supporttheunderdog wrote:I do not accpet you have proven the point at all. the two news paper items do not prove anything: only the 2002 Weale study contained any data on the genetics but this study was flawed in that it was not representative of the country as whole and it was superceeded by the far more extensive Capelli study in 2003, and the Sykes study, which showed, from a far larger sample pool, that the impact on the Geene pool of the Anglo Saxons was generally small, with some local variations.
Rather it is you who are being selective and completely failing to provide comprehensive and substantive evidencem based on comprehensive and extensive sampling accross the whole of the UK, as oppossed to very local sampling which is unrepresentative of the country as a whole.
Seems you have
forgotten your own main point. Therefore, I shall remind you:
supporttheunderdog wrote:the English are NOT predominantly descended from the Angles, Saxons and Jutes, but probably from Neolithic peoples who probably came to Brtitain from Spain
... reiterated as:
supporttheunderdog wrote:The Anglo-Saxon genetic origin is something of myth.
... and you went on to give us your selected "current" supportive "evidence" from much flawed and outdated studies from 2002/3.
supporttheunderdog wrote: MY POSTION REMAINS THAT BASED ON POST 2002 STUDIES THE PRIMARY ANCESTORS OF THE BRITISH ARE THE PEOPLES WHO ORIGINATED IN THE IBERIAN PENINSULAR
The Neolithic-Iberian connection was dis-proved to you with this later study:
European Journal of Human Genetics (2005) 13, 1293–1302.
The place of the Basques in the European Y-chromosome diversity landscape
Abstract
There is a trend to consider the gene pool of the Basques as a 'living fossil' of the earliest modern humans that colonized Europe. To investigate this assumption, we have typed 45 binary markers and five short tandem repeat loci of the Y chromosome in a set of 168 male Basques. Results on these combined haplotypes were analyzed in the context of matching data belonging to approximately 3000 individuals from over 20 European, Near East and North African populations, which were compiled from the literature. Our results place the low Y-chromosome diversity of Basques within the European diversity landscape. This low diversity seems to be the result of a lower effective population size maintained through generations. At least some lineages of Y chromosome in modern Basques originated and have been evolving since pre-Neolithic times. However, the strong genetic drift experienced by the Basques does not allow us to consider Basques either the only or the best representatives of the ancestral European gene pool. Contrary to previous suggestions, we do not observe any particular link between Basques and Celtic populations beyond that provided by the Paleolithic ancestry common to European populations, nor we find evidence supporting Basques as the focus of major population expansions.[/b]
..........................................
Then I showed you how the most current, and to date, largest European-wide study showed that the populations of Europe follow the historical and geographic localities already well established, with no contradictions to the known good historical records and observing the barriers for Finland (known linguistic oddity) and the Alps barriers in Italy.
Funnily enough, as well as directly relating the "English" (Londoners - I already gave you the Viking relationship ongoing research from the North of "England") to their ancestors, which are:- the
Germanic tribal region - the study also went on to suggest that the British have the
least differences from the other Europeans, (the opposite of the unique Finns). Which means what? Perhaps that the "English" are the MOST mixed up/heterogenous group (least pure) of
all the Europeans.
I thought you were lazy and didn't look up the
original paper refered to in the New York Times; but maybe you
knew what was coming and didn't want to acknowledge your error in clinging to those 2002 studies
So, apologies to others, but here is the complete article:
Correlation between Genetic and Geographic Structure in Europe
CURRENT BIOLOGY Volume 18, Issue 16, 26 August
2008, Pages 1241-1248
Oscar Lao1, 22, Timothy T. Lu2, 22, Michael Nothnagel2, Olaf Junge2, Sandra Freitag-Wolf2, Amke Caliebe2, Miroslava Balascakova3, Jaume Bertranpetit4, Laurence A. Bindoff5, David Comas4, Gunilla Holmlund6, Anastasia Kouvatsi7, Milan Macek3, Isabelle Mollet8, Walther Parson9, Jukka Palo10, Rafal Ploski11, Antti Sajantila10, Adriano Tagliabracci12, Ulrik Gether13, Thomas Werge14, Fernando Rivadeneira15, 16, Albert Hofman16, André G. Uitterlinden15, 16, Christian Gieger17, 18, Heinz-Erich Wichmann17, 18, Andreas Rüther19, Stefan Schreiber19, Christian Becker20, Peter Nürnberg20, Matthew R. Nelson21, Michael Krawczak2, 23 and Manfred Kayser1, 23, ,
1 Department of Forensic Molecular Biology, Erasmus University Medical Center Rotterdam, 3000 CA Rotterdam, The Netherlands
2 Institut für Medizinische Informatik und Statistik, Christian-Albrechts University Kiel, D-24105 Kiel, Germany
3 Institute of Biology and Medical Genetics, University Hospital Motol and 2nd School of Medicine, Charles University Prague, CZ 150 06, Prague 5, Czech Republic
4 Unitat de Biologia Evolutiva, Pompeu Fabra University, 08003 Barcelona, Catalonia, Spain
5 Department of Neurology, Haukeland University Hospital and Institute of Clinical Medicine, University of Bergen, 5021 Bergen, Norway
6 Department of Forensic Genetics and Forensic Toxicology, National Board of Forensic Medicine, SE 581 33 Linkoping, Sweden
7 Department of Genetics, Development, and Molecular Biology, Aristotle University of Thessaloniki, GR-540 06 Thessaloniki, Greece
8 Laboratoire d'Empreintes Génétiques, EFS-RA site de Lyon, 69007 Lyon, France
9 Institute of Legal Medicine, Medical University Innsbruck, A-6020 Innsbruck, Austria
10 Department of Forensic Medicine, University of Helsinki, Helsinki FIN-00014, Finland
11 Department of Medical Genetics, Medical University Warsaw, 02-007 Warsaw, Poland
