September 01, 2008

Geography and Genetic structure in Europe (again)

A new letter in Nature has appeared online with an almost identical conclusion with the Current Biology paper I blogged about earlier. The new study included 3,192 individuals from POPRES, genotyped with the Affymetrix 500K chip.

The newer study included a wider sampling of populations, including Cypriots, Turks, and Eastern Slavs among others. Hence, the correspondence with the map of Europe is even stronger than before.

While in the previous study PC2 separated the Finns from the rest, the wider sampling, especially of Eastern Europeans now causes a clearer separation along the east-west axis. Note that PC2 in this study is not the same as PC2 in the previous one, as can be easily seen by e.g., (i) the fact thatPortugal and Spain are correctly placed to the west of Great Britain, and (ii) the fact that the Finnish score is about the same as that of Greeks and Yugoslavs in this study. In any case, Finland is represented by a single individual.

This underscores the often-forgotten fact that PCs are calculated from the available data and thus depend on the included populations. The inclusion of Eastern Europeans (esp. Eastern Slavs and Balts) in this study has now made the strong east-west differentiation in Europe, the most salient feature on the second PC. Unfortunately no higher PCs are presented.

From the paper:
The direction of the PC1 axis and its relative strength may reflect a special role for this geographic axis in the demographic history of Europeans (as first suggested in ref. 10). PC1 aligns north-northwest/ south-southeast (NNW/SSE, 216 degrees) and accounts for approximately twice the amount of variation as PC2 (0.30% versus 0.15%, first eigenvalue54.09, second eigenvalue52.04).

...

More robust evidence for the importance of a roughly NNW/SSE axis in Europe is that, in these same data, haplotype diversity decreases from south to north (A.A. et al., submitted).
The only deviation from geography seems to be the Slovakian individual:
There is only one obvious outlier, which is Slovakia; however Slovakia is represented in our data set by only one individual, and based on the individual’s position in PC1-PC2 space it’s possible this outlier may actually have had Italian, rather than Slovakian ancestry.
and the Russians:
The Russian Federation is less-striking as an outlier, and appears to lie too far “west” genetically, which may be a result of small sample size (n = 6) or simply that the Russians sampled here have ancestry from a location further west than the proxy location for Russia (Moscow) would suggest.

As for the Greeks (GR), once again, they are placed between Italians and their northern neighbors, with Albanians (AL), a pre-Slavic Balkan population especially close, followed by Slavomacedonians ("MK") and Bulgarians (BG). This is especially impressive given the small sample sizes (8 for Greece to 2 for Bulgaria). It appears that even individual members of ethnic groups "find their way" to the appropriate place of the map.

The Way Ahead

Both this and the previous paper have made it abundantly clear that, even in Europe, where genetic differentiation is very limited and populations are arrayed in a cline, it is possible to determine the rough geographical or ethnic origin of an unknown individual. Even for closely related groups where the precise origin can't be determined (e.g. Spanish vs. Portuguese), we can at least exclude with high probability most other European groups.

The applications of this are obvious: criminal or victim DNA can be pinpointed on the map. The ethnic origin of undocumented persons (e.g. illegal immigrants) can be ascertained with some confidence. Adoptees or persons of unknown ultimate origins (such as many inhabitants of the New World) may be able to trace their ancestry to something more than they could guess by looking at the mirror.

Indeed, while 500K markers are used in this analysis, it will turn out that fewer markers will carry most of the power of distinguishing between ethnic and geographical groups: you only have to examine 500K of them to harvest the useful ones.

This opens wide possibilities for ancestry testing; such testing previously gave one fairly obvious information ("you're 99% European") and cost a substantial amount. It will soon be possible, using a specialized panel of the most informative markers to create ancestry testing that is both affordable and informative.

The identification of ethnic differences in autosomal DNA also allows us to look at history from a new perspective, as the ethnic origins of skeletal material from the past will be ascertained with a precision unmatched by biological anthropology.

