Population Genetics of Modern Human Evolution

Transcription

1 Population Genetics of Modern Human Evolution John H Relethford, State University of New York College at Oneonta, Oneonta, New York, USA Rosalind M Harding, University of Oxford, Oxford, UK The aim of studies in human population genetics is to determine how mutation, genetic drift, gene flow and natural selection have generated patterns of genetic diversity within and between populations. One application of these studies is to questions about how modern humans evolved and the meaning of human racial variation. Introductory article. Introduction. Race. Genetic Variation. Human Evolution. Mitochondrial Eve. Y-chromosome Adam. Neanderthal Enigma. Summary Article Contents Introduction Population genetics tackles questions about genetic diversity. Approximately 0.08% of the nucleotide base pairs (bp) in human DNA vary among individuals. Why these and not others? One explanation is that selection favours functionally different DNA alleles in different circumstances. Another is that DNA variation is tolerated when the alleles of a gene are functionally equivalent. The former explanation clearly applies to some variation, but the latter explanation, formalized as neutral theory, is invoked most often. Either way, the aim of population genetics is to model the dynamics of evolutionary change within and between populations. There are four basic evolutionary forces: mutation, natural selection, genetic drift and gene flow. Mutations are copying errors during DNA replication and transcription, which introduce new alleles into the population. Natural selection is the differential transmission of alleles into the next generation due to the consequences of functional differences on an individual s survival and reproductive success. Genetic drift is the differential transmission of alleles into the next generation as a result of random sampling, and has the greatest potential impact in small populations. Gene flow spreads alleles from one population into another via migration, making them more genetically similar to each other, and countering genetic differentiation by drift. These four evolutionary forces are reflected in patterns of diversity, measured by the numbers of different alleles at a gene locus, the frequencies of each allele, and the interrelatedness of each allele to the others present at the same time. To understand why genetic diversity accumulates to a particular level, why it has its observed distribution, and how its turnover occurs, we investigate the interactions of these four evolutionary forces using population genetic models. Population genetic models are mathematical objects that allow us to interpret genetic diversity when the assumptions on which these models rest are acceptable. To capture the idea of a basic model, the following analogy may be helpful. Think of a population as a leaky bucket where the level of water in the bucket reflects the balance between the drip from the tap and the size of a leak. We might ask how big the bucket has to be to hold a given measure of water and how long the water has been standing. The equivalent population genetic questions concern the evolutionary size of a population expected for a given measure of genetic diversity and the age of this diversity. Mutation is analogous to the inflow from dripping water. Genetic drift is analogous to the water leaking away. Natural selection is a control mechanism on the leak that can either act to slow down or speed up the loss. When populations are connected to each other, an additional force to consider is gene flow, which would be analogous to water flowing between several buckets. When one of the four evolutionary forces is stronger than the others, it is easy to predict the outcome, but when they are in a balance the consequences for patterns of genetic variation are much more difficult to evaluate. The study of human population genetics (also known as anthropological genetics) is concerned with explaining the causes of human diversity in the world today and the evolutionary history that has generated this diversity. Studies of anthropological genetics include efforts to describe population structure, to reconstruct population history, and to understand adaptation to local environments. Many studies of anthropological genetics have focused on local populations, but more attention has recently been directed toward understanding the evolutionary history of the human species. Anthropological genetic investigation of global human diversity leads to several related questions: 1. What is the magnitude and pattern of genetic variation within and among populations in our species today? 2. What are the relative roles of population history, adaptation, and cultural behaviours on structuring diversity? 3. How well do patterns of genetic variation fit a traditional model of racial divergence? ENCYCLOPEDIA OF LIFE SCIENCES 2001, John Wiley & Sons, Ltd. 1

2 4. What can we say about the recent evolutionary history of our species based on analyses of genetic variation in living humans? This final question is of particular interest to anthropologists grappling with the question of modern human origins did modern humans arise as a new species in Africa within the past years, or did our modern form evolve within a population that was subdivided between the Old World continents of Africa and Eurasia but united as a single species by gene flow? To answer this and related questions, anthropological geneticists rely on making inferences about the past based on the present. In this sense, our genes reflect our history. Race Humans commonly and typically classify themselves and their neighbours into a number of races. However, the everyday use of the word race implies a legitimacy that is actually at odds with much of the genetic evidence. The underlying concept of