Estimating bonobo (Panpaniscus) and chimpanzee (Pantroglodytes) evolutionary history from nucleotide site patterns.

Admixture appears increasingly ubiquitous in the evolutionary history of various taxa, including humans. Such gene flow likely also occurred among our closest living relatives: bonobos (Pan paniscus) and chimpanzees (Pan troglodytes). However, our understanding of their evolutionary history has been limited by studies that do not consider all Pan lineages or do not analyze all lineages simultaneously, resulting in conflicting demographic models. Here, we investigate this gap in knowledge using nucleotide site patterns calculated from whole-genome sequences from the autosomes of 71 bonobos and chimpanzees, representing all five extant Pan lineages. We estimated demographic parameters and compared all previously proposed demographic models for this clade. We further considered sex bias in Pan evolutionary history by analyzing the site patterns from the X chromosome. We show that 1) 21% of autosomal DNA in eastern chimpanzees derives from western chimpanzee introgression and that 2) all four chimpanzee lineages share a common ancestor about 987,000 y ago, much earlier than previous estimates. In addition, we suggest that 3) there was male reproductive skew throughout Pan evolutionary history and find evidence of 4) male-biased dispersal from western to eastern chimpanzees. Collectively, these results offer insight into bonobo and chimpanzee evolutionary history and suggest considerable differences between current and historic chimpanzee biogeography.

Early history of Neanderthals and Denisovans.

Extensive DNA sequence data have made it possible to reconstruct human evolutionary history in unprecedented detail. We introduce a method to study the past several hundred thousand years. Our results show that (i) the Neanderthal-Denisovan lineage declined to a small size just after separating from the modern lineage, (ii) Neanderthals and Denisovans separated soon thereafter, and (iii) the subsequent Neanderthal population was large and deeply subdivided. They also (iv) support previous estimates of gene flow from Neanderthals into modern Eurasians. These results suggest an archaic human diaspora early in the Middle Pleistocene.

Evolutionary history of Tibetans inferred from whole-genome sequencing.

The indigenous people of the Tibetan Plateau have been the subject of much recent interest because of their unique genetic adaptations to high altitude. Recent studies have demonstrated that the Tibetan EPAS1 haplotype is involved in high altitude-adaptation and originated in an archaic Denisovan-related population. We sequenced the whole-genomes of 27 Tibetans and conducted analyses to infer a detailed history of demography and natural selection of this population. We detected evidence of population structure between the ancestral Han and Tibetan subpopulations as early as 44 to 58 thousand years ago, but with high rates of gene flow until approximately 9 thousand years ago. The CMS test ranked EPAS1 and EGLN1 as the top two positive selection candidates, and in addition identified PTGIS, VDR, and KCTD12 as new candidate genes. The advantageous Tibetan EPAS1 haplotype shared many variants with the Denisovan genome, with an ancient gene tree divergence between the Tibetan and Denisovan haplotypes of about 1 million years ago. With the exception of EPAS1, we observed no evidence of positive selection on Denisovan-like haplotypes.

Legofit: estimating population history from genetic data.

BACKGROUND: Our current understanding of archaic admixture in humans relies on statistical methods with large biases, whose magnitudes depend on the sizes and separation times of ancestral populations. To avoid these biases, it is necessary to estimate these parameters simultaneously with those describing admixture. Genetic estimates of population histories also confront problems of statistical identifiability: different models or different combinations of parameter values may fit the data equally well. To deal with this problem, we need methods of model selection and model averaging, which are lacking from most existing software. RESULTS: The Legofit software package allows simultaneous estimation of parameters describing admixture, and the sizes and separation times of ancestral populations. It includes facilities for data manipulation, estimation, analysis of residuals, model selection, and model averaging. CONCLUSIONS: Legofit uses genetic data to study the history of a subdivided population. It is unaffected by recent history and can therefore focus on the deep history of population size, subdivision, and admixture. It outperforms several statistical methods that have been widely used to study population history and should be useful in any species for which DNA sequence data is available from several populations.

Bias in estimators of archaic admixture.

This article evaluates bias in one class of methods used to estimate archaic admixture in modern humans. These methods study the pattern of allele sharing among modern and archaic genomes. They are sensitive to "ghost" admixture, which occurs when a population receives archaic DNA from sources not acknowledged by the statistical model. The effect of ghost admixture depends on two factors: branch-length bias and population-size bias. Branch-length bias occurs because a given amount of admixture has a larger effect if the two populations have been separated for a long time. Population-size bias occurs because differences in population size distort branch lengths in the gene genealogy. In the absence of ghost admixture, these effects are small. They become important, however, in the presence of ghost admixture. Estimators differ in the pattern of response. Increasing a given parameter may inflate one estimator but deflate another. For this reason, comparisons among estimators are informative. Using such comparisons, this article supports previous findings that the archaic population was small and that Europeans received little gene flow from archaic populations other than Neanderthals. It also identifies an inconsistency in estimates of archaic admixture into Melanesia.