Of course, for the distant past, problems with obtaining authentic ancient DNA will remain, as well as the fact that accelerating human evolution may have changed allele frequencies or introduced new alleles into populations. Nonetheless, there is hope for real progress.

The real impediment will not be, in my opinion, technical, but rather psychological/political. The realization that not only major continental races, but also ethnic groups are biological entitites goes against the prevailing politically correct orthodoxy. According to this orthodoxy, European nations are artificial cultural constructions whose members share a "myth" of common origins; they are "constructed" products of the last few centuries; ethnic identification is a subjective notion of self-identity, rather than an objective notion of ancestry and homeland.

It now appears that while European nations are not races, they are, nonetheless, biological populations, occupying specific positions along the Caucasoid genetic continuum, and distinguishable from most other European nations, if not always their immediate neighbors.

Nature advance online publication 31 August 2008 | doi:10.1038/nature07331

Genes mirror geography within Europe

John Novembre et al.

Understanding the genetic structure of human populations is of fundamental interest to medical, forensic and anthropological sciences. Advances in high-throughput genotyping technology have markedly improved our understanding of global patterns of human genetic variation and suggest the potential to use large samples to uncover variation among closely spaced populations1, 2, 3, 4, 5. Here we characterize genetic variation in a sample of 3,000 European individuals genotyped at over half a million variable DNA sites in the human genome. Despite low average levels of genetic differentiation among Europeans, we find a close correspondence between genetic and geographic distances; indeed, a geographical map of Europe arises naturally as an efficient two-dimensional summary of genetic variation in Europeans. The results emphasize that when mapping the genetic basis of a disease phenotype, spurious associations can arise if genetic structure is not properly accounted for. In addition, the results are relevant to the prospects of genetic ancestry testing6; an individual's DNA can be used to infer their geographic origin with surprising accuracy—often to within a few hundred kilometres.

Link

43 comments:

Polak said...

The lack of Finns in this study is now a problem.

It will be interesting to finally see sizeable samples of Finns, Balts and Russians and how they affect the dynamics of such a PC plot.

And then we'll also need samples from different parts of larger countries, such as Russia and Germany, for the complete picture.

Obviously, the complete picture will look like a map of Europe, but with some interesting small surprises I'm sure.

Maju said...

This is interesting, in spite of being only available the PC1/PC2 simplistic graph. Unlike the other graph posted recently and like Bauchet's one, the PC1 is defined between NE and SE Europe (the latter clinal with Anatolia) and the PC2 is defined between Iberians and all Eastern populations.

Finns are marked but guess that's (like in the case of Livonians) a single individual sample, not enough to describe the population. But anyhow it's interesting that they appear, as expected, "north" of Eastern Europe. It is also interesting the somewhat distinct PL-RU cluster near them.

Another interesing element is Greek populations. Greeks appear to cluster with Balcanic peoples but Cypriots (also Greeks with all likehood) do not. I guess this represents variability within ethnic Greeks depending on geography (is the GR sample from Macedonia?).

A les important but nevertheless noticeable are the "erratics": a Brit and Swiss that look Sicilian, a Russian that looks Swiss/North Italian...

I can only imagine that the separate "westward" Italian cluster must mean Sardinians (always a distinct group), right?

Polak said...

Maju,

The Latvian is where the Finns were in the other study. That's because he's now the true Eastern European.

The Russians are also close to there, and they've now pulled the Poles that way due to common Slavic links.

The only Finn sampled probably had some Swedish admixture, and due to a lack of other Finns, who would most likely be closer to the Latvian and pull him that way, he's sought refuge to the closest larger sample that shared most markers with him (ie. Scandinavians).

Jim H. said...

I am of Slovak descent, yet my atDNA also matches to northern Italians for some reason.

My only theory is that my Slovak ancestors must have been Vlachs or Wallachians of Italian descent who migrated to Slovakia from Romania.

That begs the contentious question of to what extent generic Romanians are descended from Italians. On this new chart, Romanians and Italians are at least in the same quadrant.