race is that there are discrete and easily identifiable genetic subgroups within a species. Racial groups are expected as a consequence of an ongoing process of evolutionary divergence eventually expected to create new species. At an intermediate stage there is a prediction that the majority of variation occurs between races, and much less within races. There are problems with applying this concept of race to modern humans. The first problem is with the expectation of discrete subgroups and the second problem is with the apportionment of genetic variance. Historically, there has been little agreement about the number or definitions of human races, with many questions on where to place human populations that otherwise do not appear to fit simplistic models of racial classification. Even skin colour, one of the most highly variable traits in the human species, eludes easy classification, though virtually every racial classification scheme relies heavily on it. Skin colour does not actually appear in discrete shades such as white or black, but instead shows a continuous distribution across the species. Any decision to delineate between light medium and dark is arbitrary. Nor does similarity in skin colour necessarily reflect common genetic ancestry; the worldwide distribution of skin colour shows a very strong latitudinal correlation, with the darkest indigenous people living at or near the equator, and people being increasingly lighter with increasing north or south latitude. Ultraviolet radiation (UVR) also shows a correlation with latitude (greatest near the equator), suggesting that skin colour has been selected in response to differing UVR intensity. Thus, different populations often have similar skin colour because of adaptation to the same environmental factor and not because of close recent ancestry. Attempts to develop schemes of racial classification based on additional traits, both genetic and anatomic, have met with little success. Different traits often lead to completely different classifications. Thus, a racial classification based on lactose intolerance will be different from one based on skin colour, which will be different from one based on cranial shape or ABO blood group frequencies. An alternative approach is to define races by their geographic location, and to use these geographic units of analysis for measuring genetic variation attributed to race. We do expect neighbouring populations to share more genes than distant populations because geographic distance tends to limit gene flow. One difference between an isolation by distance model for genetic diversity compared with a racial model is that it predicts gradients rather than distinct boundaries between populations. Gradients rather than boundaries are typically observed. A further problem with geographic classification is the arbitrariness of decisions regarding the number of races, and where to draw the line separating one from the next. Should the entire continent of Africa be considered a single race? If so, then we have the problem of dealing with some of the highest levels of population diversity observed in humans subsumed under a single label. Where does one draw the line separating different regions of Eurasia? Should the peoples of the subcontinent of India be classified with Africans based on skin colour, Europeans based on facial characteristics, or in a separate race? Geographic analyses of race also illustrate the second problem regarding the expectation for greater variation between racial groups than within. No study of human genetic diversity has been able to define groups that show this pattern. Only about 10% of the total global variation in most genetic markers and craniofacial measures occurs between continental groups of populations. The remaining 90% of the variation occurs within these groups, of which most variation (85%) occurs within local populations. The partitioning of variation in the human species is exactly the opposite of that implied by the race concept. Anthropologists criticisms of the race concept have frequently been mistaken as a denial of human diversity. This is not the case there is a lot of diversity within and between human populations. The race concept is rejected as a means to describe this variation because the main apportionment of variation does not occur between discrete population clusters. While continental regions (e.g. subsaharan Africa, Europe) are often used as units of analysis, we realize that these are arbitrary units imposed on reality for the purpose of rough analysis. An analogy is human height; while height is continuous in nature, ranging from the shortest person to the tallest person, we often resort to the use of classificatory groups such as short, medium and tall, even though we realize that, in reality, there are not three distinct types of height. While the definition of populations and population groups are helpful for the description of patterns of diversity, the 2

3 evidence of genetic data denies the validity of the biological race concept for the human species. Genetic Variation Anthropological geneticists do not trace human genetic variation back to racial origins, but instead ask questions about how genetic diversity has evolved within a population genetic framework. Some of these questions are: 1. What does the level of genetic diversity imply about the evolutionary size of the human population and how old is this diversity? (This question is addressed using total global variation.) 2. Do levels of genetic variation within local populations vary, and if so, where is the greatest genetic diversity found? (The average level of total global variation within local populations is 85% and within regions is 90%.) 