This is a very important study, the first one to use both a large number of markers and a wide and fairly representative sample of Europeans from across the continent.My main observations after reading this study are:
- relative genetic homogeneity in Europe, with a fairly small percentage of variance explained by geographic differentiation
- clinal, rather than racial apportionment of European genetic variation, with no emerging separated clusters (except the Finns, who stand at some distance along the first eigenvector)
- south-north (but not east-west) decrease in genetic variation and heterozygosity indicating that Europe was populated on a south-north axis, rather than an east-west one.
- clear clustering of individuals from different ethnic groups within the European continuum, indicating that ethnic groups are not only cultural, but to some extent biological entities.
- Some ethnic groups are clearly distinguishable from each other (e.g. Swedes vs. Spaniards); some groups are partitioned into fairly disjoint sets (Spain I vs. Catalans in Spain II); others mutually overlap (e.g., British and Irish); while others overlap asymetrically (e.g., some former Yugoslavs in the Greek cluster, but not vice versa).
Interestingly Italians neighbor Spaniards on the other side; whereas former Yugoslavs neighbor Czechs.
A straightforward explanation for this pattern is that the Italian groups has mixed Western and Eastern Mediterranean affiliations; the latter stemming from either Neolithic farmers or Greek (or Etruscan, etc.) colonists.
Former Yugoslavs are mostly disjoint from Greeks, except some who seem to be Slavicized Greeks, consistent with their descent from indigenous Balkan populations on one hand and Slavic immigrants more akin to Czechs and Poles on the other. Thus, they occupy an intermediate position between Greeks and Czechs.
From the paper:
Hierarchical analysis of molecular variance (AMOVA) [17]They also refer to a few recent studies [3: Bauchet et al., 5: Tian et al., 6: Seldin et al.]:
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%).
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,
20]. Thus, both the level and the change in neutral autosomal
variation in Europe can be expected to roughly follow southernto-
northern gradients as we observed, with the possible exception
of population isolates as observed for the Finns.
Previous studies based on genome-wide SNP diversityUPDATE (Aug. 13) The Spittoon blogs about this:
reported differences between individuals of southern and
northern/central European ancestry [3, 5, 6] and, to a lesser extent,
between those of eastern and western European ancestry
[3], which were not confirmed in our study.
They confirm the findings of several recent but smaller European studies (Seldin et al, PLoS Genetics (2006); Bauchet et al, AJHG (2007); Tian et al, PLoS Genetics (2008); Price et al, PLoS Genetics (2008); Paschou et al, PLoS Genetics (2008))This is contradicted by what the authors actually say about three of these studies (see last quote above). The Spittoon also writes:
In the case of this current paper the Finns are the only nationality completely distinct from the rest of the European samples. The Finns speak a different kind of language from much of the rest of Europe, and are the only Scandinavian population represented.Actually, Finland is not usually thought of as a Scandinavian country; even if it is, it is certainly not the only Scandinavian population represented, since all three Scandinavian nations (Norway, Denmark, Sweden) were sampled.
The position of the Finns is likely due to the fact that they share quite recent ancestry with Asians, as evidenced by their possession of Y-chromosome haplogroup N-Tat.
Update (Aug. 14)
It is interesting how these results parallel those of Li et al. (2008), according to which Russians from Vologda have a membership of 86% in the main European cluster, Tuscans have 95%, and all others (except the Adygei from the Caucasus) between 99-100%. Given that Russians have a Finno-Ugrian substratum, these results parallel those of this study in which the two major genetic differences within Europe are primarily because of the Finns, and secondarily because of the Italians.
Update (Aug. 19): A previous study had discovered a substantial overlap between Greek and Italian Americans. Is the disjointness between Greeks and Italians discovered in the newer study the result of a larger number of markers, or due to the fact that northern Greeks and Central Italians seem to have been sampled? As is well known, Greek colonization of Italy originated mostly in the Peloponnese, and occurred in southern Italy and Sicily, and immigrants to America were not drawn uniformly from the territories of Italy and Greece.
Current Biology doi:10.1016/j.cub.2008.07.049
Correlation between Genetic and Geographic Structure in Europe
Oscar Lao et al.
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.
Link
137 comments:
Hey Dienekes, I don't have a copy yet. Where do the German samples come from? Also, I assume the Poles are from Warsaw yeah?
The Finns do appear to be outliers, and they are relatively speaking. But only because the Baltic states and Russians weren't sampled. If they were, they'd create a bridge from Central Europe to the Finns, almost like the Brits lnked up the Irish to the main cluster in the west.
Actually, my Irish example wasn't a great one, but meh...
I've had a look at the paper and it is very interesting. The lack of overlap between the English and Danish and the very small overlap between the English and North Germans is notable. Also the very large overlap between the Dutch and both the English and Irish is quite remarkable. I would expect thet a sample from northern France (Lyons is quite southern) would show considerable overlap with the Dutch/English/Irish.
re: urselius,
I haven't read the paper yet, but based on what Dienekes has reproduced:
(1) I'd be curious how you figured out there is a "lack of overlap between the English and Danish". Dutch individuals show up above and below the main concentration of UK individuals on the plot, but the plot fails to clearly show very many Dutch in the midst of the main UK concentration. Danish individuals, falling farther to the right in a busier section of the plot, will be expected to be less visible (likewise for DE1).
In fact, I see a red cross at about 0 on eigenvector 1 and -0.01 on eigenvector 2, which reinforces my belief that there is likely overlap between Danish and English individuals.
(2) You call the "UK" individuals "English". What does the paper actually say about the "UK" sample? (Or the "Irish" sample for that matter--I would not expect Catholic Irish, Anglo-Irish, and Northern Irish Protestant samples to be identical.)
Even if all the UK individuals were sampled in England (which I doubt), my guess is they include many with recent origins elsewhere in the UK.
(3) You seem to feel some need to separate England from the Germanic North and connect it with France. While we're speculating about where a sample from northern France would fall on this plot, why not think about sampling at finer scales in other countries as well. E.g., where would a sample restricted to East Anglia fall? Or samples from northern Jutland, southern Jutland, northwestern Germany, etc.
Finally, I notice that the green triangles seem to be densest immediately adjacent to or overlapping with continental Germanic speakers, including the Dutch, Norwegians, and Germans, and seem to be less dense as they spread out with the Irish samples along eigenvector 2 (which, by the way, I presume accounts for a much smaller fraction of variation than eigenvector 1).
The picture appears entirely consistent with a large-scale Germanic invasion and varying degrees of admixture with the previous inhabitants of the British isles--exactly what we would expect based on Y chromosomes.
I miss bayesian K-means clustering. Maybe it's in the paper?
Also some comments re. too limited samples, specially for large diverse states like France may be meaningful.
Also I noticed that the scale of the graph is not identical for eigenvectors 1 and 2. The graph should be either more compressed in the vertical direction or more expanded in the horizontal one (like 15-20% I estimate).
Former Yugoslavs are mostly disjoint from Greeks, except some who seem to be Slavicized Greeks, consistent with their descent from indigenous Balkan populations on one hand and Slavic immigrants more akin to Czechs and Poles on the other. Thus, they occupy an intermediate position between Greeks and Czechs.
Hungarians are also in the mix, in fact they overlap partly with Yugoslavians. I don't think it can be interpreted only in simplistic terms of "Greeks" and "Slavs" as people has been travelling through the Balcans (or anywhere else in Europe) since dates much older than the formation of these ethnic groups.
Based on another plot in the paper, yes, there is a very slight overlap between the UK (London) and Denmark (Copenhagen). But the English do overlap strongly with Norwegians and the Dutch. The English also are touching upon the North German (Kiel), and overlapping slightly with the South Germans (Augsburg). But then again, the Irish intermingle with the Germanics as well. So who knows if this is Germanic influence, or shared Northwest European ancestry.
Btw, interesting to note North Germans ovelapping with Poles. And as we know, Kiel was founded by Slavs.
Btw, interesting to note North Germans ovelapping with Poles. And as we know, Kiel was founded by Slavs.
IMO, knowing how ethno-cultural and political "borders" have fluctuated and in Central Europe, specially in the open plains of Germany and Poland, through the milennia, what would be really interesting would be to find no overlap: clearly distinct groups. Moreso considering that all sampled groups (except Finns) overlap with their neighbours and sometimes with peoples that are more distant.
In reply to n/a, the paper has a rather different version of the plot shown here. It has envelopes which overlap or don't overlap, single outlyers are subsumed.
The Irish and English (the paper calls these 'UK' and 'English' - go figure) overlap very greatly, both have significant overlaps with the Dutch and somewhat less with the Norwegians. The overlaps of both the insular envelopes with the Danes and North and South Germans are very very much smaller. Interestingly the Dutch have a smaller overlap with the North Germans and Danes than they do with the Norwegians.
Rather than showing evidence of incomers to the British Isles in historical times from North Germany and Denmark the results appear to show a history of interrelatedness based of geographical contigruity. The Dutch and English and Irish are related because they are close geographically.
I've just done a rough estimation of the overlap area ratios. The UK overlaps both the Danish and North German envelopes by about one third of that of the UK to Norwegian and about one twelfth of the UK to Dutch overlap. The Dutch to Danish and Dutch to North German overlaps are about half of the Dutch to Norwegian.