Maju said...

@Polak: ok. I think you are probably right.

@Jim: Romanians seem to consistently cluster with other Balcanic populations, what rather suggest low Italian admixture rate. Nevertheless both Rumanians and Bulgarians appear on this graph somewhat more to the "west" than it would be expected by mere geography, so maybe you have a point after all.

But your clustering with Northern Italians could be a "erratic", like the Brit that clusters with Sicilians. It could also mean a similar kind of admixture as most North Italians, whom I don't think make up any separate cluster of their own.

Peter said...

"The only Finn sampled probably had some Swedish admixture, and due to a lack of other Finns, who would most likely be closer to the Latvian and pull him that way, he's sought refuge to the closest larger sample that shared most markers with him (ie. Scandinavians)".

The things is that usually in these studies the Finnish sample comes from the North-East Kuusamo isolate, an isolate within an isolate. This is a tiny Finnish sub-population, most Finns, 75-80% reside in South- and SW parts of Finland which have been subjected to extensive Scandinavian input already from iron age. The recent Swedish admixture is geographically concentrated only to specific regions, although heavily populated ones.

Polak said...

^^ Well anyway, my point was that these PC plots are heavily influeced by who appears on them.

This Finn has to cluster somewhere, and minus other Finns, he's going to cluster closest to what he's most similar to.

Btw, interesting to note the Italian cluster in the far southwest, which by some people's logic here, is probably made up of Roman era African slaves. Haha.

However, in reality, these PC plots also reflect genetic isolation relative to the samples present. So if these Italians come from a village where marrying cousins has been commonplace for centuries, it might now appear like they come from Mars...but in fact, they don't.

Jim H. said...

maju,

"Similar kind of admixture as most north Italians" is an astute observation. The village where my oldest known Slovak ancestors are from was founded by Germans and settled in part by Germans from the Eisack Valley of the South Tirol. Maybe they were little different from northern Italians, genetically.

Peter said...

"This Finn has to cluster somewhere, and minus other Finns, he's going to cluster closest to what he's most similar to".

Interestingly,this Finnish individual is apparently closer to Germans, Dutch and Swedes than Poles (and Russians) are to Germans, Ducth and Swedes. So what´s the fuss with geography?

Polak said...

^^ It's a generalization. And one sample either way doesn't break up the trend too much.

Funnily enough, if they sampled only one of the far right Poles here, everyone would be screaming "Asians", but if they'd only sampled the Pole to the far left, just below the Finn here, they'd all go "Poles are Germans speaking Polish".

In reality, that area where Poles and Germans meet would be totally filled in if samples of 50 or so were taken from North Germany, East Germany, Western Poland and Northern Poland. And if southern Poland was sampled, we'd be overlaping fully with Czechs and Hungarians, etc.

Polak said...

Whoops, actually I can see two Polish samples further west than the one Finn.

I guess we win. LOL

miz RAND BLOWTON said...

Are they gonna use any of these tests on real people? All this talk ,but are they gonna use it on folks to give them a definite Racial/and ethnic Label and put it on an I.D. Card?-just so it isn't easily forgotten after all the trouble one went to, to I.D. a person.But I'm wouldn't use it to limit the person-maybe democracy should hopefully protect that.But on the contrary,it's a perfect honor to be identified as a specific thing ,even if your whole group isn't wonderful like you are.

Crimson Guard said...

Some of the French samples that are intermingled with the Italian ones, could be Southern French and Corsican(which was only relatively recently become "French"). The Great Britain one thats near the Sicilians, could be the Welsh or perhaps the Picts maybe.

The Sardinians always seem to be the odd men out, since they are so far South-West there, but somewhat nearer to the Iberians..this could be also since they received alotve Iberian immigrants during the last 500 years or so. The Balkans and Greece were occupied by Roman Italy for something close to a 2,000 years, same with the Swiss, so seeing some closeness isnt surprising.

