rate and as part of a realistic distribution of fitness effects (DFE), as opposed to a single, recent, isolated event on a purely neutral background as is more commonly modelled. Here we use forward-in-time simulations to study the performance of commonly used sweep statistics, within the context of more realistic evolutionary baseline models incorporating purifying and background selection, population size change, and mutation and recombination rate heterogeneity. Results demonstrate the important interplay of these processes, necessitating caution when interpreting selection scans; specifically, false positive rates are in excess of true positive across much of the evaluated parameter space, and selective sweeps are often undetectable unless the strength of selection is exceptionally strong.

Teaser Text

Outlier-based genomic scans have proven a popular approach for identifying loci that have potentially experienced recent positive selection. However, it has previously been shown that an evolutionarily appropriate baseline model that incorporates non-equilibrium population histories, purifying and background selection, and variation in mutation and recombination rates is necessary to reduce often extreme false positive rates when performing genomic scans. Here we evaluate the power to detect recurrent selective sweeps using common SFS-based and haplotype-based methods under these increasingly realistic models. We find that while these appropriate evolutionary baselines are essential to reduce false positive rates, the power to accurately detect recurrent selective sweeps is generally low across much of the biologically relevant parameter space.

Interpreting the pervasive observation of U-shaped Site Frequency Spectra

The standard neutral model of molecular evolution has traditionally been used as the null model for population genomics. We gathered a collection of 45 genome-wide site frequency spectra from a diverse set of species, most of which display an excess of low and high frequency variants compared to the expectation of the standard neutral model, resulting in U-shaped spectra. We show that multiple merger coalescent models often provide a better fit to these observations than the standard Kingman coalescent. Hence, in many circumstances these under-utilized models may serve as the more appropriate reference for genomic analyses. We further discuss the underlying evolutionary processes that may result in the widespread U-shape of frequency spectra.

How Can We Resolve Lewontin's Paradox?

We discuss the genetic, demographic, and selective forces that are likely to be at play in restricting observed levels of DNA sequence variation in natural populations to a much smaller range of values than would be expected from the distribution of census population sizes alone-Lewontin's Paradox. While several processes that have previously been strongly emphasized must be involved, including the effects of direct selection and genetic hitchhiking, it seems unlikely that they are sufficient to explain this observation without contributions from other factors. We highlight a potentially important role for the less-appreciated contribution of population size change; specifically, the likelihood that many species and populations may be quite far from reaching the relatively high equilibrium diversity values that would be expected given their current census sizes.

Effects of Selection at Linked Sites on Patterns of Genetic Variability

Patterns of variation and evolution at a given site in a genome can be strongly influenced by the effects of selection at genetically linked sites. In particular, the recombination rates of genomic regions correlate with their amount of within-population genetic variability, the degree to which the frequency distributions of DNA sequence variants differ from their neutral expectations, and the levels of adaptation of their functional components. We review the major population genetic processes that are thought to lead to these patterns, focusing on their effects on patterns of variability: selective sweeps, background selection, associative overdominance, and Hill–Robertson interference among deleterious mutations. We emphasize the difficulties in distinguishing among the footprints of these processes and disentangling them from the effects of purely demographic factors such as population size changes. We also discuss how interactions between selective and demographic processes can significantly affect patterns of variability within genomes.