Can someone send me the PDF?
Does the article say where the Norwegian sample came from? It says "Haukeland University Hospital"; would this mean that only people from Bergen were sampled? Otherwise it would be surprising to see that Norwegian appear to be closer to Danes than to Swedes. Ydna seems to conflict somewhat with this.
The Norwegian sampel comes from Forden on the western coast.
I was just looking at some anthropological maps, and that region is rather broad headed and red haired. Kinda "Irish" if you will.
Well, I might be getting ahead of myself, but the Norway, Denmark, Holland, UK/Ireland grouping there does look like something North Atlantic, and Keltic...or rather pre-Keltic.
I am quickly glancing over this new one. This chart is structured oddly. UK and English? Dont get it. As for the Italian samples, they are together, but not sure why differentiating them. Are they Sardinians or Tuscan's? Tuscan's are the only ones to come up partly "Middle Eastern" on previous graph, but this chart doesnt show any such thing. The Yugoslavs are quite distinct from the Italians, however Roman occupation shouldnt be ruled out with some things shown or as Maju said, earlier movements.
BTW, what code is being used for Greece, I dont see GR.
Lol, Crimson's here...
Well, the UK sample comes from London (ethnic English). Both of the Itaian samples are from Central Italy, from from different localities. And Middle Eastern samples aren't included, so it's not something we can comment on.
Moreover, the Yugoslavs aren't all that distinct from Italians. But their distance is in line with geography, which is what this study is all about. It looks like ancient populations movements had a much greater imact than more recent historic ones.
And the code being used for Greece is El..Ellas.
The Norwegian sampel comes from Forden on the western coast.
I was just looking at some anthropological maps, and that region is rather broad headed and red haired. Kinda "Irish" if you will.
Well, I might be getting ahead of myself, but the Norway, Denmark, Holland, UK/Ireland grouping there does look like something North Atlantic, and Keltic...or rather pre-Keltic.
I suppose this must mean Forde (foerde/Førde) in Sogn og Fjorande. This was one of 3 county (fylke) that "Geographical heterogeneity of Y-chromosomal lineages in Norway" found significant from all other counties in spite of the low sample size.
"..The counties Sogn og Fjordane in the west, Aust-Agder in the south and Finnmark in the north were significantly different from 19, 12 and 9 other counties respectively..."
"...In this study population, substructuring is demonstrated in Aust-Agder (south), Sogn og Fjordane (west) and Finnmark (north). This is consistent with other genetic studies reporting regional differences. The frequency of K+ in the Kell blood group is 4.61 % in Aust-Agder and 10.36 % in Sogn og Fjordane. Both counties differ from 8 % which is the Norwegian average....."
Interesting that Netherlands and German sample is so far south and amongst the Italians, maybe its Roman or from Spanish occupied Netherlands. The Italian samples are taken just from Italy, the Italian-2 looks like from Tuscany region is all.. well its terribly not important. Possibly, some of what we're seeing is more Roman era derived but theres been some recent Italian colonies in the balkans and some recent migrations to Italy even. Can see a Romanian sample close to an Italian one there, which could support Trajan's victory in Thrace/Dacia. And Italian sample close the Yugoslav and Greek one, could be Neolithic. But this is alotve speculations. Still, the Balkans/Greeks and Italians are still pretty neatly separated on this one, the very negligible overlap or drift.
Paul,
Yes, I can easily see Norwegians from Oslo and the East custering with Central Sweden.
And I can also easily see southern Sweden clustering with Denmark, and northern Sweden close to Finland.
I wonder where my home town Poznan would end up? Probably just left of this Warsaw sample...I guess.
Polak wrote: Well, I might be getting ahead of myself, but the Norway, Denmark, Holland, UK/Ireland grouping there does look like something North Atlantic, and Keltic...or rather pre-Keltic.
The areas of the North Sea area have "always" been in contact, either by land or sea. The (epi-)Paleolithic substrate is roughly the same and they had contacts in Neolithic as well (Megalithism, Bell Beaker). Later we have all those IE migrations (Celts, Germanics).
Still, remember that the graph needs a 20% expansion of the horizontal axis to be proportionate and that, in this study at least, Brits+Irish and Northern Europeans do not almost overlap. They do form a continuum anyhow.
...
Crimson Guard wrote: Interesting that Netherlands and German sample is so far south and amongst the Italians, maybe its Roman or from Spanish occupied Netherlands.
Certainly in Roman and later times, Germany and Italy were interacting demographically. First the Italians (and others) garrisoned the Roman, then the Germans invaded, later Italy and germany were part of the same macro-monarchy, what often moved armies across the Alps in one direction or the other. The "Spanish" presence in Netherlands was tied to Italy and Austria: it was all for a time part of the last epysode of the medieval dream of "the Empire".
...
I understand that ES2 is from Catalonia. Where is ES1 from?
Maju,
Most of the variance (31.6%) is on PC1. So why the need to make the plot more "proportional"? PC2 has 17.3%.
And yeah, ES2 is from Barca, while ES1 from Central Spain (that's what the map shows, but probably just means ethnic Spanish).
Most of the variance (31.6%) is on PC1. So why the need to make the plot more "proportional"? PC2 has 17.3%.
I could not know that (I have not acess to the full article). I was thinking in terms of mere proprotion of the scales (i.e. 0.5 should be equal to 0.5 in principle).
And yeah, ES2 is from Barca, while ES1 from Central Spain (that's what the map shows, but probably just means ethnic Spanish).
Barna (colloquially). Barça is the football club (and only that).
Barcelona could well not be representative of Catalans anyhow. It is a major industrial metropolis that has many many immigrants from Southern Spain. If they have made no ethnic pre-selection, they should be a very mixed bunch.
Guess the other sample site is Madrid.
Btw, Polak: what do you think of the apparent E-W axis in the northern European pool? Baucher 2007, in contrast, suggests a marked E-W Mediterranean axis instead, with little differentiation in the north instead.
Is it mere sample bias?
urselius:
Rather than showing evidence of incomers to the British Isles in historical times from North Germany and Denmark the results appear to show a history of interrelatedness based of geographical contigruity. The Dutch and English and Irish are related because they are close geographically.
Geography matters because proximity facilitates the movement of people and genes.
The "North German" and "Danish" samples come from Baltic Sea ports--i.e., about as far from the North Sea as you can get in north-central Germany or Denmark. Most likely, samples from anywhere in Jutland or northwestern Germany will show more overlap with England and the Netherlands.
Also note that on eigenvector 2, the distance from the left-most Irish sample to the right-most UK samples is about the same as the distance from the right-most UK sample to the right-most Polish sample. This suggests, to me at least, the existence of a one-time population isolate in the British Isles which has undergone admixture from a continental source. Only the Finns--a distinctive population with Swedish admixture on the opposite side of northern Europe--show a larger spread on ev2.
Bottom line is that the data in this study are far from ideal for quantifying the effects of the Anglo-Saxon invasion of England, but they are entirely consistent with significant if not complete replacement in parts of the UK.
Maju,
Apparently they made sure to sample ethnic Catalans. They say so in the study, but obviously it's never possible to get all 100%pure ethnic samples.
And yeah, the Bauchet study did show some of that west-east cline. Its just easier to pick up with so many more markers.
I think we'd see more overlapping if other populations were sampled, like west and north Poland, east Germany, north Sweden, the Baltic states, etc. But that cline would certainly still be there.
If many groups overlap it could be said then that the nations they live in are places with a certain culture that they like to live at,not that they are geneticelly different.If the English.Irish,and Dutch are the same genetically,then the difference must be that their nations choose to express their life stlyes differently,like have different legends,and stories and maybe they like different food,but not that they are all that different genetically.
Can anyone who has access to the paper say where the Irish samples were obtained, as there should be regional substructure?
Eastern Ireland has had substantial immigration from England and Wales over the last few centuries, and before that was one of the most invasion prone area of the country, due to the best land.
South Eastern Ireland was the epicenter of the Norman (together with Flemish) settlement, and before that Belgae.
South Western Ireland, has people who are much darker in hair and eye color than the rest of the country.
North Midlands has an area of 4 counties which have a huge amount of Viking heritage.
Doesn't say where the Irish are from, as far as I can tell. Just says Ireland.
But the location map shows a dot in eastern Ireland.
The light blue squares are listed as IE. Does IE represent the Irish or some other group?
IE=Irish
These Greeks clearly show some Central European admixture.
They are clearly differentiated from Central Europeans on both principal components. If this chart can be interpreted as evidence of Greeks having Central European admixture, then everyone on it has such admixture, except for the Finns and a few Italians at the bottom.
If they were just Southern Europeans, they would not be situated north of central Italians.
Central Italians aren't some yardstick of Southern European ancestry.
Dienekes,
The Central Europeans here range from Germans and Czechs to Southern Yugoslavs.
And yes, the only groups with very little or no Central Euro admixture are the Italians, Finns and possibly some of the Brits. Half these Greeks must be full of it, if they're so far north.
Like I said, pity we don't have a structure diagram, and no samples from Armenia, Southern Italy, and Southern Greece.