The Cypriots seem(despite being distant yet) to be closer to the Italians and to that British group, than Greeks or even Turks(who are supposedly something like 95% Greek or whatever). Even the Turks dont appear very Greek in this study and are clearly separated.

Maju said...

@MRB: hopefully you are just being sarcastic.

It's some sort of archaeological curiosity. Bio-archaeology. Sure there's always the odd (or maybe not so odd) wannabe neonazi lurking around... but for most people this is about history: the demic side of (pre-)history.

Maju said...

@Crimson:

The Sardinians (...) somewhat nearer to the Iberians..this could be also since they received alotve Iberian immigrants during the last 500 years or so.

Not probably. Remember that Sardinians are quite odd also via Y-DNA (and anyhow there was no such Iberian immigration in Sardinia at any time, the Catalan-speaking town of Algher excepted maybe). They don't seem to be product of any recent input. They may be a "frozen" Italian Aboriginal and/or West Balcanic Neolithic product.

They do not seem to lean towards Iberians specially (they do not cluster with them at all) but towards the active side of PC2, towards the Western component in general.

Notice that they may be well dominated by a distinct component of their own, maybe specifically Sardinian or generically Italian but, if this component is low overall, it won't influence the PC graph, no matter how important is it locally.

That's why a deep K-means clustering analysis is necessary in order to understand well the real structure. The deeper, the more regionally unique components we should see (the more shallow the more we will see the widespread components only, no matter they are minoritary for most populations).

Ebizur said...

maju said,

"Not probably. Remember that Sardinians are quite odd also via Y-DNA (and anyhow there was no such Iberian immigration in Sardinia at any time, the Catalan-speaking town of Algher excepted maybe). They don't seem to be product of any recent input. They may be a "frozen" Italian Aboriginal and/or West Balcanic Neolithic product.

They do not seem to lean towards Iberians specially (they do not cluster with them at all) but towards the active side of PC2, towards the Western component in general.

Notice that they may be well dominated by a distinct component of their own, maybe specifically Sardinian or generically Italian but, if this component is low overall, it won't influence the PC graph, no matter how important is it locally."

Maju, it is time for you to wake up and see the light. The Sardinians obviously do not show any particular relationship to Italians; if they did, we wouldn't be able to tell at a glance who the Sardinians in this study's "IT" (Italy) sample were.

On the contrary, it is quite obvious that the Spanish and Portuguese are genetically intermediate between the French on the "West European" side and the Sardinians on the "Paleo-Iberian" side.

miz RAND BLOWTON said...

Actually maybe all groups are bossy,Scots are-you gotta wear plaid near them and do everything they want,and they aren't minorities.Red headIrish people act as if all potatoes belong to them(I know all the poison ones thay killed them did),and believe me they can keep those.As soon as I eat a potato here they come,but before that I'm notRedHeadIrish ,only after I eat a potato.I'd like take a baseball bat to that RedHead of theirs,because they aren't friendly unless they see a potato-then your butt is theirs.Your RedPotato Butt,that is.

miz RAND BLOWTON said...

I'm not really being mean-it's just an observation.

miz RAND BLOWTON said...

Oh an another thing -this usually happens because there is no one else around,so all there are are Poles,Jews,Negros,Italians,Latinos,and Asians -I really don't know anybody else.And if there are one or two of your people they usually don't have or own anything or they're real stupid acting.So you almost have to side with one of the other ones. I guess my ethnicity is the Jerk race...I'm last Martian on earth

Crimson Guard said...

Well from what I understand Catalan immigration into Sardinia was pretty extensive during the Argonese/Spanish rule. Some towns or cities were completely repopulated by Spanish people.

Anyway, according to archeology and history, Sardinia should be more Italian, settled from people from mainland Italy, with some input via the Achaeans possibly or some other Eastern Euro-Mediterranean people(Etruscans?). Their Y Chromosomes should be somewhat Anatolian or even closer to Scandinavia or Balkan( Y Chromosome I)..but they never cluster near any of these other groups as far as I remember.

