Demographic Events and Evolutionary Forces Shaping European Genetic Diversity

Abstract

Europeans have been the focus of some of the largest studies of genetic diversity in any species to date. Recent genome-wide data have reinforced the hypothesis that present-day European genetic diversity is strongly correlated with geography. The remaining challenge now is to understand more precisely how patterns of diversity in Europe reflect ancient demographic events such as postglacial expansions or the spread of farming. It is likely that recent advances in paleogenetics will give us some of these answers. There has also been progress in identifying specific segments of European genomes that reflect adaptations to selective pressures from the physical environment, disease, and dietary shifts. A growing understanding of how modern European genetic diversity has been shaped by demographic and evolutionary forces is not only of basic historical and anthropological interest but also aids genetic studies of disease.

Studies of European genetic diversity are some of the largest to have been performed in any population. They provide insight into how people settled and migrated through Europe and adapted to their environment.

During classical antiquity, writers such as Herodotus chronicled the expansion and contraction of empires, as well as the traditions of the peoples associated with them. Julius Caesar's memoirs from the Roman conquests of Gaul detail his encounters with foreign tribes such as the Helvetii and the Belgae. Such accounts were fascinating to peoples of that era as humans lived largely in ignorance of other cultures beyond their relatively small geographical vicinity. In the modern world, the barriers to acquiring knowledge of other contemporary societies are small; we can now easily learn about populations from across the world through an abundance of sources. Instead, the major challenge is to discern whom the peoples of the past were. From the perspective of genetics, we are especially curious about how past demographic and evolutionary events influenced the genetic diversity in humans today. However, peering into the past poses major challenges, and, in some ways, we stand much like Herodotus and Caesar, trying to piece together an understanding of distant populations from limited contact and partial experiences.

For geneticists, Europe represents a uniquely well-studied region of the world. On the one hand, it has a richness of accessible sources. We have already mentioned historical accounts beginning with the ancient works of Herodotus; there has also been an abundance of archaeological, anthropological, and linguistic studies. More recently there has been substantial interest in understanding the genetic history of modern Europeans. Indeed, many of the largest studies of the genetics of human populations have taken place in Europe. This is, in large part, because of the availability of European universities and biomedical centers, which have provided the infrastructure for such “big science” studies that other regions have traditionally lacked. On the other hand, European human diversity has at various times been highly politicized, which has led to deeply misguided perspectives on the subject of genetic superiority and some of the most atrocious abuses to human life—the genocides and eugenics of the first half of the 20th century.

Contemporary genetic studies in Europe still work under the shadow of such views that are now understood as being scientifically without merit as well as ethically wrong, and, as today's scientists, we must be sensitive to the potential future misuse of findings regarding genetic diversity. That said, the field has been reinvigorated during the past approximately 50 years as perspectives on human diversity, both cultural and genetic, have matured. Scientifically, it is now appreciated that the genetic differences among humans are, in absolute terms, small as first identified by Lewontin (1972) (also see Chakravati 2014). Simplistic notions of genetic determinism have also fallen aside as most human traits are now thought to be driven by complex interactions between multiple environmental and genetic factors. Culturally, there is a wider appreciation that diversity makes a positive contribution to society. And finally, it is now recognized that understanding background patterns of genetic diversity is an essential component for combating heritable and infectious diseases.

Thanks to the growing interest in human population genetics, the scale of recent studies of European genetic diversity has grown to a staggering extent. Studies involving Europeans are some of the largest to have been performed in any population, regardless of the species. As a result, research on genetic diversity in Europe is of interest not just to scientists examining other human populations around the world but to all students of genetic diversity.

MAJOR HYPOTHESIZED DEMOGRAPHIC FACTORS SHAPING GENETIC DIVERSITY IN EUROPE

In a landmark review, Barbujani and Goldstein (2004) highlighted three major prehistoric demographic events that are likely to have had an impact on European genetic diversity (Fig. 1). The first event they described was the initial colonization of Europe by hunter–gatherers approximately 40,000 years ago. These travelers, likely consisting of only one or a few small groups that split off from established populations in northeast Africa near the Levant, therefore, represented only a small subset of the total human genetic diversity present within Africa. The entryway for humans into Europe is thought to have been via the Near East through modern-day Turkey, after which they expanded to the northwest either using a direct route or by first traveling along the coasts of the Mediterranean before turning northward. One plausible genetic signature of this process would be the reduction of genetic diversity as one moves northward and westward in Europe.

Figure 1.

Schematic map of Europe showing the major migrations events that may have influenced modern-day European genetic diversity. The map shows (1) the initial colonization of Europe approximately 40,000 years ago from the Middle East (blue arrows), (2) the contraction of humans into four major refugia during the LGM approximately 18,000 years ago followed by the subsequent recolonization of Europe (green arrows), (3) the movement of Neolithic famers approximately 10,000 years ago from the Fertile Crescent (red arrows), (4) various barbarian migrations into the Roman Empire during the migration period of ∼400–800 CE (dashed yellow arrows), and (5) continued gene flow across the Mediterranean Sea and between the Iberian Peninsula and North Africa throughout human history (dashed orange arrows). (This map is adapted from that presented in Renfrew 2010, which was originally made by Alessandro Achilli and Antonio Torroni.)

Second, Barbujani and Goldstein highlighted the constriction of northern European populations southward caused by the expansion of ice sheets during the last glacial maximum (LGM) at the end of the Pleistocene era (approximately 18,000 years ago). Evidence for the impact of Pleistocene glacial movements on genetic diversity has been observed in studies of a myriad of animal and plant species (Hewitt 1999). It seems unlikely that humans would have been exempt from the effects of these dramatic climatic shifts. The genetic signatures that may characterize this period may be differentiation among modern European populations reflecting the geographic locations of the different ancient glacial refugia, as well as less genetic diversity in northern Europe because of larger, more stable population sizes in the south and postglacial recolonization northward (Fig. 1).

The third possible event noted by Barbujani and Goldstein (2004) was the expansion of the first farmers into Europe following the emergence of agriculture in the Near East approximately 10,000 years ago, which signified the dawn of the Neolithic era. The spread of agriculture observed in the archaeological record may have been strictly cultural, but many researchers have argued that this was accompanied by a demographic expansion in which incoming agricultural communities displaced hunter–gatherer groups. Genetic signatures of this “demic diffusion” hypothesis would be a gradient in gene (allele) frequencies from the Near East toward Western Europe, as well as that of mixing between immigrant Neolithic and resident Paleolithic populations.

At the time of the Barbujani and Goldstein (2004) review, there was little evidence for the importance of interbreeding between modern humans emerging out of Africa and closely related forms of archaic humans, such as Neanderthals, who were already established across much of Europe. However, the sequencing of a Neanderthal genome (Green et al. 2010) has since provided evidence that up to 4% of current non-African human genetic ancestry (see Box 1) is from Neanderthal sources. However, whether there was admixture specifically with Paleolithic Europeans or whether the Neanderthal ancestry found in modern Europeans derives from gene flow that occurred in the ancestors of all non-Africans shortly after they left Africa is still largely unknown. It will be interesting to see if more detailed studies in the future can shed further light on this question. For example, there is emerging evidence of a distinct period of gene flow with the ancestors of contemporary Asians (Wall et al. 2013).

BOX 1. GENETIC ANCESTRY.

Throughout the article, we often refer to the concept of the “genetic ancestry” of an individual. Although it is a somewhat ambiguous and often loosely applied term, scientifically, genetic ancestry is based on the idea that any particular individual is part of a large human pedigree and related to certain ancestors in this pedigree moving back in time (i.e., two parents, four grandparents, eight great grandparents, etc.). In theory, the number of ancestors increases exponentially, but in reality there can only be a finite number for all persons. When genetic material is passed from an ancestor to a descendent in this pedigree, a process called recombination reshuffles the two chromosomes a particular individual inherited from their two parents, after which this new mix of chromosomes is passed on to a subsequent child. The consequence of this process is that any individual’s genome is made up of a mosaic of genetic material from many of their ancestors back in time. If we take an arbitrary point back in time, some of these ancestors may derive from a certain population, whereas the remainder may derive from another genetically distinct population. For example, for an African American individual, 75% of their genome may derive from ancestors from sub-Saharan African, whereas 25% of their genome may have been inherited from European individuals. Similarly, a modern-day European individual could possess genetic ancestry from both northwestern and southeastern European ancestors.

Beyond these events, three other demographic factors also likely have played at least some role in shaping modern European genetic diversity. One is what historians refer to as the migration period of Europe from ∼400–800 CE (Common Era). Historical records suggest that there was an extensive invasion of the Roman Empire by barbarian tribes such as Goths, Lombards, and Slavs, and many modern-day nation states trace their identity to these peoples. However, historians such as Patrick Geary have called into question whether mass movements of large cultural units actually took place (Geary 2003). Instead, he suggests that the ethnogenesis of many groups represented the influence of a military or aristocratic elite that opportunistically drew adherents from local populations; thus, there would have likely been very little impact on modern genetic diversity.

Another factor is the contact of European peoples with those from neighboring geographical regions. Although the borders of Europe are generally considered the Mediterranean Sea and Caucasus Mountains, these are not strict impediments to gene flow. Admixture between North Africans and Europeans almost certainly occurred during the Moorish occupation of the Iberian Peninsula, and there is an age-old complex system of commerce across the Mediterranean. Likewise, contacts with western Asia and the Middle East via trade, or via more ancient contacts, likely also contributed to European genetic diversity.

Finally, among all of these major migrations, one must not lose sight of the behavior of the typical individual. Like most humans from across the world, the average European would likely find a mate locally. In some cultures, an average male may have moved less, as marriage practices may be distinct between the genders and show a patrilocal pattern in which wives move further from their birthplace by relocating to the birth location of their husbands. The long-term consequences of this local, sex-biased dispersal to a specific region would be a genetic similarity among distinct geographic neighbors, as well as differences in genetic material inherited through male and female lines (i.e., the Y chromosome and mitochondrial DNA).

How the various demographic factors described above have meshed to create the current diversity of modern Europeans is a fascinating yet complicated question to answer. For instance, our summary has highlighted several events that all predict a higher degree of genetic diversity in southern than northern Europe, making it difficult to identify which specific events truly took place as the genetic signatures of the events would overlap. This situation has eloquently been compared with a “palimpsest,” an ancient manuscript that has been partially cleaned for novel writing, such that traces of the original writing remain (Jobling et al. 2003). Therefore, we should recognize that it is simplistic to think that genetic signatures of these distinct events can be clearly delineated. With these complexities in mind, we now turn to what has been learned about patterns of human genetic diversity from contemporary Europeans.

GENES AND GEOGRAPHY IN EUROPE VIA GENOME-WIDE DATA

In summarizing the patterns of diversity in European genetic data, we will take a personal perspective based on some of the studies we have been a part of, and then we will link out to studies that have preceded and followed our own work. In 2005–2006, DNA microarrays (a technology that uses a large number of DNA probes arrayed at microscopic scale on a solid surface) began to emerge as an efficient technology to determine which genetic sequence variants an individual carried (their genotype) at hundreds of thousands of individual locations in their genome in a single experiment. The costs became low enough that these arrays could be tested on thousands of individual samples, opening the way for large-scale population-based genetic studies that attempted to link disease susceptibility to particular genomic regions. As these population-based studies were being performed, a major concern was whether subtle mismatches in genetic ancestry between disease cases and disease-free or random controls would lead to false-positive disease-causing alleles being implicated because of natural allele frequency differences between these two groups. To help avoid this, one might try to match every case individual with a matched control of the same genetic ancestry. One effort was to produce a large reference data set of healthy controls: the POPRES study (Nelson et al. 2008). The European subsample of the POPRES data set contained genotypes at 500,568 locations in the genomes of 3192 individuals from 37 different populations (which mostly represented countries of birth for the individual or their grandparents).

One of the major patterns to immediately emerge from a study on this data by one of us (J.N.) was that genetic relationships among individuals were strongly correlated with their geography, such that a visual summary of the genetic data looked strikingly similar to the geographic map of Europe (Fig. 2) (Novembre et al. 2008). Notably, we did not see discrete subgroups of individuals, which might arise, for instance, if a major factor affecting modern European genetic diversity had been its subdivision into separate glacial refugia or if major language groups (e.g., Romance, Slavic, Germanic) had created substantial long-term barriers to local mating. The only major genetic outliers were five Italian samples, which, based on later analyses, we now suspect are from Sardinia, an isolated island population in the Mediterranean (see below). At roughly the same time, other groups were finding similar results (Heath et al. 2008; Lao et al. 2008). The overall pattern to emerge from these studies of Europe is of a subtle genetic continuum among modern Europeans that is determined by geography. Recent work by Ralph and Coop (2013) suggests that much of this signal can be observed even while focusing on recent common ancestry, and that the most geographically separated Europeans have shared hundreds of common ancestors within the last 3000 years.

Figure 2.

A statistical summary of genetic data from 1387 Europeans based on the two components that explain the greatest variation in the data. Small colored labels represent individuals and large colored points represent median coordinate values for each country. The inset map provides a key to the labels. The axes are rotated to emphasize the similarity to the geographic map of Europe.

From this broad view of genetic variation in Europe, it is not clear what impact the many historical and demographic factors alluded to at the beginning of this article have had on their genomes. If one looks more carefully, it is possible to uncover details that point to the importance of demographic processes involving the north/south (N/S) axis of Europe. For instance, when conducting a principal component (PC) analysis (a statistical method of summarizing highly multidimensional data into the most important elements of variation), one output is the relative amount of variation explained by the individual PC axes (dimensions). In our analysis, PC1 explained 0.30% of the variation, whereas PC2 explained 0.15%. Both of these numbers are small by typical standards in population genetics and reflect the fact that there is very little variation among European subgroups (recall humans are all very genetically similar; see Chakravarti 2014). However, the percentage for PC1 is twice that for PC2, arguing that the north northwest/south southeast axis explains more of the genetic diversity. However, these results must be interpreted with great caution. PC axis directions are very difficult to interpret, in part, because they can be very sensitive to how uniformly individuals are sampled and analyzed with regard to genetic ancestry (Novembre and Stephens 2008; McVean 2009). Indeed, a study that shortly followed ours found PC1 ran west/east, largely because they had larger numbers of individuals sampled along an east–west axis (Heath et al. 2008). A more robust indicator of a roughly N/S axis are gradients in haplotype diversity (haplotypes are particular combinations of sequence variants at multiple locations along the sequence and tend to have greater resolution with regard to the inference of demographic processes compared with variants at any single site). In a second article on the POPRES sample (Auton et al. 2009), we described a N/S gradient in haplotype diversity and a similar result was reported by Lao et al. (2008).

Such a gradient in haplotype diversity could be consistent with (a) serial founder effects in an initial northward expansion from the south of Europe during the Paleolithic, (b) subsequent expansions from the south after the retreat of the glaciers at the end of the LGM, (c) a northward expansion of agriculturalists in the early Neolithic, (d) admixture across the Mediterranean from North Africa, or (e) simply higher sustained population sizes through time in the south attributable to better long-term survival conditions. In our article, we argued that admixture across the Mediterranean certainly contributes to the N/S pattern (process d), and more extensive work by others supports this hypothesis (Botigué et al. 2013). To what extent this admixture is layered on top of other patterns (caused by processes a, b, c, and e listed above) is still a very open question and shows the challenges of interpreting the results of a palimpsest pattern of genetic variation.

Although, generally, the level of genetic similarity among European individuals tends to be highly correlated with their geographic origin, there are certain groups of individuals, termed “population isolates,” such as Sardinians, who because of particular features (e.g., a distinct cultural trait or topographical barrier) may be substantially more isolated from their neighbors than average. One of the exciting aspects of genome-wide data has been the ability to analyze these isolates in more detail and tease out patterns of genetic isolation not clearly visible using smaller amounts of genetic data. In one study, we aimed to genetically characterize one of the less-studied population isolates in Europe, the Sorbs of eastern Germany (Veeramah et al. 2011). The Sorbs are of interest to historians as they speak a West Slavic language and have maintained their traditional Slavic customs but have found themselves surrounded by Germanic speakers because of the complex population movements that occurred during the migration period of Europe. To provide a control, we compared the Sorbs to two relatively well-studied population isolates, the French Basques and Sardinians. Basque country is a mountainous region along the France/Spain border that lies in the epicenter of one the major refugia of the LGM (Achilli et al. 2004). Approximately 25% of people from the region speak a language isolate, Euskara, which has been proposed to descend from the languages present in Western Europe before the arrival of Indo-European speakers. As such, it has been hypothesized that the Basque people may have remnants of the genetic diversity of the upper Paleolithic era (Cavalli-Sforza 1988). Similar hypotheses have been advanced for the people on Sardinia (Cavalli-Sforza et al. 1994), where the Mediterranean Sea has presented a formidable barrier to large-scale immigration.

Consistent with their medieval origins, the Sorbs showed greater genetic affinity to individuals from other western Slavic–speaking populations, such as Poles and Czechs, than they did to neighboring Germans. However, the evidence for genetic isolation in the Sorbs was very subtle compared with that observed for the Basques and Sardinians, who were clear outliers from the general European pattern of genetic variation and consistent with the presence of much stronger and older barriers to gene flow. The finding of Basque genetic isolation was particularly interesting given that two previous studies (Garagnani et al. 2009; Laayouni et al. 2010) could not find any evidence of differentiation of Basque populations from their Spanish and French neighbors using a much smaller number of genetic markers (approximately 100 compared with the approximately 30,000 we used). Thus, this study showed the considerable power of genomic data for studying human European populations.

CLASSICAL STUDIES OF EUROPEAN GENETIC DIVERSITY USING UNIPARENTAL MARKERS

Genome-wide array data have clearly proved to be a very powerful tool for studying European genetic diversity. However, building on earlier studies that examined classical markers, such as blood groups and protein polymorphisms (Cavalli-Sforza et al. 1994), most of our understanding of the peopling of Europe based on genetic studies from the last approximately 25 years comes from studying the nonrecombining portion of the Y chromosome (NRY) and mitochondrial DNA (mtDNA). These markers are simple to study because they are transmitted exclusively down the paternal and maternal lines, respectively (they are “uniparental” markers).

A common framework for examining uniparental data is to assign individuals to “haplogroups” (a reference haplotype) based on whether they share a common set of genetic variants along the known NRY or mtDNA tree. The distribution of mtDNA haplogroups across Europe has been shown to be remarkably uniform with the presence of three major clades: (H,V), (J,T), and (U3,U4,U5,K) (Torroni et al. 2006; Underhill and Kivisild 2007). Haplogroup H is the most abundant across Europe with frequencies ranging from 40% to 60%. Interestingly, its frequency is greatest in Basques, and it is believed that subclades such as H1 and H3 all originated from this single region and expanded out across Europe at the end of the LGM (Achilli et al. 2004). The NRY appears to be comparatively much more geographically structured with certain haplogroups having a peak frequency in a particular geographical region, perhaps reflecting the greater level of patrilocality touched on earlier (Seielstad et al. 1998). For example, the Y-chromosome haplogroup R1b has frequencies that are very high in Western Europe, whereas haplogroup R1a tends to be more common in the east. Because of their apparent clinal distribution from east to west, haplogroups J, E1b1b (Semino et al. 2004), and R1b (Balaresque et al. 2010) have all been suggested to have once been carried by male Neolithic famers, although such theories are not necessarily universally accepted (e.g., see Busby et al. 2012).

The NRY and mtDNA have also proved useful for examining much more specific migration events from the historical (rather than prehistorical) era (Jobling 2012). For example, Zalloua and colleagues (2008) argue that the Phoenicians, a culture that exerted great influence on the Mediterranean during the 1st century BCE (Before the Common Era), contributed >6% of paternal lineages to modern-day populations with a history of Phoenician contact.

ANCIENT DNA STUDIES IN EUROPE

Although there are considerable challenges with regard to degradation and contamination of DNA samples, the study of ancient DNA (paleogenetics) (Pääbo et al. 2004) is a promising avenue for understanding the peopling of Europe because it provides a direct account of diversity in the past. The vast majority of studies to date have exclusively examined mtDNA because, as thousands of copies are present per cell as compared with just one copy for nuclear DNA, it tends to be much more abundant in ancient remains. Generally, it appears there is substantial discontinuity between both Paleolithic and Neolithic as well as (though to a lesser extent) Neolithic and modern mtDNA pools in central and northern Europe (Bramanti et al. 2009; Malmström et al. 2009; Haak et al. 2010). There has, therefore, likely been significant post-Neolithic reshaping of European maternal lineages (e.g., during the metal ages) (Deguilloux et al. 2012), although greater mtDNA Neolithic continuity has been observed in Western Europe (Sampietro et al. 2007; Lacan et al. 2011).

However, mtDNA offers a view of genetic diversity that is limited to maternal lineages only. Fortunately, the development of next-generation sequencing has been revolutionary for paleogenetics (Stoneking and Krause 2011). Whole genomes from a Neanderthal (Green et al. 2010) and other archaic humans (Reich et al. 2010), as well as an approximately 4000-year-old Paleo-Eskimo from Greenland (Rasmussen et al. 2010), have now been sequenced; very recently, we saw the publication of the complete genome of a human European mummy commonly referred to as the Tyrolean Iceman or “Ötzi” (Keller et al. 2012).

Ötzi was discovered in the Eastern Alps near the Austrio-Italian border and dates back to the Late Neolithic/Early Copper Age (5350–5100 thousand years ago). Interestingly, Ötzi’s genome clustered with Sardinians, some 500 miles away from where the sample was found. Thus, the study points to a significant change in the geographic distribution of genetic diversity since the Late Neolithic, and that modern Sardinians have likely conserved some of this ancient heritage. Two other research groups have since managed to sequence more than 10 million base pairs from late hunter–gatherers and Neolithic farmers in Scandinavia (Skoglund et al. 2012), as well as two Mesolithic hunter–gatherers in northwestern Spain (Sánchez-Quinto et al. 2012). It seems likely that other such studies will emerge soon giving us an unparalleled view of the processes that have led to the current distribution of European genetic diversity.

DARWINIAN SELECTION WITHIN EUROPE

Thus far, we have discussed the major patterns of variation that we observe when looking at genetic regions that are mostly presumed to be unaffected by positive natural selection. As mentioned above, the vast majority of human genetic diversity is thought to be shaped primarily by demographic processes such as population growth, dispersal patterns, and migrations. Some very small fraction of variants are, however, new mutations that rapidly increase in frequency because they confer some reproductive advantage to individuals that carry them resulting in their higher evolutionary fitness. One hypothesis for their scarcity in our genome is that the probability of a mutation being beneficial for an organism, rather than being deleterious or selectively neutral, is very small. However, the scale of population genetic surveys over the past decade has made it possible for researchers to identify some of these selectively adaptive variants among the much larger sea of variants that are mainly impacted by demographic history.

The selected variants discovered thus far seem to fall into three major biological categories: variants affecting human immunity, human external morphology (such as hair and eye color), and human dietary metabolism. For example, one of the most dramatic variants to have been discovered in Europeans occurs in a locus affecting lactase persistence, LCT. Using metrics that describe allele frequency differences and/or reduced haplotype diversity, the pattern of genetic diversity at this gene sticks out among the backdrop of European diversity like a skyscraper from a prairie (Fig. 3). In this region, there is a single haplotype that is very common in many populations (Bersaglieri et al. 2004; Itan et al. 2010), presumably because it carries an advantageous variant. Specifically, it is thought the underlying functional mutation confers to its carriers the considerable nutritional benefit of the ability to digest milk more easily as adults.

Figure 3.

Maps showing the frequency distribution of individual genetic variants in European populations. The top row shows three variants in the genes LCT, SLC45A2, and CCR5 that are thought to be under positive selection (selected variants are the green wedges). The bottom row shows three randomly chosen variants (A–C) with a minor allele frequency of >5% in all Europeans. It is noteworthy that the top row shows a high level of regional structuring (i.e., the selected allele for LCT is generally at much higher frequency in the northwest), whereas variants in the bottom row show a relatively flat distribution across the continent. The LCT data is from Itan et al. (2010), SLC45A2 data is from Lucotte and Yuasa (2011), and CCR5 data is from Novembre et al. (2005). In addition, all selected variants are supplemented by data from the ALFRED database (Kidd et al. 2003). The three random variants are based on frequencies in the POPRES sample.

A handful of genes affecting external morphological traits, such as eye color and skin pigmentation, also show similarly extreme patterns. For example, SLC45A2 (Lucotte and Yuasa 2011) affects skin pigmentation and possesses a genetic variant that is found at very different frequencies across the globe. In these cases, novel variants have been driven to high frequency because they facilitate the interaction of their carriers with their physical environment (Jablonski 2008).

There have also been many pathogens that have ravaged Europe across time, and variants at genes underlying adaptive immunity, such as the MHC (Meyer and Thomson 2001), and innate immunity, such as TLR6 (Pickrell et al. 2009), show signatures of positive selection, which are likely the result of adaptation to these unique pathogen exposures. A particular gene of interest is CCR5, which harbors a 32-base pair deletion known to confer resistance to HIV in humans, and in which the highest local frequency is found in Europe (16%). Some have suggested that its unique frequency in Europe is because the variant also conferred resistance to the bubonic plague that ravaged Europe during the Middle Ages, but the story appears to be much more complex and may not involve resistance to an ancestral pathogen at all (see review in Novembre and Han 2012).

The selection pressures just described are not unique to Europe. For example, in East Africa, other instances of the lactase persistence variant have arisen in the LCT genomic region, and, in high latitudes of Asia, other variants in different genes that lighten skin pigmentation have spread to high frequency (see review in Gomez et al. 2014). Therefore, it seems that when distant human populations are faced with similar selection pressures, independent evolutionary responses can take place.

CONCLUDING REMARKS

Looking to the future, we expect that studies of European genetic diversity will continue to untangle how people settled and migrated through Europe and adapted to their environment. Although this could all be considered fascinating simply from a purely historical or anthropological perspective, there are also important current practical applications of this knowledge. As previously mentioned, the development and widespread use of high-throughput array-based genotyping was largely driven by research comparing sets of disease cases with sets of control individuals at variants from across the genome. Differences in genetic ancestry, even minor ones (Novembre et al. 2008), between cases and controls can lead to significant numbers of spurious signals of apparent disease association, especially when examining many genetic markers (Hirschhorn and Daly 2005). One of the first examples of this for Europeans was a study that showed highly significant association between height and the variant most commonly associated with lactase persistence in Europeans, LCT C-13910T. However, this result was simply an artifact of population genetic stratification because Europeans of both northwest and southeast genetic ancestry were present in the cohort, and accounting for this feature resolved this artifact (Campbell et al. 2005). Thus, a good understanding of the patterns of genetic diversity observed in present-day Europeans (as well as others of recent European ancestry) is invaluable, both for statistically correcting for differences in ancestry and the initial choice of study population(s).

Understanding European genetic diversity is also important for predicting the ancestry of genetic data from samples of unknown European origin. Clearly accurate inference is of great value for forensic applications, but there is also considerable interest from the public, from a more personal or recreational viewpoint, with the geographic ancestry testing services of companies such as Family Tree DNA, 23andMe, and Ancestry.com. Although the issues involved are too complex to discuss here (e.g., Royal et al. 2010), in light of the unregulated nature of this industry, great care needs to be taken in the interpretation of these ancestry results. This is especially true today as many companies have shifted from examining NRY and mtDNA lineages only, which are limited in power and scope but yield relatively simple genealogical interpretations, to ancestry estimation from whole genome data. Although powerful insights can be gained regarding populations, inferences drawn from the analysis of any one particular individual’s data can be highly misleading; for example, an individual interpreted as being of central European ancestry can, in fact, be of mixed northern and southern European ancestry. However, if applied appropriately, there are potentially exciting opportunities that can empower individuals who desire more information about their ancestors beyond that which is available from more traditional historical sources.