Genetic relatedness to sisters' children has been underestimated.

Males of many species help in the care and provisioning of offspring, and these investments often correlate with genetic relatedness. For example, many human males invest in the children of sisters, and this is especially so where men are less likely to share genes with children of wives. Although this makes qualitative sense, it has been difficult to support quantitatively. The prevailing model predicts investment in children of sisters only when paternity confidence falls below 0.268. This value is often seen as too low to be credible; so investment in sisters' children represents an unsolved problem. I show here that the prevailing model rests on a series of restrictive assumptions that underestimate relatedness to sisters' children. For this reason, it understates the fitness payoff to men who invest in these children. This effect can be substantial, especially in societies with low confidence in paternity. But this effect cannot be estimated solely from confidence in paternity. One must also estimate the probability that two siblings share the same father.

Mobile elements reveal small population size in the ancient ancestors of Homo sapiens.

The genealogies of different genetic loci vary in depth. The deeper the genealogy, the greater the chance that it will include a rare event, such as the insertion of a mobile element. Therefore, the genealogy of a region that contains a mobile element is on average older than that of the rest of the genome. In a simple demographic model, the expected time to most recent common ancestor (TMRCA) is doubled if a rare insertion is present. We test this expectation by examining single nucleotide polymorphisms around polymorphic Alu insertions from two completely sequenced human genomes. The estimated TMRCA for regions containing a polymorphic insertion is two times larger than the genomic average (P < <10(-30)), as predicted. Because genealogies that contain polymorphic mobile elements are old, they are shaped largely by the forces of ancient population history and are insensitive to recent demographic events, such as bottlenecks and expansions. Remarkably, the information in just two human DNA sequences provides substantial information about ancient human population size. By comparing the likelihood of various demographic models, we estimate that the effective population size of human ancestors living before 1.2 million years ago was 18,500, and we can reject all models where the ancient effective population size was larger than 26,000. This result implies an unusually small population for a species spread across the entire Old World, particularly in light of the effective population sizes of chimpanzees (21,000) and gorillas (25,000), which each inhabit only one part of a single continent.

Detecting positive selection from genome scans of linkage disequilibrium.

BACKGROUND: Though a variety of linkage disequilibrium tests have recently been introduced to measure the signal of recent positive selection, the statistical properties of the various methods have not been directly compared. While most applications of these tests have suggested that positive selection has played an important role in recent human history, the results of these tests have varied dramatically. RESULTS: Here, we evaluate the performance of three statistics designed to detect incomplete selective sweeps, LRH and iHS, and ALnLH. To analyze the properties of these tests, we introduce a new computational method that can model complex population histories with migration and changing population sizes to simulate gene trees influenced by recent positive selection. We demonstrate that iHS performs substantially better than the other two statistics, with power of up to 0.74 at the 0.01 level for the variation best suited for full genome scans and a power of over 0.8 at the 0.01 level for the variation best suited for candidate gene tests. The performance of the iHS statistic was robust to complex demographic histories and variable recombination rates. Genome scans involving the other two statistics suffer from low power and high false positive rates, with false discovery rates of up to 0.96 for ALnLH. The difference in performance between iHS and ALnLH, did not result from the properties of the statistics, but instead from the different methods for mitigating the multiple comparison problem inherent in full genome scans. CONCLUSIONS: We introduce a new method for simulating genealogies influenced by positive selection with complex demographic scenarios. In a power analysis based on this method, iHS outperformed LRH and ALnLH in detecting incomplete selective sweeps. We also show that the single-site iHS statistic is more powerful in a candidate gene test than the multi-site statistic, but that the multi-site statistic maintains a low false discovery rate with only a minor loss of power when applied to a scan of the entire genome. Our results highlight the need for careful consideration of multiple comparison problems when evaluating and interpreting the results of full genome scans for positive selection.

Linkage disequilibrium between loci with unknown phase.

Linkage disequilibrium is often measured by two statistics, D and r, which can be interpreted as the covariance and the correlation between loci and across gametes. When data consist of diploid genotypes, however, gametes cannot be identified. A variety of iterative statistical methods are used in such cases, all of which assume random mating. Previous work has shown that D and r can be expressed as covariances and correlations across diploid genotypes, provided that mating is random. We show here that this result also holds approximately when mating is nonrandom. This provides a means of estimating these parameters without iteration and without assuming random mating. This estimator is nearly as accurate as the widely used EM estimator and is many times faster.