They always seem to be distinct from other Italians and others they should be technically closer too(like say the Corsicans), but arent. They namely seem closer to modern Iberians for some reason.

I suppose a more modern explaniantion has to be Spanish Dominance of the island, probably altered the island from the one antiquity. It should be remembered also despite being the 2nd largest island in the Mediterranean, its population is less than 2 million people, compared to the largest island in the Mediterranean, Sicily which has a population of 5 million.

PS

The Iberians are neatly separated from the British peoples, which I think undermines the theory they were populated by paleo-Iberians or such. British seem closer to some of the French, possibly because of the Gauls.

Maju said...

Maju, it is time for you to wake up and see the light. The Sardinians obviously do not show any particular relationship to Italians; if they did, we wouldn't be able to tell at a glance who the Sardinians in this study's "IT" (Italy) sample were.

On the contrary, it is quite obvious that the Spanish and Portuguese are genetically intermediate between the French on the "West European" side and the Sardinians on the "Paleo-Iberian" side.


Uh?

Let's see: Sardinia was colonized first of all by Neolithic colonists of the Cardium Pottery culture. Agreed? These were either UP Italians or West Balcanic Neolithic peoples or maybe even from further East. This is probably the only significative colonization of the island, specially the interior "refuge".

There migh have been an Iberian penetration with the Nuraghi but that's all.

And, by the way, we do not know for sure if that IT cluster are Sardinians, at least I don't. I was the first to suggest it and the rest followed suit. Maybe someone with access to the paper could clarify because there is no specific tag on that cluster in the graph (Scotland or Sicily are tagged instead).

Maju said...

@Crimson: I may be missing something but there is only one town where Catalan is spoken in Sardinia (Algher or maybe Alghero). I do not exclude there may have been some further settlement in other coastal spots but surely not more than other colonial inputs, like Carthaginian or Roman.

Corsica seems to have a different population history. Not sure now about the details (just vague memories right now).

Agreed re. Italian and/or Balcanic origins of Neolithic colonization but notice that there may have been important founder effects (they are pretty obvious in the Y-DNA).

We do not even know for sure if that cluster is Sardinian (anyone?) but in any case the appearence of it leaning towards the "Iberian" end of the PC2 axis may be a mere mask caused by a locally minor component. It happens too often.

The Iberians are neatly separated from the British peoples, which I think undermines the theory they were populated by paleo-Iberians or such. British seem closer to some of the French, possibly because of the Gauls.

Almost agree. Since it became obvious that British and Irish mtDNA was closer to Mid-Western Europe than anywhere else, it became very apparent that the Iberian or Basque hypothesis were at fault.

But I would not blame the Gauls as much as the older Paleolithic component from the same area, specially because in Y-DNA the islands appear much more "conservative" (R1b) than the continent (because they suffered much lesser impact from more recent immigration, be it Neolithic or Indoeuropean).

Anonymous said...

@ Jim h

Valacs lived in east Moravia up to 1600 preserving their customs and their language, before to be defeated by the armies of the Ausburg and the loss of their national identity. I don't know how much the colonization had been strong from Italy, in the age of the roman empire, but perhaps it was more strong I thought. Commercial activity probably attracted a lot of people in the zone. Extraordinary as the cultivation of the grapevine is maintained up to our days.

Crimson Guard said...

Yeah, something doesnt make sense. As its not going well with history altogether as we know.

I was just looking at these charts, according to this one, Sardinians on the Paternal(Y Chromosome)side, are closer to the Spanish:

http://racialreality.110mb.com/sicily/index_files/ychromosomes.jpg

Nearly same story with the maternal(mtDNA), although not much differences can be seen.

http://racialreality.110mb.com/sicily/index_files/mtdna.jpg

Autosomes, shows that Sicilians are closer to the Central Spanish and even French, while the Sardinians are more Eastern Mediterranean but still distant from the other groups.

http://racialreality.110mb.com/sicily/index_files/autosomes.jpg

Alternatively, on these ones from 2005 & 2007, show that Y chromosomes, Sicilians are closer to West Italians and Hungarians, while mtDNA, Sardinians are close to North-East Spain.

http://dienekes.110mb.com/articles/fallmerayer/mtdna.jpg

http://dienekes.110mb.com/articles/fallmerayer/ystr.gif


Overall though, Sardinians are a bad bad population to study for some unexplained reason according to geneticists, subject to genetic drift or whatever it is. As they almost always autosomally end up being distinct.