12 Istituto di Medicina Legale, University of Ancona, I-60020 Ancona, Italy
13 Molecular Neuropharmacology Group and Center for Pharmacogenomics Department of Neuroscience and Pharmacology, University of Copenhagen, 2200 Copenhagen, Denmark
14 Research Institute of Biological Psychiatry and Center for Pharmacogenomics, Mental Health Center Sct. Hans, Copenhagen University Hospital, DK-4000 Roskilde, Denmark
15 Department of Internal Medicine, Genetics Laboratory, Erasmus University Medical Center Rotterdam, 3000 CA Rotterdam, The Netherlands
16 Department of Epidemiology, Erasmus University Medical Center Rotterdam, 3000 CA Rotterdam, The Netherlands
17 Institute of Epidemiology, Helmholtz Zentrum München - German Research Center for Environmental Health, D-85764 Neuherberg, Germany
18 Institute of Medical Informatics, Biometry and Epidemiology, Ludwig-Maximilians University, D-81377 Munich, Germany
19 Institut für Medizinische Molekularbiologie, Christian-Albrechts University Kiel, D-24105 Kiel, Germany
20 Cologne Center for Genomics and Institut für Genetik, University of Cologne, D-50674 Cologne, Germany
21 Genetics, GlaxoSmithKline, Research Triangle Park, North Carolina 27709
Summary
Understanding the genetic structure of the European population is important, not only from a historical perspective, but also for the appropriate design and interpretation of genetic epidemiological studies. Previous population genetic analyses with autosomal markers in Europe either had a wide geographic but narrow genomic coverage [1] and [2], or vice versa [3], [4], [5] and [6]. We therefore investigated Affymetrix GeneChip 500K genotype data from 2,514 individuals belonging to 23 different subpopulations, widely spread over Europe. Although we found only a low level of genetic differentiation between subpopulations, the existing differences were characterized by a strong continent-wide correlation between geographic and genetic distance. Furthermore, mean heterozygosity was larger, and mean linkage disequilibrium smaller, in southern as compared to northern Europe. Both parameters clearly showed a clinal distribution that provided
evidence for a spatial continuity of genetic diversity in Europe. Our comprehensive genetic data are thus compatible with expectations based upon European population history, including the hypotheses of a south-north expansion and/or a larger effective population size in southern than in northern Europe. By including the widely used CEPH from Utah (CEU) samples into our analysis, we could show that these individuals represent northern and western Europeans reasonably well, thereby confirming their assumed regional ancestry.
According to current theory, the autosomal gene pool of extant human populations in Europe lacks sharp discontinuities [1] and [2], with the exception of known isolates such as the Finns [6] and [7]. For classical genetic markers including, for example, erythrocyte antigens, changes in population genetic structure have been observed to follow a predominantly southeast-northwest gradient [1] and [2], thereby apparently matching the Pleistocene settlement of Europe, the Neolithic expansion from the Fertile Crescent, and (at least in part) the postglacial resettlement of Europe during the Mesolithic. Such gradient was also observed with particular haplogroups derived from the nonrecombining part of the Y chromosome (NRY), but other NRY data revealed additional population structure in Europe that has been associated with various demographic events in prehistoric, historic, and modern times [8], [9] and [10]. In contrast, the European mitochondrial DNA pool has been found to be rather homogeneous [11]. Here, we investigated the genetic structure of the European population by using 309,790 single-nucleotide polymorphisms (SNPs) in 2,457 individuals, ascertained at 23 sampling sites (henceforth referred to as “subpopulations”) in 20 different European countries. The data emerged from the genotyping of 2,514 European samples with the GeneChip Human Mapping 500K Array, followed by stringent quality control (see Table 1 and Experimental Procedures for details) and represent the largest Europe-wide genetic study to date.
Table 1. European Subpopulation Summary Statistics
Subpopulation Code Total No. Samples Final No. Samples* Sex Ratio
Norway (Førde) NO 52 52 1.74
Sweden (Uppsala) SE 50 46 all male
Finland (Helsinki) FI 47 47 0.74
Ireland IE 37 35 4.29
UK (London) UK 197 194 8.85
Denmark (Copenhagen) DK 60 59 1.22
Netherlands (Rotterdam) NL 292 280 all female
Germany I (Kiel) DE1 500 494 1.08
Germany II (Augsburg) DE2 500 489 1.02
Austria (Tyrol) AT 50 50 all male
Switzerland (Lausanne) CH 134 133 0.81
France (Lyon) FR 50 50 2.13
Portugal PT 16 16 0.78
Spain I ES1 83 81 1.02
Spain II (Barcelona) ES2 48 47 0.71
Italy I IT1 107 106 1.38
Italy II (Marches) IT2 50 49 all male
Former Yugoslavia YU 58 55 1.90
Northern Greece EL 51 51 1.43
Hungary HU 17 17 0.54
Romania RO 12 12 1.00
Poland (Warsaw) PO 50 49 all male
Czech Republic (Prague) CZ 53 45 0.96
Total 2,514 2,457
Total number of samples, final number of samples after data cleaning, and the sex ratio (male:female) of the final sample data set for each subpopulation. * is after stringent quality control.
First, we quantified the amount of information that each SNP could potentially provide about an individual's subpopulation affiliation by using the ancestry informativeness index In (Figure S1 available online) [12]. The maximum In value (0.09) was observed for rs6730157 in the RAB3GAP1 gene located about 68 kb away from the Lactase (LCT) gene. Furthermore, nine of the 20 (45%) most ancestry-informative SNPs, and 17 of the top 100 (Table S1), were from the LCT region and previously showed signatures of a selective sweep in CEU (Centre d'Etude du Polymorphism Humain from Utah) samples [13]. The average In across markers was 0.0064 (standard deviation: 0.0032), which represents only 0.93% of the maximum possible In of 0.69 in our study. (Note that this maximum would be attained if a SNP was fixed for one allele in 12 subpopulations and for the other allele in the remaining 11 subpopulations).