3. Do genetic distances between populations vary, and if so, what does this imply for the evolution of diversity among continental regions? (The average level of variation between populations is 5% within continents and 10% between continents.) Polymorphic variation when evaluated in globally representative samples occurs in about 0.08% of DNA base pairs in the nuclear genome, giving an estimate of about for the evolutionary population size of humans, assuming a simple population genetic model. (Remember the leaky bucket and dripping tap.) This estimate for the evolutionary population size seems small compared with the number of people living on the planet now, but it reflects only a rough estimate of numbers of potential ancestors living in past generations. Our ancestors have the remarkable attribute of an unbroken line of descent into the current generation, in at least one gene. They lived in populations with many other individuals whose lines of descent died out before the present generation. Many selective and random reasons contribute to this variability in long-term reproductive success and account for why estimates of evolutionary population size from genetic diversity, for all species, are small compared with census sizes. Population genetic models have also been used to estimate the time depth of typical diversity in the human nuclear genome, and it is suggested to be approximately years. This is an estimate for the nonfunctional (neutral) diversity at an average gene locus. The time depth of diversity found in any particular locus may be much more or much less than this expectation for the average, not only because of selection, but also because of the stochastic nature of genetic drift, which may generate huge variability in persistence times. For most loci, the consequences of adaptation and genetic drift are a high rate of turnover, so that typical genetic diversity preserves information only about our recent Pleistocene history. For some loci, diversity is very recent. The global diversity found on the Y-chromosome, for example, may have all been generated within the last years. Outside of the nuclear genome is a special locus, of which there are many copies in each cell, one in each mitochondrion. Global diversity in the human mitochondrial genome (mtdna) appears to trace back to a glacial maximum at years ago. A little more explanation of these very recent estimates is given below. Studies of how genetic diversity is distributed geographically show that common polymorphisms are often found globally and it is usually only polymorphisms with a minor allele frequency less than about 15% that are specific to populations. Population differences are mainly due to the presence of low-frequency alleles that have not diffused far from their geographic place of origin. Most lowfrequency alleles are young and are destined to be lost by genetic drift before ever becoming common. However, common polymorphisms also succumb to drift, and it may be the old ancestral allele that eventually is reduced to low frequency. One way to determine which allele at a locus is likely to be the ancestral variant is to make a comparison with a chimpanzee sequence. There will be many human chimpanzee differences at sites not polymorphic in humans, but typically, the polymorphic sites will include a base pair variant present in the chimp, allowing the ancestral sequence to be reconstructed. Interestingly, when these ancestral allelic sequences are low in frequency, they are more often found within subsaharan African populations than in Europe or Asia. A variety of different traits, including craniometric measures, blood groups and microsatellites (loci with variable numbers of repeated 2 5 bp motifs) as well as DNA sequence polymorphisms, have been studied to compare levels of genetic diversity between African populations with the populations of other geographic regions. When these studies focus on common polymorphisms they do not find significantly greater diversity in subsaharan Africa, but when they incorporate information from low-frequency alleles, these contribute more to diversity in populations of sub-saharan Africa than in Europe or Asia. The most likely explanation for higher African diversity is that there has been less genetic drift over evolutionary time in subsaharan Africa than elsewhere. Many traits also show a pattern where the largest interpopulation genetic distances are between subsaharan African populations and a variety of either European or Asian populations. This pattern reflects two aspects of human demographic history. First, accumulation of diversity within subsaharan Africa contributes a larger component to population differences than differentiation by drift among European and Asian populations. Second, patterns of both diversity accumulation and low levels of 3

4 population differentiation reflect high rates of gene flow within continental regions and between Europe and Asia. There has been comparatively less gene flow between Africa and Eurasia, probably because of isolation imposed by the Sahara Desert during the greater part of Pleistocene interglacials. Human Evolution Questions about human evolution are addressed through analysis of the fossil and archaeological records, combined with analyses of diversity in living human populations. The study of recent human evolution focuses on the transition between fossils referred to as archaic humans, with brain size similar to our own but with some morphological differences, and the fossils of anatomically modern humans. A variety of archaic fossils with dates between and years are found throughout the Old World. These fossils are related within a taxonomic group that first spread from subsaharan Africa into the Old World close to 2 million years ago. The oldest fossils showing modern human characteristics have been found in Africa and date to about years ago. The oldest related modern human fossils outside of Africa appear in the Middle East, dating from about years