UncleTomRuckusInGoodWhiteWorld said...

"The applications of this are obvious: criminal or victim DNA can be pinpointed on the map. The ethnic origin of undocumented persons (e.g. illegal immigrants) can be ascertained with some confidence."


LOL...this doesn't work well in a multiethnic society or one that had considerable past immigration...this such as the UK, Scotland, France, even the Netherlands.

I would assume even in certain border areas of Greece this would be sketchy.

Bismarck said...

Dear Dienekes,

I posted a comment on an ancient blog:

http://dienekes.blogspot.com/2005/08/haplogroup-frequency-correlations-in.html

I would be very grateful for any comments you would have time to make. Thank you very much in advance,

Erik Nelson

Polak said...

In regards to the British outliers...

The one way down south probably has Maltese ancestry, while those near the Poles probably have Polish grandfathers...soldiers or pilots from WWII.

And btw, being clearly seperated on these plots doesn't mean no shared ancestry or even not originating from the same population. It might just mean breeding isolation for along time.

Oh, and I think it's painfully obvious now that these PC plots react to who's on them, and in what strength. But unfortuanetly, few people understand this.

Maju said...

@Crimson: Error 403. Your files (graphs) are not accesible because you account lacks of index page (or so it says).

The similitude between Italian and Spanish mtDNA must be due to the fact that they are all dominated by haplogroup H. But that's way too common in Europe.

As for the Y-DNA, Sardinians are notable for the dominance of two rare clades: I2a2 (M26) and R1b1a. It connects specifically to nowhere AFAIK, except that I2a (mostly as I2a1) is dominant in the Western Balcans, where Cardium Pottery originated or at least "coalesced" before expansion, what strongly suggests that there was some sort of founder effect with the Neolithic colonization.

Any similitude in autosomal PC graphs is probably due to locally minor components. In fact, in the graph at this post, that cluster we suspect Sardinian looks extreme in both PCs: as southeastern as Turks or most Italians (PC1) and as western as Spaniards (PC2). This is probably only because the third vertice, the NE component of PC1 is almost totally lacking, significatively more than among Spaniards, who are almost neutral on PC1.

The graph actually says nothing more than that. If most populations would be 80% locally specific and 20% of the most widely distributed components, they would still get positioned in that graph, sometimes extremely, even if they are actually 80% out of it.

Maju said...

Oh, and I think it's painfully obvious now that these PC plots react to who's on them, and in what strength. But unfortuanetly, few people understand this.

Exactly! It maps the most widely distributed ancestries, not the most important ones locally. The ones that weight the most at continental level, not in each population.

For example this graph surely says little about the Irish, who are actually (as per Bauchet's K-means clustering) dominated by a Central/North European component that is not probably weighted in this case. Their position is surely determined by the peculiar balance among them of three "foreign" components: Northeastern, Northwestern and Southwestern, which are all minoritary among them. It also affects the same way to other populations like Poles, who are also (per Bauchet'07) dominated by the NW component.

Instead in the previous Euro PC graph posted by Dienekes some days ago, it was the NW component the one highlighed in PC2, what caused Spaniards to cluster with Italians, as the SW component, dominant among the former, was not weighting.

Peter said...

"Whoops, actually I can see two Polish samples further west than the one Finn.

I guess we win. LOL"

Well if the direction of Hungarians and Checks constitute "west" to you, then maybe...

Polak said...

West is left here along PC2. No mystery about it.