But don't worry, I'm sure more samples and studies are coming. So we won't have to speculate, and you'll realise I'm right.
Another interesting feature is the Romanians appear to be very variable with at least 3 sub-populations, though the sample number for Romanians was very small.
Also the North Germans and Bavarians appear to be almost separate populations.
The Irish and English not only overlap but they do so throughout, not only in the part of their envelopes mutually overlapping the Dutch but also at the opposite end of their range. Had the Irish/English convergence been the result of shared continental influence: Anglo-Saxons, Danes Normans, Flemmings in historical times, then their overlap would be expected to only cover the area also overlapping the Dutch.
?Had the Irish/English convergence been the result of shared continental influence: Anglo-Saxons, Danes Normans, Flemmings in historical times, then their overlap would be expected to only cover the area also overlapping the Dutch.
No, it wouldn't.
And stop calling the UK sample "English" unless you can establish it is in fact limited to ethnic English.
According to the paper, the London sample is a "population reference sample" from a medical study, which suggests an indiscriminate sampling of (white) Londoners. Which means the sample most likely includes many descendants of Irish, Welsh, and Scottish immigrants.
"Irish ancestry
The highest concentration is in London where more than three-quarters (77%) claim to have Irish roots"
http://news.bbc.co.uk/2/hi/uk_news/1224611.stm
" Howes (2004) cites a recent Greater London
Authority (GLA) survey which revealed that 11% Londoners have Irish parents and
19% have Irish grandparents. "
http://www.charityblog.org.uk/wordpress/wp-content/uploads/2006/03/BIASreportjan06.pdf
The paper uses 'UK' and 'English' without discrimination. Unsurprisingly the majority of Londoners (excluding non-European immigrant populations, which the paper does) are of English origins. By no means are all Londoners called Murphy. Your argument sounds like special pleading - if the London sample group (which is one of the larger ones - and so should be more representative) is so contaminated then so should many of the others and the whole paper become utterly worthless. By the way the Scots and Welsh are British and not Irish. There are many Finns with Swedish surnames and ancestry but this seems to have had little effect on the Finnish results.
And "yes it would," try a tad of logical analysis.
"Irish ancestry
The highest concentration is in London where more than three-quarters (77%) claim to have Irish roots"
Bejabers! there aren't any English left - so why talk about them at all?
The survey wasn't taken in pubs on St Patrick's Day perchance?
Dienekes,
The markers I'm referring to are markers that evolved in Central Europe (something like the green in the Bauchet study).
These Greeks obviously carry them, and this means they have Central European admixture.
The markers I'm referring to are markers that evolved in Central Europe (something like the green in the Bauchet study).
Did you see these markers in your dream, because there are certainly no markers identified in this study as having evolved in Central Europe.
And yes, the only groups with very little or no Central Euro admixture are the Italians, Finns and possibly some of the Brits.
There is no way you know admixture from that graph, it just shows affinity by means of two statistically relevant variables, not clusters.
It is very possible that, as appears in other studies, the Greek/Sudbalcanic/Aegean genetic pool is largely a principal component of European one. Central Europeans should be the most likely to be "admixed", if anyone, as they are in the middle of all, suffering immigration waves at Neolithic from the Balcans and later also in Chalcolithic from Eastern Europe - not to mention Scythians, Romans, Germans, Slavs, etc.
Ok, they also spreaded their genes around... after all they were always in the middle of everything. The graph anyhow shows a continuum, a cline, not any marked clusters (excepting Finns).
The markers I'm referring to are markers that evolved in Central Europe (something like the green in the Bauchet study).
Bauchet 2007 yielded a quite different bidimensional PC graph, with greater southern E-W differences and smaller northern ones. And Greeks showed little influence of the CE "green" component in any case, clustering basically with West Asians (Armenians, Turks, Askhenazim) and some Italians (southerners?).
Maju you always overthink eveyrthing.
These are Northern Greeks, not just Greeks.
I bet Southern Greeks would cluster further down here, as far as the Italians at least.
Btw, I know it's hard to infer admixture rates from this plot. That's why I said a structure diagram is necessary.
But anyway, what this study shows is that the elements of European substructure were laid out in ancien times - from the Paleolithic to the Neolithic. After that we mainly mixed with our kin and near neighbours, homogenizing the original landscape and blurring any real clusters.
Those Italians show a strong Neolihic component IMO, and it's much weaker in the Northern Greek sample. I would explain that by backflow from Central Europe (the green in Bauchet). Surely you'd think that out of all the Europeans, the Greeks would have the most Neolithic admixture, and this is precisely why in every genetic sudy to date we've seen a NW or N to SE cline...the Greeks being at that SE tip.
The word "English" appears in the paper once, possibly through carelessness on the part of the authors (versus 6 times for "UK").
The true proportion of Londoners with Irish ancestry is most likely bracketed by the 77% reportedly claiming Irish ancestry and the 19% with Irish grandparents.
The issue at hand is what, if anything, this study says about the Anglo-Saxon invasion. If one wants to make claims about the relatedness of the English and Irish, one needs to sample Englishmen--not some undefined mix of English, Irish, Scottish, and Welsh.
"By the way the Scots and Welsh are British and not Irish."
Dude, what the fuck are you going on about? Who said they were Irish?
The Scots and Welsh are British, but they are not English. Genetically, they are most likely intermediate between the English and Irish.
But anyway, what this study shows is that the elements of European substructure were laid out in ancien times - from the Paleolithic to the Neolithic.
There is nothing about the timing of the formation of these genetic differences in the paper, certainly nothing about them being of Paleolithic antiquity. Indeed, the strongest difference in the LCT gene is certainly of Neolithic age.
Btw, I know it's hard to infer admixture rates from this plot. That's why I said a structure diagram is necessary.
Admixture analysis is appropriate when populations form by admixture of pre-existing and differentiated components, e.g. Caucasoids and Mongoloids in Central Asia. But, in the case of Europe the existence of such differentiated components and their "admixture" is not evident.
So, even though an admixture analysis (e.g. using STRUCTURE) can be performed, its results can't be interpreted as proportions from distinct ancestral groups.
Dienekes, start reading at the end of page 3...last paragraph.
The report talks about migrations during the Palolithic, Mesolithic and Neolithic as the main culprits in forming the foundations of European genetic substructure.
Obviously these migrations resulted in different clusters in Europe because they were thousands of years apart. But these clusters were blurred together by more recent historic mixing, usually between neighboring groups. That's why...doh...Swedes don't overlap with Italians, etc. But every nation here does overlap with its neighbours.
So guess what...the populations that went north during the Mesolithic developed a new character there...and when they came back south duirng the Middle Ages, they gave some of that pimpjuice to the Northern Greeks.
And correct me if I'm wrong, but wasn't Slavic spoken in Northern Greece until very recently?
The report talks about migrations during the Palolithic, Mesolithic and Neolithic as the main culprits in forming the foundations of European genetic substructure.
They note that the migrations that shaped Europeans all proceeded in the expected geographical direction (S->N or SE->NW); they do not however make any claim about the relative importance of these migrations.
So, the notion that the differences are of Mesolithic or Paleolithic origin are not supported by this paper.
WTF?
They say specifically that the "major" migrations during the Paleolithic, Mesolithic and Neolithic set the tone in Europe.
And then, more recent historic events have had, and I quote, "a more homogenizing effect upon the previously established genetic landscape".
Obviously we can still make out that "previously established genetic landscape" even tho these days it's made up of clines rather than clear clusters.
A structure diagram would show the different components of this genetic landscape, even tho today no population is carying just one.
So the question is Dienekes, why need a homogenizing effect for something that is, according to you, already homogenous? Hey?
They say specifically that the "major" migrations during the Paleolithic, Mesolithic and Neolithic set the tone in Europe.
That is archaeology not genetics. There are no dates on this paper; no attempt to show when the genetic differentiation established itself.
You haven't answered my quesion...
Why the need for a homogenizing effect for something that is, according to you, already homogenous?
I criticized your claim that genetic differentiation in Europe today stems from the Paleolithic and Mesolithic. Most genes with salient differences within Europe aren't really that old.
The fact that recent population movements in Europe homogenized the pre-existing landscape is irrelevant to the question of when the pre-existing landscape formed itself.
There is genetic substructure in Europe, even though it's been homogenized by recent historic events, and is now seen in the form of clines rather than clusters.
So the upshot is, those Northern Greeks carry Central European markers IMO.
"The issue at hand is what, if anything, this study says about the Anglo-Saxon invasion. If one wants to make claims about the relatedness of the English and Irish, one needs to sample Englishmen--not some undefined mix of English, Irish, Scottish, and Welsh.
"By the way the Scots and Welsh are British and not Irish."
Dude, what the fuck are you going on about? Who said they were Irish?
The Scots and Welsh are British, but they are not English. Genetically, they are most likely intermediate between the English and Irish."
Presumably some of the London sample are English, stands to reason - I have London ancestry going back to the mid 1600s, before that they came from Kent, not Kerry. If so why the lack of significant overlap with North Germans and Danes? Maybe the Londoners on the end of the envelope away from the continentals are all Irish or Welsh, but why do the ones on the other end overlap the Dutch and not the historical prime suspects.