Second, we performed a principal-component analysis (PCA) in which the first two PCs were found to account for 31.6% and 17.3%, respectively, of the total variation, an amount similar to that reported in previous studies [1] and [5].
In our study, the first two PCs revealed a SNP-based grouping of European subpopulations that was strongly reminiscent of the geographic map of Europe (Figure 1; Figure S2). The first PC aligned subpopulations according to latitude, with the two Italian subpopulations at one end and the Finnish subpopulation at the other.
The second PC tended to separate subpopulations more according to longitude, with the Finnish subpopulation showing the largest values and the Irish and UK subpopulations showing the lowest values. The apparent geographic footing of the two PCs received additional support from an observed statistically significant positive correlation (Pearson r2 = 0.632, two-tailed p < 10−15) between the genetic distance (Euclidian distance between the median first two eigenvectors of the PCA) and the geographic (great-circle) distance between the analyzed subpopulations.
Figure 1.
SNP-Based PCA of 2,457 European Individuals from 23 Subpopulations
(A) Kernel density plot of the first two dimensions of a SNP-based PCA using those 309,790 SNPs from the GeneChip Human Mapping 500K Array Set (Affymetrix) that passed quality control.
(B) Geographic distribution of the 23 subpopulations; capitals were used as the respective landmark if location information was either unspecific or lacking (see Table 1 for further sample details).
Third, we searched for genetic barriers [14] in our dataset by using the same genetic and geographic distance matrices. This analysis identified two statistically significant barriers for the 23 subpopulations. One barrier was observed between the Finnish and all other subpopulations (first PC considering FI against the rest: r2 = 0.074, two-tailed p < 10−15; second PC considering FI against the rest: r2 = 0.33, two-tailed p < 10−15) and the other one between the two Italian and all other subpopulations (first PC considering IT1 and IT2 against the rest: r2 = 0.37, two-tailed p < 10−15; second PC considering IT1 and IT2 against the rest: r2 = 0.014, two-tailed p = 2.31 × 10−9).
Fourth, we studied the geographic distribution of genetic diversity by computing mean heterozygosity and mean linkage disequilibrium (LD) based upon HR2 [15] between markers at a distance < 10 kb for each subpopulation. Results from both analyses showed that the genetic diversity tended to be larger, and the LD smaller, in southern Europe as compared to northern Europe (Figure 2). Moreover, both analyses supported a genetic gradient of south-north orientation (r2 adjusted for the number of data points between the mean observed heterozygosity and latitude: 0.76, p = 3.80 × 10−8; adjusted r2 between HR2 and latitude: 0.71, two-tailed p = 4.33 × 10−7) but not of west-east orientation (adjusted r2 between heterozygosity and longitude: 0.03, two-tailed p = 0.416; adjusted r2 between HR2 and longitude: 0.099, two-tailed p = 0.078). Spatial autocorrelation analysis of both variables revealed statistically significant (p < 0.05) patterns compatible with a clinal distribution as indicated by the presence of positive and statistically significant autocorrelation values for small pair-wise distances and negative and statistically significant Moran's I values for large distances (see Figure 2). Bearing analysis [16] revealed for the heterozygosity measure the maximal angular correlations (r = 0.69) at 87° and the minimal (r = −0.153) at 165°, as well as for HR2 the maximal at 55° (r = 0.67) and the minimal (r = −0.167) at 160°, thus also suggesting a south-to-north spatial distribution of both variable. These results are compatible with larger effective population sizes in the south than in the north of Europe and/or a population expansion from southern toward northern Europe. Hierarchical analysis of molecular variance (AMOVA) [17] revealed that clustering the individuals according to four geographic groups—north (NO, SE, FI), north-west/central (IE, UK, DK, NL, DE1, DE2, AT, CH, FR), east (HU, RO, PO, CZ), and south (PT, ES1, ES2, IT1, IT2, YU, EL)—explained an average of 0.17% (95% coefficient interval: 0.0% to 0.91%) of the total genetic variance, whereas individual subpopulation affiliation explained 0.25% (95% coefficient interval: 0.0% to 1.25%).
Figure 2.
Geographic Distribution of Two Measures of Genetic Diversity across the European Population
(A and B) Isoline map (A) of Europe based on the mean observed heterozygosity in each of 23 European subpopulations with (B) corresponding spatial autocorrelogram.
(C and D) Isoline map (C) of Europe based on the mean observed linkage disequilibrium based on HR2 in each of 23 European subpopulations with (D) corresponding spatial autocorrelogram. Both spatial autocorrelograms showed statistically significant departures from randomness (p < 0.05). For each distance class, the number of subpopulation pairs included and the statistical significance (*, p < 0.05; **, p < 0.01; ***, p < 0.001) are provided.