Anonymous said...

About the Sardinians you have to consider that it is a people that lived isolated for many centuries. It is people with a strong cultural identity and their mixture with others it is very low. The foreigners always settled only in few cities on the coasts, while the Sardinians have always lived inside the island, for this they has developed very special genetic characteristics.

just passing by said...

What DNA is the paper referring to? I assume it is either mtDNA or Y-DNA. I doubt that this would be a sampling of autosomal DNA.

Only one person sampled in Slovakia? It used to be part of the Austria-Hungary Empire, which included northern Italy in its heyday. So obviously more samples from Slovakia are needed.

Slavs spread eastward into what is now Russia from eastern Europe. So it shouldn't be surprising that Russians have a western bias, unless the samples come from a prehistoric Finnish strain that was there first.

R1a1 trends in a southeast - northwest axis with Norway at the northwest.

McG said...

I must be missing something here in this paper and the recent one reporting similar data. We are looking at the analysis of autosomal data. What is this supposed to tell us? I understand the Y Chromosome and I believe I can trace its evolution as it has had STR mutations. I still happen to believe that the Irish/Scottish tribes all can trace their origins to Iberia?? This presentation is not disputing that(?) but is presenting the present constituency of different regions of Europe in a different manner. To me it is saying that after a period of time, maybe hundreds or even thousands of years, the autosomal characteristics of the inhabitants tend to "mix up" such that they all appear similar to each other - "by this measurement". Is this primarily female driven by mitochondrial DNA transmission down the line?? Certainly, different "races", tend to physically resemble each other? Is that what this study is saying??

I would like a little explanation here of what this study is basically saying? thanks in advance

Dienekes said...

Slavs spread eastward into what is now Russia from eastern Europe. So it shouldn't be surprising that Russians have a western bias

That is a very good point.

McG said...

One other comment on the Finns. My wife is full-blooded Finn. All four grandparents born in Finland. She has some of the problems pure(?) Finns have; among which is a type of anemia which affects about 15% of the population. Additionally, her blood type is B-, which is fairly rare in western europe.

I don't have the reference but I read somewhere that a significant part of the male population was wiped out at one time (probably by the Tatars?); It was a linguistic paper and pointed out that the children of the invaders were raised by the Finnish mothers and no language discontinuity was observed! I believe there are two basic ethnic groups in Finland, the Taditch (sp), and the Swede-Finns. I would expect their Y chromosome should reflect some of this???

Maju said...

What DNA is the paper referring to? I assume it is either mtDNA or Y-DNA. I doubt that this would be a sampling of autosomal DNA.

It is autosomal DNA, that is what "genotyped with the Affymetrix 500K chip" means. It's one of several standarized systems of testing nuclear DNA. Would it be haploid DNA, they would be looking for defining SNPs along genealogical branches.

Slavs spread eastward into what is now Russia from eastern Europe. So it shouldn't be surprising that Russians have a western bias...

They do not in that graph. They are the "easternmost" plural (+/- representative) sample in their quadrant.

Slavs spawned from Ukraine, it seems now.

...

We are looking at the analysis of autosomal data. What is this supposed to tell us?

Overall genetic affinities. It makes no intention of assuring wether these are neutral or adaptative markers. And logically it tends to become homogeneous over time in physical and culturally contiguous regions (due to recombination, "reshuffling", with every generation).

I still happen to believe that the Irish/Scottish tribes all can trace their origins to Iberia?? This presentation is not disputing that(?) but is presenting the present constituency of different regions of Europe in a different manner.

This presentation may be missing much locally important stuff. They should only plot populations for which the PC+PC2>50% (i.e. clearly meaningful), but that would leave the graph nearly blank.

But from other studies (again Bauchet'07) it does not seem that Irish cluster better with Iberians than with Central/North Europeans. What strongly suggest (specially Ireland being an island, and hence somewhat isolated) that the main component of Irish (as well as Brits) is from Central Europe. They do show some relevant Iberian or Basque components anyhow but not enough to change the previous conclussion.