Either Britain was invaded by a few boatloads of pirates, not thousands of peasants, from Angeln, Jutland and Old Saxony, or Bede was very geographically challenged and by "Angles Saxons and Jutes" he really meant "Frisians and Batavians."
I'm afraid this paper in no way supports recent historical events having any real effect on the genetics of the British Isles and every support to a view that The British Isles and the nearest parts of the continent having very ancient history of genetic continuity.
Mind the language, you were asserting that London was a mixed population because of Scottish and Welsh immigration, but this is just from other areas of the UK (one of the most populous areas of Scotland, Lothian and Borders, was part of the English kingdom of Northumbria anyway).
The only surprice here is the distance between ALL the Finnish and the Swedish samples. Basicly, considering the recorded population history one would expect that at least 2 or 3 Finnish samples would cluster very closely to the Swedish ones, instead of one loosely connected. I remember reading preliminary findings of a study that used a sample from Kuusamo, a place with a very peculiar population history, with a dual founder effect in the past millenia and possible bottlenecks. If this was the one, it would explain a lot.
The only surprice here is the distance between ALL the Finnish and the Swedish samples.
That also happened in Baucher-2007: though there were no other Scandinavian samples, Finns stood apart from Germans anyhow, that in this study strongly overlap with Swedes. And that after discounting the Siberian (Altaian-like) component.
I guess that they have sampled only ethical Finns, not Swedes from Finland. The likely reason is that Finland was originally colonized from Eastern Europe, while Scandinavia was from Central Europe.
I am talking of Epipaleolithic, of course - though the same applies for Neolithic influences anyhow.
As Eastern Europe is absent from both studies, Finland excepted, Finns tends to appear isolated.
But that doesn't mean there is no connection between Finland and Central/Northern Europe. In Baucher's K-means clustering the genetic component clearly dominant among Finns (blue) was also somewhat important in Central/NW Europe, specially in Poland, with traces found in all Europe.
Inversely, as you predict, some Finns showed relevant Central/Northern component as well (green).
And (side comment), now that I think of it, the Basque component (orange) was oddly common among all European populations (but not Jews or Armenians). So good for those who claim that European genetic structure should be dated to the Neolithic! It's impossible that Basques influenced significatively Italy or Germany so late in history, much less Greece or Finland.
maju said,
"And (side comment), now that I think of it, the Basque component (orange) was oddly common among all European populations (but not Jews or Armenians). So good for those who claim that European genetic structure should be dated to the Neolithic! It's impossible that Basques influenced significatively Italy or Germany so late in history, much less Greece or Finland."
Unfortunately for maju and his delusions of a Basque origin of Y-DNA haplogroup R1b (or at least R1b1b2-M269), haplogroup R1b is the most common Y-DNA haplogroup among Armenians and it also has a significant presence among Jews, so the fact that the Jews and Armenians appear to have the least Basque-like influence in their autosomal DNA is a strike against the Basque origin of R1b theory. Perhaps the true Basque Y-DNA is some form of haplogroup I-M170; that would explain why a Basque-like autosomal component forms a significant part of the genetic composition of all European populations, but not SW Asian ones like Jews or Armenians.
urselius,
I'm trying very hard to be patient here, but you're either very slow or dishonest.
Presumably some of the London sample are English, stands to reason
One would hope so. But it's impossible to know how many, if any, "pure" English are present in the UK sample, or to which part of England they trace their ancestry.
I have London ancestry going back to the mid 1600s, before that they came from Kent, not Kerry
Is this you?
"This user is of English ancestry."
"This user is of Irish and British ancestry."
Tends to prove my point.
(Tending to confirm another suspicion: "This user is very interested in Napoléon.")
but why do the ones on the other end overlap the Dutch and not the historical prime suspects.
Re-read this thread. As I've already pointed out, what you insist on referring to as the "North German" sample comes from a town facing the Baltic (founded long after the 5th century, by the way), and the Danish sample is from Copenhagen. Do you understand where Copenhagen is? It's on an island. On the opposite side of Jutland from Britain.
a few boatloads of pirates, not thousands of peasants, from Angeln, Jutland and Old Saxony
Might help to actually sample descendants of peasants from Jutland and northwestern Germany before making declarations like this.
Other sources (e.g. Cavalli-Sforza) show Danes to be among the closest genetic neighbors of the English.
I'm afraid this paper in no way supports recent historical events having any real effect on the genetics of the British Isles
I'm afraid repeating something doesn't make it true.
This study was not designed to answer that question. The results are entirely consistent with a historical mass-migration of Germanic-speakers. They may also be consistent with a version of history you'd prefer. It's not possible to strongly discriminate between the two possibilities using these results.
Look at the "People of the British Isles" project if you want to see a study that may have something useful to say on the subject in coming years.
@Ebizur: I said nothing about R1b here. We have discussed it elsewhere and I don't think it ha room here. I just mentioned the Basque autosomal component as per Baucher et al, 2007.
Surely the very meaningful presence of such component among Italians (likely North Italians) is because of tourism to Venice and the not despicable presence of it among Greeks is because of the Almogavars. Equally the more relevant presence of the Basque cluster among Germans and Irish is surely because of cod fishermen... and well, the presence of some 5% of that Basque component among Poles is of course caused by Jesuitic promiscuity. Honestly, I cannot think of any other possible Basque-Polish connection reason... except, doh!, Paleolithic elements.
But, well, it's of course just my feverish mind. How can I even dare to think that Paleolithic had any relevance in European genetics?! How did I dare to notice that the Franco-Cantabrian region was the most important European area once upon a time. After all they were just hunter-gatherers, c'mon!
n/a said Other sources (e.g. Cavalli-Sforza) show Danes to be among the closest genetic neighbors of the English.
Do you mean his famous PC maps? There's no dominant component in them for the Brits. He seems to have totally missed the North European component of Baucher.
I also think that the historical Scandinavian and Low German influence among Brits was meaningful but limited. Even Y-DNA ancestry studies show that the North European influence is limited and clinal - and certainly Vikings were not Valkirias. And their genetic impact should not have been much bigger than that of Celts or older waves, including the decicissive original Epipaleolithic colonization of all those areas, that must have a partly common origin (Doggerland? Rhineland?).
Do you mean his famous PC maps?
No. I mean Table 5.5.1 in HGHG.
The area where NO, DK, and DE1 intersect seems to be the "Germanic core". DE1 stretches away from this area in the direction of PO, SE in the direction of FI, DE2 in the direction of CH and AT and UK to the left.
This is consistent with the origin of Germanics in N Germany/Denmark/S Sweden. Norwegians, being on the Atlantic coast and neighboring with other Germanics are predictably part of the "core". So are the Dutch, who are also neighbors only of other Germanic people.
DE1, DE2, UK, and SE who are further from the source region and/or neighbor non-Germanic peoples are predictably further away from the "core".
Dear n/a,
To quote:
"I'm trying very hard to be patient here, but you're either very slow or dishonest."
and:
"Is this you?
"This user is of English ancestry."
"This user is of Irish and British ancestry.""
What a busy weasel you are! I guess you have answered your question, to your own satisfaction at least.
You will have noted that I have a very good knowledge of history and that I'm a molecular biologist by trade. This is unfortunately not a common combination, historians normally have a poor grasp of scientific method and scientists (as evidenced by Oppenhiemer) have, in general a feeble depth of historical knowledge. With my background, by your lights, as I'm demonstrbly not slow, I must be dishonest.
To quote:
"Presumably some of the London sample are English, stands to reason
One would hope so. But it's impossible to know how many, if any, "pure" English are present in the UK sample, or to which part of England they trace their ancestry.
I have London ancestry going back to the mid 1600s, before that they came from Kent, not Kerry
Tends to prove my point."
But my London ancestors had no Irish ancestry at all and possessed an unusual surname of early Anglo-Saxon origins (an 'ingas' name) - my Irish ancestry came from Ireland.
quote:
"(Tending to confirm another suspicion: "This user is very interested in Napoléon.")"
Wha?? What suspicion is this? If you look through my wiki additions they are mostly on British military figures of the Napoleonic period. OK, one had French ancestry, but seeing as he died, sword in hand, leading his men in the destruction of half a dozen French battalions at Salamanca, I think his - shudder - Gallic ancestry can be forgiven.
Quote:
Re-read this thread. As I've already pointed out, what you insist on referring to as the "North German" sample comes from a town facing the Baltic (founded long after the 5th century, by the way), and the Danish sample is from Copenhagen. Do you understand where Copenhagen is? It's on an island. On the opposite side of Jutland from Britain.
Well Kiel is the place used for the N. German sampling, and you can spit from Kiel to Angeln where the Angles are supposed to originate. Kiel is a damned site closer to all the areas of supposed Anglo-Saxon origins than Rotterdam is.
Quote:
"a few boatloads of pirates, not thousands of peasants, from Angeln, Jutland and Old Saxony
Might help to actually sample descendants of peasants from Jutland and northwestern Germany before making declarations like this."
Good idea, and an even better one is include samples from other areas of the near continent at the same time. If the London group had sampled from Belgium and northern France as well as England and Wales and Denmark and N.Germany then their results would have been more complete and meaningful.
Quote:
""I'm afraid this paper in no way supports recent historical events having any real effect on the genetics of the British Isles"
I'm afraid repeating something doesn't make it true.
This study was not designed to answer that question. The results are entirely consistent with a historical mass-migration of Germanic-speakers. They may also be consistent with a version of history you'd prefer. It's not possible to strongly discriminate between the two possibilities using these results.""
Y-e-e-s - because of the huge congruency between the UK envelope and those of the Danes and N. Germans.
Keep taking the pills.
Dienekes:
"This is consistent with the origin of Germanics in N Germany/Denmark/S Sweden. Norwegians, being on the Atlantic coast and neighboring with other Germanics are predictably part of the "core". So are the Dutch, who are also neighbors only of other Germanic people."
I think you have confused linguistic origins and ethnic origins. Germanic seems to have had a common homeland in southern Scandinavia and northern Germany. However, the people who speak Germanic languages today away from the linguistic homelands undoubtedly include many genetic influences from people once speaking other language groups. It has been admitted for many years that High German has a probable Celtic basis for its differences in pronunciation from Low German, for instance. There is also evidence for an ancient non-Germanic and non-Celtic language being spoken in the Netherlands and surrounding areas.
And the Dutch sample overlaps the Irish considerably, so saying it doesn't neighbour non-Germanic groups is not really tenable.
And the Dutch sample overlaps the Irish considerably, so saying it doesn't neighbour non-Germanic groups is not really tenable.
The envelope of the Dutch cluster is small, and close to the Germanic core region, whereas the envelope of the Irish is wide, and stretches further to the left.
The Irish do have some Germanic ancestry, but they are more variable than the Dutch who form a tight spherical cluster very slightly little off from the "core region".
It's a matter of interpretation, certainly.
I don't really see a "Germanic cluster" as such, I see three overlapping clusters with a mixture of Germanic and some non-Germanic languages.
First I see an "Alpine cluster" - France, Switzerland, Austria and S. Germany with this overlapping, to a degree, a "South Baltic cluster" - N. Germany, Denmark, and Sweden and finally a "North Sea cluster" of UK, Ireland and Holland with Norway being intermediate between the Baltic and North Sea clusters.
I agree with Dienekes for a change...well, to a point.
The true Germanic cluster is where Norway, Denmark, Sweden and DE1 (Kiel) overlap.
The Dutch sample is not a good example of anything strongly Germanic, because it comes from Rotterdam. So yeah, most of the Dutch here, as well as many of the British show Germanic influence, but aren't part of the core.
On the right, individual Germans overlap almost completely with the Poles from Warsaw, which is kinda in east Poland (enlarge the PDF verson of the diagram in the original report to see it properly). That would seem hell weird, but not really, because as we know Kiel was founded by Slavs, and there were even Slavic settlements as far as Bavaria. Pity that Poles/Pomeranians from the Gdansk region weren't sampled. I reckon they'd pull to the left quite a bit.
So we can see some ancient links there between the Celtics, Germanics and Slavs in the north, and also some recent mixing between the Germans and Celts on one hand, and Germans and Slavs on the other.
Btw, Tuulia, the Finns come from Helsinki, with parents and g-parents from all over the country.
Samples from North Sweden would probably be closer to Finland. So would the Baltic States. On the other and, Finland Swedes would probably be close to Sweden.
Cheers
By the way, here are some interesting comments from NY Times on this study...
"The map also identifies the existence of two genetic barriers within Europe. One is between the Finns (light blue, upper right) and other Europeans. It arose because the Finnish population was at one time very small and then expanded, bearing the atypical genetics of its few founders.
The other is between Italians (yellow, bottom center) and the rest. This may reflect the role of the Alps in impeding free flow of people between Italy and the rest of Europe."
http://www.nytimes.com/2008/08/13/science/13visual.html?_r=1&oref=slogin
>> The true Germanic cluster is where Norway, Denmark, Sweden and DE1 (Kiel) overlap.
Sweden doesn't overlap with Denmark and Norway, they only overlap with DE1. They are away, on the axis towards FI, not surprisingly given that ~15% of Sweidsh Y-chromosomes are N1c. A different Swedish sample might be closer to the core, however.
Sweden does overlap with Denmark and Norway. You can see it both on the individual and group diagrams. Just enlarge the group plot on your PDF, and you'll see a secondary Swedish cluster.
Anyway, that's beside the point. If they had samples of Eastern Norway, that Germanic core cluster would enlarge to fit in most of this current Swedish sample...although a Northern Swedish sample would fall out of the core and just be Germanic influenced IMO. And that would be because of Finnish influence.
However, this stuff doesn't have much to do with N. Forget the Uralic admix here Dienekes, this is all about genome wide markers of intra-European value. Those Finns aren't on the far right becaue of any N and Uralics, they're there because their West Eurasian genes were isolaed for a long time.
The position of the Finns here is all due to isolation and drift.
They too mostly originated in the south of Europe, but a small number of founders brought only a limited amount of variation with them...and then they got stuck way up in the north for a long time.
The fact that they migrated through Poland and the Baltic States, means that they now show similarity to Eastern Europe. They also pull towards the Swedes somewhat because of more recent Swedish influence.
But like I said, their unusual position has nothing to do with Uralic admixture. This is like the second part of the Bauchet study, after the Altaina/Uralic/Asian genes were taken out and only intra-European variation tested.
The position of the Finns here is all due to isolation and drift.
They too mostly originated in the south of Europe, but a small number of founders brought only a limited amount of variation with them...and then they got stuck way up in the north for a long time.
More likely due to the fact that a major component of their ancestry originated in Siberia about 4 thousand years ago.
If they had samples of Eastern Norway, that Germanic core cluster would enlarge to fit in most of this current Swedish sample
Eastern Norway only has 3% of N1c, unlike Sweden where it is ~15%. Finno-Ugrian admixture unsurprisingly moves this Swedish sample away from the core and towards FI. Perhaps other Swedish regions (like the 4 isolated individuals in the tiny SE cluster to the left) might form part of the core, however.
"More likely due to the fact that a major component of their ancestry originated in Siberia about 4 thousand years ago."
Dienekes, if this only was about the y-chromosome markers, Finns would definitely cluster closer to Swedes. Not only because of the presence of N1c in Sweden, but also because of the presence of I1a in Sweden. So, I too tend to see isolation as the main explanation for the difference.
And while there probably is an "Siberian" element in Finnish (and
any North Eastern European, for that matter) ancestry, I thought the latest consensus among the genesists was that it's much deeper than 4 000 years you suggest.
Polak said: The position of the Finns here is all due to isolation and drift.
Dienekes replied: More likely due to the fact that a major component of their ancestry originated in Siberia about 4 thousand years ago.
I rather think that it's due to their main nDNA component being Eastern European. Unlike Scandinavia, Finland was originally colonized from Eastern Europe and most later influences (except those from Sweden) arrived from there.
That would also explain better the relatively high presence of the typical Finnic component among other Europeans, specially Northern ones, found in Bauchet'07. This is not easy to explain if we consider that component merely "Finnic" but makes sense if it's relatively common in Eastern Europe.
The Siberian component among Finns was not anyhow more than 10-15% in Bauchet'07. What made Finns stand out among the other Europeans sampled was a component that looked from Europe, that was clearly majoritarian among Finns and that I understand that reflects some Eastern European component; Eastern Europeans were not sampled in either case but Poles showed a big deal of that component too.
Not only because of the presence of N1c in Sweden, but also because of the presence of I1a in Sweden.
But at quite different frequencies.
I thought the latest consensus among the genesists was that it's much deeper than 4 000 years you suggest.
The two most recent studies on this by Lappalainen et al. and Derenko et al. have both used the "evolutionary mutation rate" that is 3.6x slower than the germline mutation rate. See also this; there is also a comment by Tuuli Lappalainen on it.
The Siberian component among Finns was not anyhow more than 10-15% in Bauchet'07.
10-15% of the distance between Europeans and Siberians is more than enough to move Finns to the right.
But they are also different because of other factors, the "Siberian" component is of lesser importance in cmparison with them. Notice that again in this paper it is also the Polish who appear closer to the Finns but the Polish did not show significative "Siberian" component in Bauchet'07 but they showed high (even if secondary) "Finnic" component.
Maju is right here. Dienekes is way off.
The Siberian component in the Finns is irrelevant in this case.
The only relevant thing is that Finland was populated via Poland and Eastern Europe by a small group of founders, who were then isolated.
So they're over to the far right because they're ultra-Eastern Europeans, but with a strong left-right cline due to more recent admixture from Swedes (Central European influence).
After the Finns, Poles have most of this Eastern European component, then the Eastern Germanics.
The Atlantic fringe populations, and thus the Western Germanics, show very little of this component, alhought it is there, even as far as Ireland.
So let me reiterate, this has nothing to do with N1c. Dienekes is so wrong it hurts.
Hey Dienekes,
The region of Poland sampled here has 1.7% of N1c (Kayser 2005).
Polak: I am glad that we agree on something for a change. Nevertheless:
... Finland was populated via Poland and Eastern Europe...