Estimating the age of retrotransposon subfamilies using maximum likelihood.

We present a maximum likelihood model to estimate the age of retrotransposon subfamilies. This method is designed around a master gene model which assumes a constant retrotransposition rate. The statistical properties of this model and an ad hoc estimation procedure are compared using two simulated data sets. We also test whether each estimation procedure is robust to violation of the master gene model. According to our results, both estimation procedures are accurate under the master gene model. While both methods tend to overestimate ages under the intermediate model, the maximum likelihood estimate is significantly less inflated than the ad hoc estimate. We estimate the ages of two subfamilies of human-specific LINE-I insertions using both estimation procedures. By calculating confidence intervals around the maximum likelihood estimate, our model can both provide an estimate of retrotransposon subfamily age and describe the range of subfamily ages consistent with the data.

Neanderthal-Denisovan ancestors interbred with a distantly related hominin.

Previous research has shown that modern Eurasians interbred with their Neanderthal and Denisovan predecessors. We show here that hundreds of thousands of years earlier, the ancestors of Neanderthals and Denisovans interbred with their own Eurasian predecessors-members of a "superarchaic" population that separated from other humans about 2 million years ago. The superarchaic population was large, with an effective size between 20 and 50 thousand individuals. We confirm previous findings that (i) Denisovans also interbred with superarchaics, (ii) Neanderthals and Denisovans separated early in the middle Pleistocene, (iii) their ancestors endured a bottleneck of population size, and (iv) the Neanderthal population was large at first but then declined in size. We provide qualified support for the view that (v) Neanderthals interbred with the ancestors of modern humans.

Genetic analysis of lice supports direct contact between modern and archaic humans.

Parasites can be used as unique markers to investigate host evolutionary history, independent of host data. Here we show that modern human head lice, Pediculus humanus, are composed of two ancient lineages, whose origin predates modern Homo sapiens by an order of magnitude (ca. 1.18 million years). One of the two louse lineages has a worldwide distribution and appears to have undergone a population bottleneck ca. 100,000 years ago along with its modern H. sapiens host. Phylogenetic and population genetic data suggest that the other lineage, found only in the New World, has remained isolated from the worldwide lineage for the last 1.18 million years. The ancient divergence between these two lice is contemporaneous with splits among early species of Homo, and cospeciation analyses suggest that the two louse lineages codiverged with a now extinct species of Homo and the lineage leading to modern H. sapiens. If these lice indeed codiverged with their hosts ca. 1.18 million years ago, then a recent host switch from an archaic species of Homo to modern H. sapiens is required to explain the occurrence of both lineages on modern H. sapiens. Such a host switch would require direct physical contact between modern and archaic forms of Homo.

Modeling the amplification dynamics of human Alu retrotransposons.

Retrotransposons have had a considerable impact on the overall architecture of the human genome. Currently, there are three lineages of retrotransposons (Alu, L1, and SVA) that are believed to be actively replicating in humans. While estimates of their copy number, sequence diversity, and levels of insertion polymorphism can readily be obtained from existing genomic sequence data and population sampling, a detailed understanding of the temporal pattern of retrotransposon amplification remains elusive. Here we pose the question of whether, using genomic sequence and population frequency data from extant taxa, one can adequately reconstruct historical amplification patterns. To this end, we developed a computer simulation that incorporates several known aspects of primate Alu retrotransposon biology and accommodates sampling effects resulting from the methods by which mobile elements are typically discovered and characterized. By modeling a number of amplification scenarios and comparing simulation-generated expectations to empirical data gathered from existing Alu subfamilies, we were able to statistically reject a number of amplification scenarios for individual subfamilies, including that of a rapid expansion or explosion of Alu amplification at the time of human-chimpanzee divergence.

How population growth affects linkage disequilibrium.

The "LD curve" relates the linkage disequilibrium (LD) between pairs of nucleotide sites to the distance that separates them along the chromosome. The shape of this curve reflects natural selection, admixture between populations, and the history of population size. This article derives new results about the last of these effects. When a population expands in size, the LD curve grows steeper, and this effect is especially pronounced following a bottleneck in population size. When a population shrinks, the LD curve rises but remains relatively flat. As LD converges toward a new equilibrium, its time path may not be monotonic. Following an episode of growth, for example, it declines to a low value before rising toward the new equilibrium. These changes happen at different rates for different LD statistics. They are especially slow for estimates of [Formula: see text], which therefore allow inferences about ancient population history. For the human population of Europe, these results suggest a history of population growth.

Ancestral alleles and population origins: inferences depend on mutation rate.