This doesn't mean that maybe at sufficient depth of analysis an Irish-specific or even maybe "Atlantic" component would not appear. It is possible but so far it has not been detected.

Personally I understand that these homogeneity sets (as seen by K-means clustering) reflect the Late UP/Epipaleolithic reality and therefore I would rather expect an Italian-specific component much before an Irish-specific one.

I don't have the reference but I read somewhere that a significant part of the male population was wiped out at one time (probably by the Tatars?);

That surely means that Finns have a strong founder effect, specially on the male side, most of whose lineages appear to be ultimately original from Siberia (haplogropup N). Nothing to do with Tatars anyhow.

Polak said...
This comment has been removed by the author.
Polak said...

"Slavs spawned from Ukraine, it seems now."

Most likely.

But it's hard to make firm conclusions based on a PC plot with such small samples for Eastern Europe.

IMO when more Finnic and Baltic samples are included, the Russians will break away from the Poles.

eurologist said...

Given how these graphs react to extremes, I think one could learn a lot more by doing geographically isolated groups and then removing certain principal components.

Say, you are interested in figuring out central European relationships and the impact of either including or excluding a slav component. In that case, I would first look at what signifies Iberian (by concentrating on, say, Spain, Portugal, southern France, western Switzerland, and north-western Italy), and then subtracting that component from all data. Likewise with Sami, Turkish, and southern Mediterranean components, if necessary. Then, with what is left (and, hopefully, more data points from south-east central Europe), it would be easier to see, e.g., if there is more than one important central European component (besides a putative post glacial spread from southern France, and later Slavic migrations).

Similarly, other questions could be better addressed by first excluding minority outlier components.

McG said...

I read a comment on rootsweb by Ann Cousins re: this work. She states that this is a very early chip, nothing like DeCodeme for example. It does include a very small number of Y SNP's though. Newer packages contain more Y SNP's and are more tailored to popgen questions. It looks like we will be seeing more papers of this ilk in the future as researchers learn to use the larger chips and can tailor them to specific applications/questions?
Health vice popgen will continue to be emphasized because that's where the funding is.

Mauri said...

"The only Finn sampled probably had some Swedish admixture, and due to a lack of other Finns, who would most likely be closer to the Latvian and pull him that way, he's sought refuge to the closest larger sample that shared most markers with him (ie. Scandinavians)".

Not necessarily. Finns have main two genetic heredity. This issue is very poor understood generally. To be pedantic the baltian heredity has represented only by one group from three, the other are eastern and western.

There is one big confusion in many researches and the eastern part has much gained, because all researches uses governemental distribution instead of trying to find real geographic results for genetic distribution. As some earlier said for instance line from west to east, to almost pure eastern areas as Kuusamo and Kainuu have count to the west. This is only result from indolence and use big county hospitals in collecting samples.

Mauri said...

MCG; "One other comment on the Finns. My wife is full-blooded Finn. All four grandparents born in Finland. She has some of the problems pure(?) Finns have; among which is a type of anemia which affects about 15% of the population. Additionally, her blood type is B-, which is fairly rare in western europe.

I don't have the reference but I read somewhere that a significant part of the male population was wiped out at one time (probably by the Tatars?); It was a linguistic paper and pointed out that the children of the invaders were raised by the Finnish mothers and no language discontinuity was observed! I believe there are two basic ethnic groups in Finland, the Taditch (sp), and the Swede-Finns. I would expect their Y chromosome should reflect some of this???"

I am Finn and never heard about this anemia.

About one third of Finns have HG I1 and almost two parts have N1. The most widespread Mitochondrial Hg is H.

Male Finns genetic age is about 2000-2500 years, calculated by TMRCA value. I think it is result from the colonization of Finland. When speaking about Finn's little genetic variance we look autosomal dna and by it we cannot define the population age. Younger geographic population may have more genetic variance than older one, depemding on how big was the population in the begining.