I rather tend to think in terms of "via Eastern Europe" exclusively. I don't understand why you include Poland in this. Poland probably recieved that Eastern component from a similar source (the East) but in Epipaleolithic times it belonged to the Central/Northern ethno-cultural area an that's why it mostly clusters with them. The Eastern component in Poles is surely of later arrival (Dniepr-Don northern offshots, IEs, Slavs...).
Maju,
You might be right.
I was just speculating that a specific group of early southerners may have moved via Central-Eastern Europe up to Finland, thus leaving their legacy in Poland along the way. Remember that the Saami have mtDNA lineages found in Baques and Berbers.
That "Eastern European" admixture, as you have observed as well, is found all the way from the Atlantic to Finland (and probably well into Russia). If it didn't originaly make its way north through Central-East Europe somwhere, it would have to have been transported west much later. And yeah, maybe that did happen?
Poles and Germans may have got that influence from Slavic speakers, the Swedes and Norwegians from Finnic speakers moving down (fe. Swedish Forest Finns), and the Irish and British from Vikings.
But I don't know. Seems to me this influence is really high in Eastern Germanic regions. Maybe too high even for later Finnic and Slavic admixture to be enough of an explanation?
But I think if we sample more regions, and analyse STRCUCTURE as well, we'll be able to figure out the real distribution, frequency, and thus source of this EE component...which is certainly not assciated with N1c carriers.
Here's a theory...
The blue markers come from Eastern Europe, and are certainly of European origin. Here they are from Bauchet 2007...
http://img246.imageshack.us/img246/6762/intraeuropeangeneticstrok6.jpg
They're found in high frequencies in populations carrying a lot of R1a1, all the way from the Ukranian steppes to Western Siberia.
They're carried to Central Europe along with R1a1 with pre-Slavic and Slavic migrations. The Eastern Germanics thus absorb a fair bit of the stuff, and the Poles even more.
The early Baltic Finns are actually Eastern Germanics...Goths if you will, who carry I1a, some Central European R1a1, and a lot of the green and blue.
However, they're overrun by migrants from the east carrying heaps of N1c. But despite this Uralic marker, these people are largely Caucasoid as a result of mixing with R1a1 Caucasoid Siberians earlier on, and thus loaded with the EE blue markers.
So, in the end, the Baltic Finns end up with a lot of N1c, tonnes of the blue stuff, and only 10-15% Siberian genome wide influence. Also, their earlier "Gothic" I1a is supplemented by more recent Swedish I1a.
How's that?
Here's a theory...
The blue markers come from Eastern Europe, and are certainly of European origin. Here they are from Bauchet 2007...
http://img246.imageshack.us/img246/6762/intraeuropeangeneticstrok6.jpg
They're found in high frequencies in populations carrying a lot of R1a1, all the way from the Ukranian steppes to Western Siberia.
Ok so far. Except that we have no data on Eastern Europe other than Finland.
They're carried to Central Europe along with R1a1 with pre-Slavic and Slavic migrations. The Eastern Germanics thus absorb a fair bit of the stuff, and the Poles even more.
The early Baltic Finns are actually Eastern Germanics...Goths if you will, who carry I1a, some Central European R1a1, and a lot of the green and blue.
This part is less of my like. Goths probably played just a very minor role in all the picture. I understand that most people have poor knowledge of Neolithic and specially Chalcolithic processes in Europe and therefore tend to seek explanations in later, better understood, periods, whose "nations" already have names we can comprehend easily (names like "Goths", "Celts", "Greeks", etc.). But Chalcolithic migrations (and those before them as well) were at least as important if not even more.
For instance an migration that almost nobody mentions is that f pre-IE "Ukranians" into Baltic Europe, influencing both the Eastern Baltic and Scandinavia/Low Germany. Of course this migration is largely similar to the later ones by IEs but it also has a more specifically Baltic scope and IMO can account for Nordic Y-DNA I - and, in the contextof this discussion, maybe also for the peculiar northern distribution of the "Finnic" or "Eastern" nDNA cluster, less important in Germany for instance.
Just a guess anyhow. Indo-European migrations can also explain it largely, though they seem more oriented to Central Europe than Northern Europe as such. Of course, latter migrations (Slavs, Scythians, maybe even Goths to a minor extent) may have also added up but I would not think of them as so crucial factors.
My two cents anyhow.
Yeah, I just mentioned the Goths because I know that Finnish has a lot of proto-Gothic in it. In fact, some commonly used Finnish sentences basically sound Gothic.
The Finns are a simple mix of aboriginal northern Europeans and Siberian Uralic speakers
Due to their language, remoteness, and isolation, they did not receive many southern elements in more recent times. Thus, they occupy the top position in the first eigenvector which shows the clinal variation between northern and southern Europeans.
On the second eigenvector though, they are strongly differentiated from other Europeans. Along this eigenvector their distance to Poles is about the same as the distance of Poles to the Portuguese, even though the Poles are much closer to the Finns geographically (and longitudinally) than they are to the Portuguese.
Thus, the second eigenvector is not dominated by a cline from western to eastern Europe, but its most salient feature is the differentiation between Finns and the rest.
The discontinuity is evidenced also in Y-chromosome distribution where the Finno-Ugrian element reaches high frequency among Finno-Ugrians, Balts and northern Russians, and drops sharply among Poles.
It is this discontinuity, the result of the arrival of Uralic speakers on a north European substratum that is responsible for the special position of the Finns. Finns are at some distance from the main European "blob" not due to any mystery East European component, but because, like the northern Russians of Li et al. (2008) they are not fully of Caucasoid origin genetically but also have components aligning them to East Eurasians.
As for Bauchet et al., his results are consistent with Asian admixture being the contributing factor to the position of the Finns.
http://dienekes.110mb.com/articles/greekadna/pca.jpg
If genetic drift was responsible, then we would not expect the Finns to deviate from Europeans in the direction of East Eurasians, as is evident; this is consistent with East Eurasian admixture being the cause of the Finns' distance from the European cluster.
dienekes said,
"If genetic drift was responsible, then we would not expect the Finns to deviate from Europeans in the direction of East Eurasians, as is evident; this is consistent with East Eurasian admixture being the cause of the Finns' distance from the European cluster."
On the contrary, we would not expect the Finns' minor deviation from the major cluster of modern Europeans in the direction of modern East Eurasians if the East Eurasian admixture in the Finns were directly correlated with the presence of haplogroup N1c in the Finnish population; this haplogroup is found in 60% to 70% of Finns versus less than 10% of Altaians, for example. Polak and the others are correct; haplogroup N1c definitely has nothing to do with East Eurasian admixture in the Finns and other Eastern Europeans.
Finns are distant from Europeans because of their East Eurasian admixture. Under the theory of "drift due to isolation", the Finns might be distant from other Europeans, but not specifically at an East Eurasian direction.
It is irrelevant that Altaians may not have a high percentage of N1c. N1c is the evidence of East Eurasians migrating to Europe; it does not follow that all East Eurasians should possess N1c.
If, for example an East Eurasian population mixed with Portuguese or Poles, they would introduce different Y-chromosome types to the East Eurasian population, yet in both cases autosomally they would become closer to Europeans.
It is certainly not irrelevant that the Altaians do not have a high percentage of N1c; if haplogroup N1c is found in a greater percentage of Finns and Slavic Russians than Turkic Altaians, Kyrgyz, Kazakhs, etc., then it can only be related to East Eurasian admixture in the Eastern European populations by a scenario that incorporates a great deal of genetic drift, or else through the simple admixture process that you have envisioned followed by a subsequent replacement of the original N1c Y-DNA of the purported East Eurasian source populations with some other Y-DNA haplogroup. This is not difficult to understand, dienekes, and I'm really sad to see that you are disputing such an obvious fact.
Are you disputing the fact that N1c came from East Eurasians to Europeans and not vice versa?
Are you disputing the fact that Finns and Northern Russians have East Eurasian autosomal admixture?
Are you saying that by some magical accident the populations who brought N1c into Europe from Asia did not carry East Eurasian autosomal DNA with them? If not, then how did the Finns and Northern Russians acquire their East Eurasian autosomal admixture?
I still think intra-European diversity is the key here.
If the Finns are being pushed out right because of Siberian influence, why are the Brits being pushed left? Sub-Saharan Admixture? Of course not, it's just isolation and drift.
Some of these North European populations have very little genetic diversity indeed. The Southerners have much more.
Also, East Eurasian populations weren't sampled here. So even if these Finns carry something exotic, like 10-15% of that orange in Bauchet, how can that be interpreted by the plot?
Once Bauchet dropped the exotic stuff, the Finns still showed up in the same place as in this study...above right of the North Europeans.
>> If the Finns are being pushed out right because of Siberian influence, why are the Brits being pushed left?
It is not right to think that a single process is responsible for each of the principal components. As the authors say, referring to the 2nd one, multiple processes may have contributed to the north-south cline. So, the fact that PC2 differentiates between W and E Europe (e.g. Spain vs. Greece or Britain vs. Poland), doesn't mean that its most salient feature isn't the differentiation between Finns and the rest.
Ebizur wrote: "we would not expect the Finns' minor deviation from the major cluster of modern Europeans in the direction of modern East Eurasians if the East Eurasian admixture in the Finns were directly correlated with the presence of haplogroup N1c in the Finnish population".
The relevant word here is 'directly'.
As Dienekes wrote: "If, for example an East Eurasian population mixed with Portuguese or Poles, they would introduce different Y-chromosome types to the East Eurasian population, yet in both cases autosomally they would become closer to Europeans".
Once again we see an example of the relative independence of Y-chromosome haplogroup and autosomal DNA. In other words haplogroups can flow through a population relatively independent of aDNA. And yet people many people are still prepared to construct major theories about our evolution based simply on the distribution of haplogroups.
It is certainly an unfortunate example, nevertheless haplogroups more often than not seem show correlation with nDNA and X-DNA. The correlation is not lineal, of course, but it does exist. This is particularly true for mtDNA but it also applies to an extent to Y-DNA, in spite of gender bias.
A good and well studied example is Latin America. The exact correlations between the different markers (haplogroups, X-DNA, nDNA) vary somewhat but they all suggest and evidence the same kind of demic alchemy. IMO, we should try to consider haplogroup genetics in Y-DNA/mtDNA pairs, in any case (and of the two mtDNA is often more informative).
the fact that PC2 differentiates between W and E Europe (e.g. Spain vs. Greece or Britain vs. Poland), doesn't mean that its most salient feature isn't the differentiation between Finns and the rest.
This was not so in Bauchet'07. In that study PC2 was structured between Basques and Spanish on one side and Finns and Armenians and Sicilians in the other. Finns and Armenians made up the PC1, so these two suggested a triangular main structure.
Just for the record, Bauchet's PC3 was about the Basque vs. Finnish+Spanish axis, PC4 Finns vs. English+Irish, PC5 Italians vs. Armenians+Basque+Polish+Irish and PC6 was not well defined in regional terms.
Anyhow, it's not like anyone of those simplified graphs is going to show the "exact" strutucture: differences in sample sizes and their relative weight in the PC-determination may alter the whole graph a lot. Certainly a bidimensional graph cannot show the multidimensional complexity of genetic structure properly, just describe the most noticeable elements - and that's all.
Also it's becoming quite evident that ignoring Eastern Europe and West Asia in European genetic studies is not very helpful.
terryt,
"we would not expect the Finns' minor deviation from the major cluster of modern Europeans in the direction of modern East Eurasians if the East Eurasian admixture in the Finns were directly correlated with the presence of haplogroup N1c in the Finnish population"
Actually, the relevant word here is "minor"; perhaps I should have written this sentence as "we would not expect the Finns' deviation from the major cluster of modern Europeans to be so minor if the deviation were due to admixture with East Eurasians bearing haplogroup N1c Y-DNA."
The problem is that somewhere between 60% and 70% of the Finns belong to haplogroup N1c, whereas approximately 1% of the Uralic peoples' "East Eurasian" neighbors, the Turkic peoples, belong to haplogroup N1c. On average, even Slavic and Germanic peoples have more haplogroup N1c than Turkic peoples have. So, I reiterate, haplogroup N1c cannot be directly correlated with East Eurasian autosomal influence in the Finns or any other European population. That is not to deny the possibility of East Eurasian autosomal admixture in some eastern European populations.
perhaps I should have written this sentence as "we would not expect the Finns' deviation from the major cluster of modern Europeans to be so minor if the deviation were due to admixture with East Eurasians bearing haplogroup N1c Y-DNA."
A single genetic marker provides a very noisy estimate of genetic admixture. Thus, no, we wouldn't expect the extent of non-European admixture to be accurately quantified by the frequency of N1c.
N1c is useful because the phylogegraphic determination of its origin establishes the fact of migration of eastern Eurasians into eastern Europe. This fact is consistent with the non-Caucasoid admixture in some eastern European populations. This non-Caucasoid genomic admixture (~1/10) predictably moves these populations away from the European cluster, in the direction of East Eurasians.
Northern Swedes do not cluster with Finns. Swedish-speaking Ostrobothnians do cluster with Swedes.
Well, it's interesting that in this study there are Swedish samples a lot closer to Finns than this Uppsala sample. They're not in the North, but they are closer.
http://www.biomedcentral.com/content/pdf/1471-2156-9-54.pdf
@Polak:
Fig.2c shows Swedes clustering exremely distant from either Finnish population. If Sweden-Eastern Finland defines the PC1, Western Finland alone defines the PC2 (both vs. Sweden and Eastern Finland).
Fig.2d can lead to totally different conclussions though but I still find odd that SSO Fino-Swedes cluster best with SKA Swedes (Danes by origin) than with any other Swedish sample.
In any case, the difference between these two figures, made up with the same data and software, suggests me to remain very cautious re. PC graphs.
They're only using 34 SNPs. That might explain the randomness of the Swedish samples.
In other studies, using a few hundred SNPs, there's almost no difference at all between Northern Europeans, except Finns (see Price et al. 2008). But clines do appear after a few thousand markers, like in this Lao study.
In other studies, using a few hundred SNPs, there's almost no difference at all between Northern Europeans, except Finns (see Price et al. 2008).
Which makes it all the more noteworthy that Northern Swedes cluster separately from Finns.
A little birdie just told me that yes, that is correct. Northern Swedes aren't all that close to Finns.
But the Finnish Swedes do cluster in between the Finns and Swedes. There's a new study coming out on that.
Still, my point before was that Uppsala Swedes aren't the closest Swedish sample to Finland. There are other regions closer to Finland, although not extremely close.
"nevertheless haplogroups more often than not seem show correlation with nDNA and X-DNA". I think we should play it safe and just say the jury is still out on that one. There are certainly plenty of exceptions and as far as I know no definite examples of correlation.
Terry: I would not like to start a discussion on that after 107 comments. But certainly there are clear examples of correlation, like that study on American Mestizos a few months ago. Obviously if you look only at the Y-DNA, you may get confused and think in terms of native Europeans or something close but overall both haploid lines are pretty much representative of what you find by other means. Y-DNA alone isn't.
Still, my point before was that Uppsala Swedes aren't the closest Swedish sample to Finland. There are other regions closer to Finland, although not extremely close.
I wouldn't place too much emphasis on small differences at the county level, which may just be noise.
In addition to the smaller sample sizes at the county level, in the county-level PCA, the first two PCs combine to explain only 31% of variation, vs. 69% for the first PC alone in the regional PCA. At the regional level, Central Sweden is closest to Finland. This could again just be noise, or could be the result of historical Finnish immigration.
"But certainly there are clear examples of correlation, like that study on American Mestizos a few months ago".
Hang on Maju. You're not saying that in this case the Y-chromosomes and mtDNA lines involved came in together are you?
Sure, if this particular combination were to now expand to some other region and remain genetically isolated from their neighbours we would then have an example of the kind of correlation we're looking for but all we have here so far is an example of mixing, a temporary correlation of aDNA and haplogroups.
I suspect it is most likely that the sort of mixing we have in the Carribean is exactly the same sort of mixing that has been happening since Australopithecus first appeared, in fact presumably long before that time.
n/a,
I mentioned the Forest Finns in a post above. They're the reason why I think Eastern Norway and Western Sweden might cluster a bit closer to the Finns. But we won't really know until they're tested.
IT2 sample is the italian region Marches. This region is on the west cost, but it overlap with the italians western neighbors, while IT1 has a small overlap with a distant country as Romania and with the oriental neighbors.
the Emilia and the Marches are the regions occupied by the gauls Boi and from the Senonis in historical age and this can explain the thing. The overlap with Romania is amazing, I had never believed that there had been a remarkable colonization from the Italian peninsula in Dacia, but is better wait for more tests.
pardon, Marches is on the east coast, of course.
This is my final take on Northern Europe here...
There appear to be three main population isolates from which North/Central Europe was populated in ancient times...what I would today call the Atlantic, Central and Baltic clusters.
Finns are absolutely full of the Baltic stuff, and have very litte of the Central and Atlantic components. Although this does vary, depending on which part of Finland they come from, so that's why there's a strong left-right cline in the Finnish sample.
Poles and Swedes are the next up with the Baltic influence, but the Swedes show more Atlantic admixture than Poles, and Poles more Central than Swedes. The Northern Germans are a really mixed bunch, showing a lot of Atlantic, Central and Baltic influence. The Danes and Western Norwegians are basically like more Atlantic Swedes, while the Dutch are like the more Atlantic Northern Germans.
Finally, the most Atlantic and least Baltic are the Brits and Irish. However, some Germanic influence (and hence Central and Baltic admixture) is seen in some of the Brits. At the same time, some of the British samples look to be purely "Atlantic". In fact, they are almost as Atlantic as the Finns are Baltic, hence their isolation on the far left.
There, too easy...
There appear to be three main population isolates from which North/Central Europe was populated in ancient times...what I would today call the Atlantic, Central and Baltic clusters.
You are forgetting Siberia, the source of Finns' language, 50%+ of their patrilineages and ~1/10 of their autosomal genes.
Dienekes, seriously, this is an intra-continental perspective. It's like that second part of the Bauchet study, once the Altaians were dropped.
The Finns don't need anything exotic to push them out that far. They're a small country, originating from a small number of isolated founders. A lot like the west coast Brits.
The distances on this PC plot are actually a lot less interesting then we think. They're more about people getting in on with their cousins (local breeding isolates) than an illustration of historic events and great migrations.