Overall, our study showed that the autosomal gene pool in Europe is comparatively homogeneous but at the same time revealed that the small genetic differentiation that is present between subpopulations is characterized by a significant correlation between genetic and geographic distance. Furthermore, the qualitative nature of these results is in close agreement with expectations based on human migration history in Europe. The major prehistoric waves of human migration in Europe followed south and southeastern to north and northwestern directions [1], including the first Paleolithic settlement of the continent by anatomically modern humans [18], most of the postglacial resettlement during the Mesolithic [19], and the farming-related population expansion during the Neolithic [18] and [20]. Thus, both the level and the change in neutral autosomal variation in Europe can be expected to roughly follow southern-to-northern gradients as we observed, with the possible exception of population isolates as observed for the Finns. On the other hand, migration events in more recent (i.e., historic) times are presumed to have had a more homogenizing effect upon the previously established genetic landscape, as a result of their sporadic nature and haphazard geographic orientation [2]. This implies that genetic differences between extant European subpopulations can be expected to be small indeed. The genetic landscape described by the 300,000 autosomal SNPs analyzed here closely resembles that previously obtained with 128 alleles from 49 classical markers (see Table 1.3.1 in [1]). This similarity is highlighted by a significant correlation (r = 0.516; two-tailed Mantel test p = 0.0042, performed with 10,000 Monte Carlo permutations) between the pair-wise FST values [21] computed for the 19 European subpopulations that overlapped between the two datasets (Danish, Dutch, Yugoslavian, Hungarian, Irish, Italian, Portuguese, Spanish, Swiss, English, German, Austrian, Finnish, French, Greek, Norwegian, Polish, Swedish, and Czechoslovakian). This notwithstanding, a stronger correlation between FST and great-circle geographic distances was observed for the subpopulations when the SNPs from our study were used (r = 0.661; two-tailed Mantel test p = 0.00010, performed with 10,000 Monte Carlo permutations) as compared to the classical markers (r = 0.503, two-tailed Mantel test p = 0.00020, performed with 10,000 Monte Carlo permutations).
Previous studies based on genome-wide SNP diversity reported differences between individuals of southern and northern/central European ancestry [3], [5] and [6] and, to a lesser extent, between those of eastern and western European ancestry [3], which were not confirmed in our study. They mostly relied on the analysis of European Americans whose geographic assignment was determined from self-reported family records. Although genetic studies using European Americans can reveal important information about the genetic structure of the European ancestry of European Americans, caution must be exercised when drawing conclusions about the current genetic structure of Europe from European Americans because (1) European migrants may not have been representative of their country of origin, (2) the temporal difference introduced by sampling second- or third-generation descendants means that allele-frequency estimates inevitably ignored recent population movements (i.e., WWII-related migrations), and (3) self-reported geographic origin is error prone [22]. Our study avoided these potential pitfalls by using large samples of individuals of genuinely European origin, as evidenced by the documentation of their respective place of birth or residence being in one of the named subpopulations, and with comprehensive continent-wide coverage.
It is of general interest to place the CEU samples, widely used in genetic epidemiological and population genetic studies as representing the European population, into the context of our findings. The CEPH-CEU panel comprises U.S. Americans who were collected in Utah in 1980 and who are assumed to have descended from migrants originating from northern and western parts of Europe [23]. The samples were also included in the International HapMap Project and formed the basis of selecting tagging SNPs used in current genome-wide association studies with Illumina SNP arrays. Whereas a previous study [3] confirmed the grouping of the CEPH-CEU samples with other northern and western European subpopulations, our study was capable of providing their most precise positioning on the European genetic map (Figure 3). It turned out that, while the CEPH-CEU panel was indeed largely representative of northwestern and central Europeans, parts of Scandinavia as well as southern and eastern Europe were not well represented by these samples (Figure 3). Estimated inflated false-positive rates for all subpopulations were largest in the Finns, followed by the two Italian subpopulations (see Table S2). This implies that researchers conducting genetic-association studies in at least these regions, using the CEPH-CEU samples as controls, may be at increased risk of false-positive associations. Our confirmation of the regional European origin of the CEPH-CEU samples also indicates that inferring the geographic origin of an unknown person from autosomal DNA markers, which is highly relevant in the forensic context, might now be feasible down to the level of European subregions, at least when a large number of genetic markers and a reference database, such as are applied here, are used.
Figure 3.
Position of CEPH-CEU Samples in a SNP-Based PCA Kernel-Density Plot of 23 European Subpopulations
CEU individuals (U.S. Americans of European descent from Utah) are plotted as open circles. For details, see Figure 1 and Table 1.
Conclusions
Our comprehensive SNP genotype data from 23 European subpopulations, providing a dense coverage at both the geographic and genomic level and
representing the largest Europe-wide genetic study to date, allowed us to describe the genetic structure of the European population with the highest resolution. Although the amount of differentiation within the European autosomal gene pool was found to be small,
the existing genetic differences nevertheless correlated well with geographic distances. Furthermore, mean heterozygosity was larger, and mean linkage disequilibrium smaller, in southern than in northern European subpopulations, and both parameters exhibited a continuous clinal distribution across Europe.
Overall, our results were compatible with expectations based on European population history, mainly the prehistoric population expansion from southern to northern Europe and/or a larger effective population size in the south as compared to the north of Europe. Our dataset also allowed placement of the widely used CEPH-CEU samples onto the European genetic landscape, essentially confirming their genetic ancestry in northern and western Europe.
Experimental Procedures
Samples and Genotyping
The GeneChip Human Mapping 500K Array Set (Affymetrix) was used to genotype 500,568 SNPs in 2,514 individuals from 23 different sampling sites (henceforth termed “subpopulations”) located in one of 20 different European countries. Genotyping according to the instructions provided by the manufacturer was carried out at one of seven specialized centers: the Cologne Center for Genomics at the University of Cologne (Germany) for DE1, NO, SE, FI, AT, FR, ES2, IT2, EL, PO, and CZ; the Helmholtz Zentrum München - German Research Center for Environmental Health for DE2; the genetics laboratory of the Department of Internal Medicine, Erasmus MC (Netherlands) for NL; and the RH Microarray Centre Rigshospitalet, Copenhagen University Hospital (Denmark) for DK (see Table 1 for abbreviation explanations). Samples from the GlaxoSmithKline-sponsored POPRES project (IE, UK, CH, PT, ES1, IT1, YU, HU, and RO) were genotyped at Expression Analysis (Durham, NC, USA) and at Gene Logic (Gaithersburg, MD, USA) (see Table 1 for abbreviation explanations). Some samples belonged to existing control population studies, with detailed descriptions available elsewhere: KORA [24] for DE2, PopGen [25] for DE1, the Rotterdam Study [26], [27] and [28] for NL, and POPRES (drawn from the LOLIPOP and CoLaus studies) for IE, UK, CH, PT, ES1, IT1, YU, HU, and RO [29], [30] and [31]. Samples were drawn randomly from these pools or, in the case of POPRES, were ascertained on the basis of sample-size requirements. European migrants from non-European regions were not included in the initial analysis. For 11 of the subpopulations (NO, SE, FI, AT, FR, ES2, IT2, EL, PO, CZ, and DK), samples were obtained from healthy unrelated volunteers: Norwegian samples (NO) from blood donors of the Førde region, Swedish samples (SE) from the Uppsala region [32], Finnish samples (FI) from the Helsinki area with parents and grandparents originating from various regions in Finland, Austrian samples (AT) from the Tyrol region with parents originating from Tyrol, French samples (FR) from blood donors of Lyon with parents originating from the Rhône Alpes area, Spanish samples (ES2) from Catalonia of blood donors from rural areas who speak Catalan as their mother tongue and who had regional Catalan ancestry for at least two generations [33], Italian samples (IT2) from blood donors of the upland of the Marches region [34], Greek samples (EL) from the north of the country [35], Polish samples (PO) from the Warsaw region of central Poland [36], Czech samples (CZ) from the central Bohemian region in and around Prague, and Danish samples (DK) from the Danish Blood Donor Corps in the Copenhagen area. In addition, GeneChip Human Mapping 500K Array data from CEPH-CEU samples were retrieved from the Affymetrix website (
http://www.affymetrix.com).
Quality Assessment and Control Procedure
Array-based SNP genotypes were subjected to stringent quality control: First, each individual was required to have a genotype call rate ≥ 93%, with the dynamic model (DM) algorithm with a confidence score of 0.26, and a per-individual call rate ≥ 95% for all individuals genotyped by the same facility, with the Bayesian robust linear model with Mahalanobis distance classifier (BRLMM) algorithm with a confidence score of 0.5. The call rate was defined here as the proportion of unambiguous genotypes among either all SNPs (per-individual call rate) or all individuals (per-marker call rate), respectively. Markers that were monomorphic (1.4% of the total), that were located on the X chromosome (2.1%), or that had a per-marker call rate ≤ 90% in at least one genotyping facility (5.7%) were excluded, as were those showing a significant (p ≤ 0.05) deviation from Hardy-Weinberg equilibrium (HWE) in at least one subpopulation (31.3%). HWE was tested by means of a χ2 test, or by Fisher's exact test when the observed or expected number of a given genotype was less than 5. This method was preferred over others that have been shown to be more powerful [37] because the computational requirements of these methods increase exponentially with sample size and were thus too resource intensive for our study. The average proportion of heterozygous genotypes at X chromosomal markers was estimated per individual in order to detect false gender assignments. Male subjects can be expected to show X chromosomal heterozygosity proportions ≤ 1%, reflecting the overall genotyping error rate, and female subjects should show proportions near the average heterozygosity (26%) of the analyzed X chromosomal SNPs. Average identity-by-state (IBS) distances were calculated for a given set of markers as the average genetic dissimilarity between pairs of individuals. Analysis of IBS values within subpopulations allowed us to detect two types of outliers: (1) cognate relatives, i.e., individuals that were genetically more similar than expected to another member of the same subpopulation, and (2) “aliens,” i.e., individuals that were far less genetically similar than expected to the rest of the subpopulation. Formally, cognate relatives were defined as pairs of individuals having a pair-wise IBS value larger than the so-called “Tukey outlier criterion” when compared with the rest of pairs of individuals of the same subpopulation, i.e., the median IBS plus three times the interquartile range (IQR) in that subpopulation. In this case, the partner with the lower call rate was excluded. Aliens were defined as individuals with at least 60% of their pair-wise IBS values below the median minus three times the IQR. These two criteria led to the exclusion of 56 individuals from further analysis (Table 1). One individual identified as female had an average proportion of heterozygous X chromosomal markers of only 0.6% and was thus excluded from further analysis. In total, quality control left 2,457 individuals (97.6%) and 309,790 markers (62.4%) for inclusion in subsequent analysis. AMOVA [17] was performed to ascertain the magnitude of variation attributable to the respective genotyping center or subpopulation. The mean amount of genetic variance explained among genotyping centers was 0.095% (95% confidence interval: 0% to 0.71%), whereas subpopulation affiliation explained 0.63% of the variance (95% confidence interval: 0% to 2.86%). As expected, the largest amount of genetic variation was explained by differences between individuals (99.72%; 95% confidence interval: 98.61% to 100.00%). Data are available on request from the authors according to the regulations of the participating studies and sample cohorts.
Statistical Data Analyses
The ancestry-informativeness index In was estimated for each marker as described eslewhere [12]. Principal-component analysis was performed with the Eigensoft program with the default settings [38]. Population-wise kernel densities were computed from the first two PCs with the adehabitat R package [39] and subjected to least-squares crossvalidation [40] that used 80% of individuals per subpopulation for training. Pearson correlation coefficients were computed for the genetic distance between the subpopulations (represented by the respective median over all individuals in that subpopulation of the first two eigenvectors) and the great-circle geographic distance. The statistical significance of these correlation coefficients was assessed by means of a Mantel test [41]. Barrier analysis was performed on the basis of the Monmonier's algorithm [14]. Locus-wise AMOVA [17] was conducted after clustering the European subpopulations by genotyping center as well as by the use of four geographic groups. Negative percentages of explained variation were settled to 0. Both mean heterozygosity and mean linkage disequilibrium computed by means of HR2 [15] were computed with a subsample of ten individuals per population in order to adjust for possible influence of sample size [42]. Spatial autocorrelation and Bearing analyses were performed with the software PASSAGE 1.1 [43]. Isoline maps were performed with the Golden Surfer 8 software [44], with the inverse-distance method used for interpolation points. Isoline levels were defined to include the value of at least one of the 23 populations with intervals of 0.001 in the case of heterozygosity and 0.002 in the case of HR2. For evaluation of the extent to which the CEPH-CEU samples are representative of the subpopulations used in the present study, marker-wise tests of association (Fisher's exact test) were performed each time with the CEPH-CEU samples as “controls” and a given subpopulation as “cases.” The false-positive rate was defined as the percentage of markers yielding a p value < 0.05. If the CEPH-CEU samples were representative of a subpopulation, the false-positive rate would be around 0.05, whereas higher false-positive rates indicate that the CEPH-CEU samples may not be representative of the respective subpopulation.
Acknowledgments
All volunteers are gratefully acknowledged for sample donation. We thank the following colleagues for their help and support: J. Kooner and J. Chambers of the LOLIPOP study and D. Waterworth, V. Mooser, G. Waeber, and P. Vollenweider of the CoLaus study for providing access to their collections via the GlaxoSmithKline-sponsored Population Reference Sample (POPRES) project; K. King for preparing the POPRES data; M. Simoons, E. Sijbrands, A. van Belkum, J. Laven, J. Lindemans, E. Knipers, and B. Stricker for their financial contribution to the generation of the Rotterdam Study dataset; P. Arp, M. Jhamai, W. van IJken, and R. van Schaik for generating the Rotterdam Study dataset; T. Meitinger, P. Lichtner, G. Eckstein, and all other members of the Helmholtz Zentrum München genotyping staff for generating the KORA Study dataset; H. von Eller-Eberstein for management of the PopGen project; F.C. Nielsen, R. Borup, C. Schjerling, H. Ullum, E. Haastrup, and numerous colleagues at the Copenhagen University Hospital Blood Bank for assistance in making the Danish data available; and S. Brauer for DNA sample management. We would additionally like to thank Affymetrix for making the GeneChip Human Mapping 500K Array genotypes of the CEPH-CEU trios publicly available and the Centre d'Etude du Polymorphisme Humain (CEPH) for the original sample collection. We are grateful to three anonymous reviewers for their comments, which stimulated us to improve the manuscript. This work was supported by the Netherlands Forensic Institute to M.Ka.; Affymetrix to M.Ka. and M.Kr.; the German National Genome Research Network and the German Federal Ministry of Education and Research to H.-E.W., S.S., M.Kr., and P.N. (01GR0416 to P.N.); the Helmholtz Zentrum München - German Research Center for Environmental Health, Neuherberg, and the Munich Center of Health Sciences as part of LMUinnovativ to H.-E.W.; the Netherlands Organization for Scientific Research (NWO 175.010.2005.011) to A.G.U.; the European Commission to A.G.U. (GEFOS; 201865) and A.S. (LD Europe; QLG2-CT-2001-00916); the Czech Ministry of Health (VZFNM 00064203 and IGA NS/9488-3) to M.M.; Helse-Vest, Regional Health Authority Norway to L.A.B., the Swedish National Board of Forensic Medicine (RMVFoU 99:22, 02:20) to G.H.; and the Academy of Finland to A.S. (80578, OMLL) and J.P. (109265 and 111713). None of the funding organizations had any influence on the design, conduct, or conclusions of the study.
References
1 L.L. Cavalli-Sforza, P. Menozzi and A. Piazza, The History and Geography of Human Genes, Princeton University Press, Princeton, NJ (1994).
2 R.R. Sokal, R.M. Harding and N.L. Oden, Spatial patterns of human gene frequencies in Europe, Am. J. Phys. Anthropol. 80 (1989), pp. 267–294. View Record in Scopus | Cited By in Scopus (77)
3 M. Bauchet, B. McEvoy, L.N. Pearson, E.E. Quillen, T. Sarkisian, K. Hovhannesyan, R. Deka, D.G. Bradley and M.D. Shriver, Measuring European population stratification with microarray genotype data, Am. J. Hum. Genet. 80 (2007), pp. 948–956. Article | PDF (721 K) | View Record in Scopus | Cited By in Scopus (66)
4 A.L. Price, J. Butler, N. Patterson, C. Capelli, V.L. Pascali, F. Scarnicci, A. Ruiz-Linares, L. Groop, A.A. Saetta and P. Korkolopoulou et al., Discerning the ancestry of European Americans in genetic association studies, PLoS Genet 4 (2008), p. e236. View Record in Scopus | Cited By in Scopus (15)
5 C. Tian, R.M. Plenge, M. Ransom, A. Lee, P. Villoslada, C. Selmi, L. Klareskog, A.E. Pulver, L. Qi and P.K. Gregersen et al., Analysis and application of European genetic substructure using 300 K SNP information, PLoS Genet 4 (2008), p. e4. View Record in Scopus | Cited By in Scopus (29)
6 M.F. Seldin, R. Shigeta, P. Villoslada, C. Selmi, J. Tuomilehto, G. Silva, J.W. Belmont, L. Klareskog and P.K. Gregersen, European population substructure: Clustering of northern and southern populations, PLoS Genet 2 (2006), p. e143. View Record in Scopus | Cited By in Scopus (40)
7 A. Sajantila, A.H. Salem, P. Savolainen, K. Bauer, C. Gierig and S. Paabo, Paternal and maternal DNA lineages reveal a bottleneck in the founding of the Finnish population, Proc. Natl. Acad. Sci. USA 93 (1996), pp. 12035–12039. View Record in Scopus | Cited By in Scopus (105)
8 L. Roewer, P.J. Croucher, S. Willuweit, T.T. Lu, M. Kayser, R. Lessig, P. de Knijff, M.A. Jobling, C. Tyler-Smith and M. Krawczak, Signature of recent historical events in the European Y-chromosomal STR haplotype distribution, Hum. Genet. 116 (2005), pp. 279–291. View Record in Scopus | Cited By in Scopus (84)
9 Z.H. Rosser, T. Zerjal, M.E. Hurles, M. Adojaan, D. Alavantic, A. Amorim, W. Amos, M. Armenteros, E. Arroyo and G. Barbujani et al., Y-chromosomal diversity in Europe is clinal and influenced primarily by geography, rather than by language, Am. J. Hum. Genet. 67 (2000), pp. 1526–1543. Article | PDF (2344 K) | View Record in Scopus | Cited By in Scopus (301)
10 M. Kayser, O. Lao, K. Anslinger, C. Augustin, G. Bargel, J. Edelmann, S. Elias, M. Heinrich, J. Henke and L. Henke et al., Significant genetic differentiation between Poland and Germany follows present-day political borders, as revealed by Y-chromosome analysis, Hum. Genet. 117 (2005), pp. 428–443. View Record in Scopus | Cited By in Scopus (34)
11 L. Simoni, F. Calafell, D. Pettener, J. Bertranpetit and G.V. Barbujani, Geographic patterns of mtDNA diversity in Europe, Am. J. Hum. Genet. 66 (2000), pp. 262–278. Article | PDF (285 K) | View Record in Scopus | Cited By in Scopus (113)
12 N.A. Rosenberg, L.M. Li, R. Ward and J.K. Pritchard, Informativeness of genetic markers for inference of ancestry, Am. J. Hum. Genet. 73 (2003), pp. 1402–1422. Article | PDF (10083 K) | View Record in Scopus | Cited By in Scopus (180)
13 B.F. Voight, S. Kudaravalli, X. Wen and J.K. Pritchard, A map of recent positive selection in the human genome, PLoS Biol. 4 (2006), p. e72. View Record in Scopus | Cited By in Scopus (99)
14 F.C. Manni, E. Guérard and G.E. Heyer, Geographic patterns of (genetic, morphologic, linguistic) variation: How barriers can be detected by “Monmonier's algorithm.”, Hum. Biol. 76 (2004), pp. 173–190. View Record in Scopus | Cited By in Scopus (186)
15 C. Sabatti and N. Risch, Homozygosity and linkage disequilibrium, Genetics 160 (2002), pp. 1707–1719. View Record in Scopus | Cited By in Scopus (27)
16 A.B. Falsetti and R.R. Sokal, Genetic structure of human populations in the British Isles, Ann. Hum. Biol. 20 (1993), pp. 215–229. View Record in Scopus | Cited By in Scopus (19)
17 L. Excoffier, P.E. Smouse and J.M.V. Quattro, Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data, Genetics 131 (1992), pp. 479–491. View Record in Scopus | Cited By in Scopus (5300)
18 E.M. Belle, P.A. Landry and G. Barbujani, Origins and evolution of the Europeans' genome: Evidence from multiple microsatellite loci, Proc Biol Sci 273 (2006), pp. 1595–1602. View Record in Scopus | Cited By in Scopus (21)
19 A. Torroni, H.J. Bandelt, V. Macaulay, M. Richards, F. Cruciani, C. Rengo, V. Martinez-Cabrera, R. Villems, T. Kivisild and E. Metspalu et al., A signal, from human mtDNA, of postglacial recolonization in Europe, Am. J. Hum. Genet. 69 (2001), pp. 844–852. Article | PDF (1531 K) | View Record in Scopus | Cited By in Scopus (128 )
20 L. Chikhi, R.A. Nichols, G. Barbujani and M.A.V. Beaumont, Y genetic data support the Neolithic demic diffusion model, Proc. Natl. Acad. Sci. USA 99 (2002), pp. 11008–11013. View Record in Scopus | Cited By in Scopus (90)
21 B.S. Weir and C.C. Cockerham, Estimating F-statistics for the analysis of population structure, Evolution Int. J. Org. Evolution 38 (1984), pp. 1358–1370. View Record in Scopus | Cited By in Scopus (5944)
22 M.S. Burnett, K.J. Strain, T.G. Lesnick, M. de Andrade, W.A. Rocca and D.M. Maraganore, Reliability of self-reported ancestry among siblings: Implications for genetic association studies, Am. J. Epidemiol. 163 (2006), pp. 486–492. View Record in Scopus | Cited By in Scopus (8 )
23 J. Dausset, H. Cann, D. Cohen, M. Lathrop, J.M. Lalouel and R. White, Centre d'etude du polymorphisme humain (CEPH): Collaborative genetic mapping of the human genome, Genomics 6 (1990), pp. 575–577. Abstract | PDF (405 K) | View Record in Scopus | Cited By in Scopus (252)
24 H. Lowel, A. Doring, A. Schneider, M. Heier, B. Thorand, C. Meisinger and M.K.S. Group, The MONICA Augsburg surveys–basis for prospective cohort studies, Gesundheitswesen 67 (Suppl 1) (2005), pp. S13–S18. View Record in Scopus | Cited By in Scopus (44)
25 M. Krawczak, S. Nikolaus, H. von Eberstein, P.J. Croucher, N.E. El Mokhtari and S. Schreiber, PopGen: Population-based recruitment of patients and controls for the analysis of complex genotype-phenotype relationships, Community Genet. 9 (2006), pp. 55–61. View Record in Scopus | Cited By in Scopus (69)
26 A. Hofman, M.M. Breteler, C.M. van Duijn, G.P. Krestin, H.A. Pols, B.H. Stricker, H. Tiemeier, A.G. Uitterlinden, J.R. Vingerling and J.C. Witteman, The Rotterdam Study: Objectives and design update, Eur. J. Epidemiol. 22 (2007), pp. 819–829. View Record in Scopus | Cited By in Scopus (176)
27 A. Hofman, D.E. Grobbee, P.T. de Jong and F.A. van den Ouweland, Determinants of disease and disability in the elderly: The Rotterdam Elderly Study, Eur. J. Epidemiol. 7 (1991), pp. 403–422. View Record in Scopus | Cited By in Scopus (752)
28 M. Kayser, F. Liu, A.C. Janssens, F. Rivadeneira, O. Lao, K. van Duijn, M. Vermeulen, P. Arp, M.M. Jhamai and W.F. van Ijcken et al., Three genome-wide association studies and a linkage analysis identify HERC2 as a human iris color gene, Am. J. Hum. Genet. 82 (2008), pp. 411–423. Article | PDF (1469 K) | View Record in Scopus | Cited By in Scopus (54)
29 J.S. Kooner, J.C. Chambers, C.A. Aguilar-Salinas, D.A. Hinds, C.L. Hyde, G.R. Warnes, F.J. Gomez Perez, K.A. Frazer, P. Elliott and J. Scott et al., Genome-wide scan identifies variation in MLXIPL associated with plasma triglycerides, Nat. Genet. 40 (2008), pp. 149–151. View Record in Scopus | Cited By in Scopus (91)
30 M.R. Nelson, S.A. Bacanu, M. Mosteller, L. Li, C.E. Bowman, A.D. Roses, E.H. Lai and M.G. Ehm, Genome-wide approaches to identify pharmacogenetic contributions to adverse drug reactions, Pharmacogenomics J. (2008) 10.1038/tpj.2008.4 in press. Published online February 26, 2008.
31 M.S. Sandhu, D.M. Waterworth, S.L. Debenham, E. Wheeler, K. Papadakis, J.H. Zhao, K. Song, X. Yuan, T. Johnson and S. Ashford et al., LDL-cholesterol concentrations: A genome-wide association study, Lancet 371 (2008), pp. 483–491. Abstract | Article | PDF (287 K) | View Record in Scopus | Cited By in Scopus (104)
32 A.O. Karlsson, T. Wallerstrom, A. Gotherstrom and G. Holmlund, Y-chromosome diversity in Sweden - a long-time perspective, Eur. J. Hum. Genet. 14 (2006), pp. 963–970. View Record in Scopus | Cited By in Scopus (23)
33 S. Plaza, F. Calafell, A. Helal, N. Bouzerna, G. Lefranc, J. Bertranpetit and D. Comas, Joining the pillars of Hercules: mtDNA sequences show multidirectional gene flow in the western Mediterranean, Ann. Hum. Genet. 67 (2003), pp. 312–328. View Record in Scopus | Cited By in Scopus (56)
34 V. Onofri, F. Alessandrini, C. Turchi, B. Fraternale, L. Buscemi, M. Pesaresi and A. Tagliabracci, Y-chromosome genetic structure in sub-Apennine populations of Central Italy by SNP and STR analysis, Int. J. Legal Med. 121 (2007), pp. 234–237. View Record in Scopus | Cited By in Scopus (15)
35 H. Kondopoulou, R. Loftus, A. Kouvatsi and C. Triantaphyllidis, Genetic studies in 5 Greek population samples using 12 highly polymorphic DNA loci, Hum. Biol. 71 (1999), pp. 27–42. View Record in Scopus | Cited By in Scopus (14)
36 R. Ploski, M. Wozniak, R. Pawlowski, D.M. Monies, W. Branicki, T. Kupiec, A. Kloosterman, T. Dobosz, E. Bosch and M. Nowak et al., Homogeneity and distinctiveness of Polish paternal lineages revealed by Y chromosome microsatellite haplotype analysis, Hum. Genet. 110 (2002), pp. 592–600. View Record in Scopus | Cited By in Scopus (42)
37 D.J. Schaid, A.J. Batzler, G.D. Jenkins and M.A. Hildebrandt, Exact tests of Hardy-Weinberg equilibrium and homogeneity of disequilibrium across strata, Am. J. Hum. Genet. 79 (2006), pp. 1071–1080. Article | PDF (263 K) | View Record in Scopus | Cited By in Scopus (14)
38 N. Patterson, A.L. Price and D. Reich, Population structure and eigenanalysis, PLoS Genet 2 (2006), p. e190.
39 C. Calenge, The package “adehabitat” for the R software: A tool for the analysis of space and habitat use by animals, Ecol. Modell. 197 (2006), pp. 516–519. Abstract | Article | PDF (244 K) | View Record in Scopus | Cited By in Scopus (84)
40 B.W. Silverman, Density Estimation for Statistics and Data Analysis, Chapman & Hall / CRC Press, Boca Raton, Florida (1986).
41 N. Mantel, The detection of disease clustering and a generalized regression approach, Cancer Res. 27 (1967), pp. 209–220. View Record in Scopus | Cited By in Scopus (3849)
42 M. Jakobsson, S.W. Scholz, P. Scheet, J.R. Gibbs, J.M. VanLiere, H.C. Fung, Z.A. Szpiech, J.H. Degnan, K. Wang and R. Guerreiro et al., Genotype, haplotype and copy-number variation in worldwide human populations, Nature 451 (2008), pp. 998–1003. View Record in Scopus | Cited By in Scopus (237)
43 Rosenberg, M.S. (2001). PASSAGE: Pattern Analysis, Spatial Statistics, and Geographic Exegesis. 1.1 Edition, A.S.U. Department of Biology, ed. (Tempe, AZ.).
44 Golden Software. (2007). Surfer Version 8.08.3267. Colorado, USA.
Supplemental Data
The MOST recent and comprehensive study offers no surprises and confirms the known history of
the English and their ancestors/closest kin; the Dutch, the Danes, the Norwegians and the Germans
Go on, dog - find me something
more recent, of equal scientific worth, which contradicts the above