Previous studies have found that at most human loci, ancestral alleles are "African," in the sense that they reach their highest frequency there. Conventional wisdom holds that this reflects a recent African origin of modern humans. This paper challenges that view by showing that the empirical pattern (of elevated allele frequencies within Africa) is not as pervasive as has been thought. We confirm this African bias in a set of mainly protein-coding loci, but find a smaller bias in Alu insertion polymorphisms, and an even smaller bias in noncoding loci. Thus, the strong bias that was originally observed must reflect some factor that varies among data sets--something other than population history. This factor may be the per-locus mutation rate: the African bias is most pronounced in loci where this rate is high. The distribution of ancestral alleles among populations has been studied using 2 methods. One of these involves comparing the fractions of loci that reach maximal frequency in each population. The other compares the average frequencies of ancestral alleles. The first of these methods reflects history in a manner that depends on the mutation rate. When that rate is high, ancestral alleles at most loci reach their highest frequency in the ancestral population. When that rate is low, the reverse is true. The other method--comparing averages--is unresponsive. Average ancestral allele frequencies are affected neither by mutation rate nor by the history of population size and migration. In the absence of selection and ascertainment bias, they should be the same everywhere. This is true of one data set, but not of 2 others. This also suggests the action of some factor, such as selection or ascertainment bias, that varies among data sets.

Genomics refutes an exclusively African origin of humans.

Ten years ago, evidence from genetics gave strong support to the "recent African origin" view of the evolution of modern humans, which posits that Homo sapiens arose as a new species in Africa and subsequently spread, leading to the extinction of other archaic human species. Subsequent data from the nuclear genome not only fail to support this model, they do not support any simple model of human demographic history. In this paper, we study a process in which the modern human phenotype originates in Africa and then advances across the world by local demic diffusion, hybridization, and natural selection. While the multiregional model of human origins posits a number of independent single locus selective sweeps, and the "out of Africa" model posits a sweep of a new species, we study the intermediate case of a phenotypic sweep. Numerical simulations of this process replicate many of the seemingly contradictory features of the genetic data, and suggest that as much as 80% of nuclear loci have assimilated genetic material from non-African archaic humans.

GENETIC EVIDENCE FOR A PLEISTOCENE POPULATION EXPLOSION.

Expansions of population size leave characteristic signatures in mitochondrial "mismatch distributions." Consequently, these distributions can inform us about the history of changes in population size. Here, I study a simple model of population history that assumes that, t generations before the present, a population grows (or shrinks) suddenly from female size N0 to female size N1 . Although this model is simple, it often provides an accurate description of data generated by complex population histories. I develop statistical methods that estimate θ0 = 2uN0 , θ1 = 2uN1 , and τ = 2ut (where u is the mutation rate), and place a confidence region around these estimates. These estimators are well behaved, and insensitive to simplifying assumptions. Finally, I apply these methods to published mitochondrial data, and infer that a major expansion of the human population occurred during the late Pleistocene.

A MODEL OF KIN-STRUCTURED MIGRATION.

When individuals disperse from one local group to another, they often do so in the company of relatives. This is known as "kin-structured migration," and its effect on genetic population structure is investigated here. It is shown that when migration is kin-structured, the ratio of between- to within-group variance is increased by a quantity that can be estimated either from behavioral or genetic data. Theoretical results indicate that kin-structured migration should be most important in populations with high mobility, and analysis of data for humans and lions suggests the kin-structured migration may have a substantial effect on genetic population structure in both species. Its effect seems to be small in a population of pine voles.

MIGRATION AND GENETIC DRIFT IN HUMAN POPULATIONS.

In humans and many other species, mortality is concentrated early in the life cycle, and is low during the ages of dispersal and reproduction. Yet precisely the opposite is assumed by classical population-genetics models of migration and genetic drift. We introduce a model in which population regulation occurs before migration. In contrast to the conventional model, our model implies that geographic variation in the allele frequencies of newborns should exceed that of adults. Thus, it is important to distinguish genetic variation of adults from that of newborns in species with human-like life cycles. Classical models deal with the variance of group allele frequencies about the allele frequency of a hypothetical "continent" or "foundation stock." Empirical studies, however, can only measure "reduced" variance, i.e., variance about the current population mean. Our model deals with reduced variance, and should therefore be more relevant to field studies. We show that reduced variance converges faster, which implies that populations are more likely to be at equilibrium with respect to reduced than unreduced variance. To summarize the effect of migration on genetic population structure, we introduce a new parameter, the effective migration rate. Unlike most population structure statistics, it does not confound the effects of mobility and population size, and it should therefore be useful for comparisons between populations. Finally, we show that the difference between geographic variation of newborn and adult allele frequencies contains information about both effective population size and effective migration rate.