Dienekes,
little stubborn there, are we?
According to the authors of the study whom my friend over HBF contacted, the Finns stood apart thanks to isolation and drift just like the NY times article addressed, not because of Siberian admixture which BTW can be detected in every European population at various rates.
Siberian influence is a completely different issue not even addressed here.
But even some Finns are now going on about Siberians and Ugrics because it's just the more interesting explanation.
Somoene should finaly do a proper study, sampling several Finnish areas, Europeans, and North Asians. Because this is all getting really boring.
Dienekes, seriously, this is an intra-continental perspective. It's like that second part of the Bauchet study, once the Altaians were dropped.
Intra-continental variation occurs both because of intra-continental processes, and extra-continental influences.
Intra-continental variation in Africa, for example, is due both to different isolated African populations, but also extra-African invasions (e.g. the Amharas). Or, in Asia, both to long-term differentiation between e.g. South and East Asians, but also to migrations from the West.
Similarly, in Europe, intra-continental variation is generated by exogenous factors, e.g. the greater West Asian influence in Southern Europe because of the Neolithic, or the Siberian influence in northeastern Europe.
According to the authors of the study whom my friend over HBF contacted, the Finns stood apart thanks to isolation and drift just like the NY times article addressed, not because of Siberian admixture which BTW can be detected in every European population at various rates.
That is inaccurate, in Bauchet et al. the only population with significant Asian admixture were the Finns, just as the Russians were in Li et al. Most Europeans don't have any detectible Asian admixture.
"Drift" doesn't explain why Finns deviate from Europeans in the direction of Asians. Both Bauchet et al. and the older classical studies concur that Finns have ~1/10 of Asian genes. This component, lacking in all the other studied populations in Lao et al. is what differentiates Finns.
"Drift" doesn't explain why Finns deviate from Europeans in the direction of Asians. Both Bauchet et al. and the older classical studies concur that Finns have ~1/10 of Asian genes. This component, lacking in all the other studied populations in Lao et al. is what differentiates Finns".
You sound like a broken parrot. What the authors of this 500k snp study said regarding Finns is very true indeed. Or are you suggesting that we should believe your Greek fairytales intead?
What do you mean with "significant"?
Based on Bauchet´s study we can concur that very few European population has Uralic admixture exceeding 5% indeed. Based on Cavalli´s estimations we can concur that the ancient Uralics were very close to caucasoids in terms of genome, nothing to compare to genetic distance with negroids and caucasoids f.e.
Anyway, AncestryByDna has come up with very interesting stuff about extra-european admixture rates of Europeans. I am sure we will know lot more in the future to the extend that your speculations will hardly matter much anymore.
But Dienekes, which direction is towards the Asians in this study?
There are no Asians to tell us.
We have some Hungarians, who should have something like 10% Uralic influence, going by Cavalli-Sforza old studies. But they're not pulling anywhere.
But Dienekes, which direction is towards the Asians in this study?
In PC2 of Bauchet et al. it was clear that Finns were positioned away from Europeans in the direction of Altaians.
This is also consistent with Li et al. who discovered that northern Russians were different from Europeans due to affiliation with East Asian and Native American groups.
So, there is no mystery Eastern European component differentiating Finnic or Finnic-admixed populations from other Europeans, but rather their unique East Eurasian admixture.
We have some Hungarians, who should have something like 10% Uralic influence, going by Cavalli-Sforza old studies. But they're not pulling anywhere.
Hungarians have very little East Eurasian Y-chromosomesor mtDNA. Predictably they are like other Europeans autosomally. Their Tat-C bearing Mongoloid ancestors did not seem to have a big impact.
Anyway, AncestryByDna has come up with very interesting stuff about extra-european admixture rates of Europeans.
None of the published scientific studies has shown any significant East Eurasian admixture in Europeans other than the Finns or Russians.
Li et al. which is the most recent one on the matter, found 6% East Asian admixture in Russians and 0% in all other Europeans (Adygei from the Caucasus had 2%).
Bauchet et al. found no important Altaian component in most of Europe.
The Portuguese appear to be similar to Spaniards and other Europeans. It is the Finns and Russians that appear to be most different from other Europeans in all the latest high-resolution autosomal published studies.
"The Portuguese appear to be similar to Spaniards and other Europeans. It is the Finns and Russians that appear to be most different from other Europeans in all the latest high-resolution autosomal published studies".
Spaniards have plenty of negroid-style dna as well but do not topple Portuguese. I shall also remind you of Cavalli-Sforzas estimations of the genetic differences of Negroid and Uralic-style gemones. The distance of English and Siberian Uralics is 66 (English 100) and English and Negroids 477. So 9% negroid impact will alter the structure of gemone way more than 9% Uralic, in terms of Europeannes rates obviously.
"No recent studies using large numbers of markers have found any substantial autosomal DNA from Sub-Saharan African in Europe. Even Near Eastern Arab populations such as the Palestinians who were thought (erroneously) to have Sub-Saharan admixture due to mtDNA turn out to have 0% Sub-Saharan admixture (Li et al.)"
As usually you like to play little semantic gimmicks with "significant" and "substantial" whenever it fits to your purpose. We all you know that your blog is very agenda driven. Your obvious goal is to portray Northern Europeans as something they are not. Finns are an easy target and from them you extend your little liturgies to rest of the Northern Europeans. It's all the same "blonds are not as beautiful as mediterranian" stories.
Don't blame the Northern Europeans that hollywood has casted Dago's as non-whites, you little complexed Greek.
Dienekes,
This is where the Finns deviate towards the Altaians in Bauchet.
http://img72.imageshack.us/img72/6245/10rv8.jpg
And this is just where they devaite as a result of the blue (Baltic?) markers. And it has nothing to do with Altaians, who are no longer part of the picture.
http://img181.imageshack.us/img181/2451/eurodnank8.jpg
The Lao study is basically like this second diagram, looking at intra-European markers and differenecs.
>> And this is just where they devaite as a result of the blue (Baltic?) markers.
Incorrect, they are using all markers.
>> looking at intra-European markers
Incorrect, they are looking at all the markers on the chip that passed quality-control.
PS: In Bauchet et al. it is clear that Finns are 10% of the way between Germans and Altaians on PC2 (Figure 2).
Here's a direct quote from the article. Looks like Polako, got it 100% correct.
"The map also identifies the existence of two genetic barriers within Europe. One is between the Finns (light blue, upper right) and other Europeans. It arose because the Finnish population was at one time very small and then expanded, bearing the atypical genetics of its few founders".
Congrats for Semitikes, docile people are what the world needs.
>> Here's a direct quote from the article.
Yes, the NY Times article, not the published one.
I think Dienekes may be right in the discussion with Polak, yet the two graphs show some differences: Finns are better aligned in PC2 with Armenians in the West Eurasian graph than in the global one.
Anyhow that blank space "south" of the Finns should be filled up with Eastern Europeans, would these be sampled. Russians for instance are likely to show some trend towards Siberian populations, even if they cluster much more strongly with Europeans. Just as Finns or maybe only somewhat like them (only further research will tell).
Semitikes,
What´s your stance on the Sami aboriginals of FennoScandinavia? These folk carry 25-35% of N1c, much less than Baltic-Finns.
A study that includes samples from a few new populations.
Latvians, Russians, and Ukrainians appear to be even more extreme than Finns on PC2, not intermediate between Swedes and Finns.
Excellent Study.
- Daryl Basarab
What a sorry character Mr Medienekes is lol. He´s quick to conclude an Asian ancestry for the Finns based on an uncomplete study on the genetic structure of European populations. Just like Polak wrote, once we add Slavic and Baltic populations to the equation the picture changes completely.
http://www.gnxp.com/blog/2008/08/genetic-map-of-europe-again.php
I can´t but help feeling sorry for Mr Medienekes who´s obsession and Asperger like Zwangsvorstellung on an Asian origin for the Finns is jeapardizing his reputation as a serious debater on the subject of human genetics.
based on an uncomplete study
Right, the position of a single Finn on a 2D map is the "complete study"...
Whatever the validity of your criticisms, Bobafetto, you should show some respect. Moreso when you are at Dienekes' home, so to say, and he's so liberal as not to outright delete your insulting comment (as I would have done very justifiedly).
Anyhow that kind of insulting attitude only shows your immaturity and damages your position in the discussion by putting everybody against you automatically. They won't even ponder your argument, as the first impression they get is: "woah, another troll!"
wow central italians are even more south on this map then the portugese and greeks are!! i wonder where sardinians woud be placed if they were sampled? maybe to the far south and west near portugese but more south! i wonder where southern italians woud be placed id assume more east south then north greeks and turks woud be farthest south and east if they were sampled! i hope more studies like this woud be released
Finns are pushed right because of their eastern european genes and due to the erroneous method in all studies. Mistakes in sample cropping are present, but are smaller factors.
Greater factor is the lack of Northern Russian samples because wide genome comparison with all surrounding areas around would pack it tight. Adding Batic, Russian, Northern Russian, Sweden and Russian samples will squeeze Finnish area, because of all other, less significant for Finns, sampled genes of those populations will take part in the placement calculation. All disparity and originality in the periphery of PC-map will be emphasized.
Post a Comment