ago. They turn up in Europe, East Asia and Australia, with dates from about to years ago. How did modern humans evolve from their archaic form? Two models polarize a debate about the origin of modern humans. One model focuses on African replacement, proposing that modern humans evolved sometime between and years ago in Africa as a new and separate species, and that this species subsequently dispersed and replaced coexisting archaic human groups elsewhere in the Old World. This out of Africa model suggests that all of our ancestors from before years ago lived in subsaharan Africa and that no living humans descend from archaic human populations such as the Neanderthals of Europe and the Middle East. An alternative view is that humans have been evolving within a single evolutionary population, which, though structured, has been prevented from divergence into a new species within the last million years by gene flow. This multiregional model suggests that the ancestors of living humans trace back to a widely dispersed archaic population living not only in Africa several hundred thousand years ago, but perhaps across the entire Old World. Patterns of human genetic variation are most often interpreted as supporting the African replacement model, but they are also compatible with a multiregional transition. For example, the small estimated evolutionary population size has been argued as incompatible with a multiregional model because individuals are too few to spread between three continents at a density that would maintain gene flow. But, this number is an estimate of the potential number of ancestors, and is not an estimate of the census size. There is no widely accepted population genetic model available for estimating census size from genetic diversity. Recent ancestry, in particular for mtdna, has been taken as evidence of a recent speciation origin, but other loci have very different time-depths. Higher genetic diversity in Africa has been said to indicate an origin in Africa, but in fact, the characteristic pattern of this diversity indicates only a larger number of ancestors, not greater time-depth, within Africa. Larger genetic distances between populations in subsaharan Africa and those in Europe or Asia generate phylogenetic trees of populations with a characteristic primary split that has been interpreted as evidence of the migration of modern humans out of Africa. However, variable rates of genetic drift and gene flow between continental regions are a more likely explanation for observed geographic patterns of diversity. The main point is that genetic diversity data can be interpreted to fit either model for modern human origins, and therefore have not resolved the debate. Mitochondrial Eve Perhaps the most controversial genetic research pertaining to modern human origins has resulted from the use of mitochondrial DNA (mtdna). Most human DNA exists in the chromosome pairs within the nucleus of the cell and is inherited through the process discovered by Gregor Mendel both parents contribute half of their DNA to their offspring. But the small mtdna genome is inherited exclusively from the mother. Without the opportunity for recombination that occurs during meiosis for nuclear DNA, the simpler uniparenta1 mode of inheritance for mtdna makes genealogies relatively easy to reconstruct. However, because mtdna traces only through the maternal line, mtdna variants are subject to greater genetic drift than diversity in the nuclear genome. Any two people share a single mtdna ancestor at some point in the past. A less-related third person also shares a single ancestor with the first two, but further back in time. For any piece of DNA that does not recombine, a single common ancestral sequence exists from which the DNA of everyone now living descends. In 1987, Rebecca Cann and colleagues looked at mtdna variations in people from around the world and estimated that the most recent common ancestor appeared to have lived about years ago in Africa (referred to as Eve by the press and some scientists). The implication was that mtdna evidence supported a recent African origin for modern humans and that the multiregional hypothesis was rejected. Subsequent studies of the pattern of diversity in mtdna suggested an expansion of the modern human population from a small speciation bottleneck about 4

5 years ago, leading to a prediction that other loci in the genome would show the same signal. However, it has become clear from many genetic studies since that the most recent common ancestral sequences for diversity in other loci trace to different ancestors in different generations. Furthermore, it seems unlikely that the number of ancestors in any single generation have ever been fewer than several thousand. Eve does not necessarily tell us anything about the origin of modern humans. Y-chromosome Adam Given the interest in tracing female ancestors back to Eve, it was inevitable that scientists would turn their attention to tracing a genetic Adam by using polymorphisms on the Y-chromosome. The sex chromosomes come in two forms, X and Y, with a usual pattern where females have two X- chromosomes, and males have one X- and one Y- chromosome. Because the Y-chromosome is inherited exclusively through the father, and for the greater part does not recombine, it is not too difficult to reconstruct its most recent common ancestor. Like mtdna, the nonrecombining part of the Y-chromosome is also subject to greater drift than other loci in the nuclear genome. When populations are in expansion phases there is less opportunity for drift, but at other times, mtdna and Y- chromosome diversity is expected to leak out faster than diversity in the recombining nuclear genome. Y-chromosome variations show relatively low levels of global diversity with a relatively high proportion differing between populations. The low global diversity is consistent with either a selective sweep or substantial genetic drift. There have been several estimates for the age of the most recent common ancestor, and the latest work suggests a date in the neighbourhood of years ago. As with analyses of mtdna, such results do not tell us about modern human origins but probably do imply fluctuations in ancestral population sizes. Patterns of Y-chromosome diversity are also compatible with gene flow in both directions, in and out, of subsaharan Africa. Y-chromosome data have also been taken to indicate that males are the sex who stay close to home and contribute less to gene flow than their partners. However, DNA variants with estimated ages less than years, are typically population-specific at all loci. The question that remains is what combination of gene flow, genetic drift, and possibly selection, accounts for the success and global distribution of this recent ancestral Y-chromosome. Until recently, there was no way of assessing genetic variation in prehistory except for inferences obtained from analyses of the fossil record. This situation changed in 1997 when mtdna was successfully extracted from the original Neanderthal specimen discovered at Feldhofer Cave in the Neander Valley, Germany. This fossil is undated but is probably from an individual who lived about years ago. In early 2000, an mtdna sequence was obtained from a second Neanderthal specimen, this time from Mezmaiskaya Cave in the northern Caucasus, and dating to years ago. Many anthropologists have long argued over the distinctiveness and evolutionary status of the Neanderthals. Some favour classifying the Neanderthals as a subspecies of Homo sapiens (H. sapiens neanderthalensis as opposed to modern humans, who are often referred to as H. sapiens sapiens), and others argue for separate species status (Homo neanderthalensis). The fate of the Neanderthals has also been fiercely debated, with views ranging from a multiregional perspective of Neanderthals being part of our ancestry, and others arguing for complete replacement in Europe. Sequence analysis of Neanderthal mtdna allows direct assessment of prehistoric genetic variation, although the small sample size of two specimens with data taken from only one gene locus must be kept in mind. When compared with the human reference standard, the Feldhofer Cave specimen differs at 27 mtdna sites and the Mezmaiskaya Cave specimen differs at 23 mtdna sites. These differences are larger than are typically found between random pairs of living humans (who average about eight site differences). It is not clear, however, exactly what this difference implies. Are Neanderthals different because they belonged to a different species, or because they lived thousands of years ago? No mtdna sequence data of a modern human from the same time period is yet available. For comparison, global mtdna diversity within modern humans has a similar magnitude to levels within geographic subspecies of chimpanzees. So it seems that Neanderthals are not related within the same subspecies as living humans. However, the difference between Neanderthals and living humans is not greater than variation between chimpanzee subspecies. This comparison suggests that human populations may have been more greatly structured in the past and that Neanderthals may have been a race of Homo sapiens. If so, there was no species barrier to prevent admixture between modern and Neanderthal groups, but we still do not know if any people living now carry genetic diversity that descends from Neanderthal ancestry. Neanderthal Enigma The studies discussed so far all attempt to make inferences about the past based on genetic variation in living humans. Summary Population genetics provides models for investigating the balance of evolutionary forces acting on genetic diversity. 5

6 Studies that use these models have found that the evolution of contemporary human genetic diversity has occurred over the past several hundred thousand years or longer. Our species is geographically widespread, but shows low levels of differences among population groups suggesting persistent levels of gene flow as well as dispersal. It is difficult to classify humans into groups by their DNA profiles, and impossible to successfully apply a biological concept of race to diversity within living human populations. The origin of modern human genetic diversity is still widely debated. Genetic data indicates the importance of Africa in modern human evolution, in line with the observations from the fossil record of the first appearance of modern anatomical form in Africa. Whether Africa is the only region that we can trace our ancestors to, or whether it is the primary region remains to be seen. Some genetic evidence does suggest ancient contributions in southern Asia, a region where the fossil evidence for replacement is equivocal. It may be the case that our origins are best described as mostly (but not exclusively) out of Africa. Further Reading Foley R (1998) The context of human evolution. Genome Research 8: Hawks J, Hunley K, Lee S-H and Wolpoff M (2000) Population bottlenecks and Pleistocene human evolution. Molecular Biology and Evolution 17: Jorde LB, Bamshad M and Rogers AR (1998) Using mitochondrial and nuclear DNA markers to reconstruct human evolution. BioEssays 20: Ovchinnikov IV, Go therstro m A, Romanova GP et al. (2000) Molecular analysis of Neanderthal DNA from the northern Caucasus. Nature 404: Przeworski M, Hudson RR and Di Rienzo A (2000) Adjusting the focus on human variation. Trends in Genetics 16: Relethford JH (1998) Genetics of modern human origins and diversity. Annual Review of Anthropology 27: l 23. Relethford JH (2001) Genetics and the Search for Modern Human Origins. New York: John Wiley & Sons. Templeton AR (1999) Human races: a genetic and evolutionary perspective. American Anthropologist 100: