The figure above comes from the an article in The New York Times, The Genetic Map of Europe, which draws from a new paper, Correlation between Genetic and Geographic Structure in Europe. The authors sampled 2,500 Europeans across 300,000 points of genetic variation, then extracted out the components of that variation, and plotted the individual data points along the two largest independent dimensions. You note that various samples tend to cluster geographically with each other; i.e., Finns tend to cluster with other Finns, Italians with Italians. This makes sense since Europe hasn’t been a random mating population, most people found mates from local regions. Sandman, Genetic Future and Dienekes have extensive comments so I’ll leave it at that. But, below the fold I’ve taken a less stylized figure, which shows all the individuals sampled as points, and added some labels to give you a better geographical intuition.
I predict that deeper analyses of other regions will result in the same trend; geography determines genetics…most of the time. But, I think the the exceptions, such as the Finns, will warrant closer examination and are interesting.
Razib Khan’s degrees are in biochemistry and biology. He has blogged about genetics since 2002, previously worked in software development, is an Unz Foundation Junior Fellow and lives in the western US. He loves habaneros.
August 14th, 2008 at 7:06 am
The only surprising feature of this map is that Italy overlaps so strongly with Portugal (who’d have thunk?) That’s the only major deviation between the graph and the geographic map of Europe.
I can’t see any clustering that cannot be accounted for by sampling effects (e.g. “France” is really the city of Lyons – no wonder FR looks like a small extension of CH!). There is a lot of variation in there, but it seems to be entirely clinal: constant gradation rather than deep, border-following chiasms.
August 14th, 2008 at 7:11 am
The Finns are an outlier so we can disregard them… just kidding! (So Suomi if you can’t take a joke!).
I don’t see how your comments are less extensive than the links you point to – all of you seem to blog about the same amount on this. You’re too modest, R!
I notice some are marked 1 & 2 (e.g. italy 1 & Italy 2). What’s up with that? In Italy’s case, is one of them actually Sicily? Is Spain 2, for example, the Basque region? Or doesn’t it work that way?
August 14th, 2008 at 8:01 am
Actually, I’d be surprised if Finland turns out to be a true outlier. Russia, Latvia, Lithuania, and Estonia don’t appear to have been sampled. I’d say that the position of the Finland samples on that graph lets us make a pretty good prediction of where those populations would most likely graph out.
August 14th, 2008 at 9:19 am
We Irish (for own own reasons) have mythically seen ourselves as “Iberians” to put a distance between us and the Protestant/ Germanic English & Scandinavians. However, looking at the map, Ireland (IE) appears an extension of the British/ Norse/ Swedish/ Dutch north-eastern peoples of Europe.
Its a case of where we might expect an outlier but don’t find it.
However, that may point up some limitations of the survey if (as it seems) the genes sampled came from people on the east coast (around Dublin). Dublin was a Scandinavian foundation (10th century) & the east cost was extensively colonised from England in the 12th century and again by a new wave in the 17th century. Conversely, millions of Irish emigrated to England in the 19th century, more than between any other two European countries, I would reckon.
So where are the ancient Celtic genes? One recent theory is that Ireland was never genetically Celtic, only culturally. We may even be related to the Basques.
The point is that this is a very broad (but still fascinating!) map, there may be nuances of great interest and significance in the details within each country.
August 14th, 2008 at 10:53 am
Sorry Toby, you Irish are just as plain and boring as us Dutch. No mythical pas for you indeed
And no relation with the Barques either
If I’m correct, a big chunk of the Spanish population is derived from Basques or their Basque-like Iberian cousins (that’s perhaps ES2 in the graph?). No connection with the Irish at all
August 14th, 2008 at 1:22 pm
I read somewhere that the biggest contributors to the UK gene pool were from the Iberian Penninsula.
August 14th, 2008 at 1:24 pm
I read somewhere that the biggest contributors to the UK gene pool were from the Iberian Penninsula.
some Y lineages, the male line, show this effect. but most of the new studies coming out don’t seem to imply this. since the british population is a compound, more or less, that does not negate the possibility. the specific work showed a relationship between welsh and basques as possible autochthonous peoples.
August 14th, 2008 at 1:37 pm
Well, there goes the Basque option
.
I know of a BBC doctumentary which found a concentration of Scandanavian genes in north-east England – no surprise since this was the region taken over temporarily by the Danes in the 9th century and called the Danelaw.
Disappontingly for Dublin (originally a Scandanavian city) there were few Norse genes left, also very few in the rest of the country – most of Ireland’s coastal cities are Norse originally.
A study to track “Celtic” genes in Ireland, Scotland, Wales, Brittany and other Celtic realms like Galicia in Spain & Wallonia in Belgium would be interesting. However, we know that the Highlands were re-populated from Ireland sometime in the Dark Ages, much the same time as there was an influx of Welsh into Brittany. All these regions have been massive migration and conquests since before the modern era, so we may not learn much.
August 14th, 2008 at 1:52 pm
Toby,
Disappointingly for Dublin (originally a Scandanavian city) there were few Norse genes left, also very few in the rest of the country
Remember that most of the Dublin Vikings and many of their Irish confederates were driven across the Irish Sea.
The area of Ireland that contains the highest Viking contribution is actually the 4 counties bordering Upper and Lower Lough Erne, here – unlike the many coastal trading and slaving settlements – were farming settlements.
August 14th, 2008 at 2:29 pm
Aw, come on Toby, if “Ireland (IE) appears an extension of the British/ Norse/ Swedish/ Dutch north-eastern peoples of Europe” that makes you different from us Scots who are probably an extension of the British/ Norse/ Swedish/ Dutch north-western peoples of Europe.
August 14th, 2008 at 3:35 pm
What I don’t like about the top graphic is that it’s incorrect in the overlap of UK and Irish, as samples from Ireland are the most extreme left, not the UK.
What I’d like to know is the ancestry of that UK sample that is isolated below the main UK cluster and to the far left of the French?
August 14th, 2008 at 3:55 pm
What I’d like to know is the ancestry of that UK sample that is isolated below the main UK cluster and to the far left of the French?
welsh?
August 14th, 2008 at 4:48 pm
Yeah, I thought Welsh or maybe Pict?
August 14th, 2008 at 5:44 pm
I don’t see why it is surprising that Italy and Portugal overlap. My guess would be that the genetic variation would have almost a one-to-one correlation with the language variation, as people took both their genes and their language with them when they moved.
August 14th, 2008 at 6:15 pm
Scott: In general, cultural domination and language replacement do not automatically imply massive population transfer or significant gene flow. That’s why the Turks are not Mongoloids. The Romans killed and deported a lot of Iberians, and there were quite a few Italian settlers, but did it really cause a massive gene flow? I’m really not sure.
One possible interpretation: for the Med, the East-West gradient is mostly caused by the Balkanic component. Non-balkanic mediterraneans did not receive it (except for some limited migration to Italy), and thus are compressed along the EW axis of the graph. Once again reckless speculation saves the day!
August 14th, 2008 at 6:24 pm
“reckless speculation saves the day!”: you should chance your arm at Climate Science, Toto.
August 14th, 2008 at 8:22 pm
My guess would be that the genetic variation would have almost a one-to-one correlation with the language variation, as people took both their genes and their language with them when they moved.
no. r is way less than 1.0. you should read some more.
August 15th, 2008 at 12:24 am
“My guess would be that the genetic variation would have almost a one-to-one correlation with the language variation, as people took both their genes and their language with them when they moved.”
The closest language related to Hungarian is Finnish. Both are non Indoeuropean languages, however according to this map, while the Finns remain unrelated to the rest of Europe, Hungary seems mixed in very well with the Germans and Slavs surrounding it.
August 15th, 2008 at 8:20 am
The methods that are used to come up with words like “closely related” or “next closest” are not the same for genetics and languages. Similar, but not close enough to make sense of earlier phases of diversification. (Or, more accurately, there will appear to be some sense as wall as anomalies and telling the spurious from the real is more of an art than a science. And we are hoping for science.)
Razib: I like the subtle LOL Cat-like modification to the graphic!!!!
August 15th, 2008 at 8:49 am
The closest language related to Hungarian is Finnish. Both are non Indoeuropean languages, however according to this map, while the Finns remain unrelated to the rest of Europe, Hungary seems mixed in very well with the Germans and Slavs surrounding it.
Khanty and Mansi, which are Siberian languages, are more closely related to Hungarian than Finnish. Hungarians and Finns are probably genetically closer to each other than either is to Khantys or Mansis, though.
August 15th, 2008 at 1:21 pm
I am not a linguist, but IMHO Finnish language is much closer to Estonian than Hungarian (at least the vocabulary is much closer). Finns are often able to understand Estonian language, but not Hungarian. Estonia is also much closer geographically. It would be interesting to see how closely Finns and Estonians cluster genetically.
Maybe the Estonian Genome project/Biobank is soon able to provide similar information of Estonians.
August 17th, 2008 at 2:12 pm
Until quite lately one fourth of the Finns spoke Swedish. We thought they had common ancestors with the Swedes but apparently they didn’t after all. From the 1930s to 70s some 500 000 Finns emmigrated to Sweden but apparently they didn’t stay.
August 18th, 2008 at 12:18 am
Voi vittu! eiks me ollakaan germaaneja.
August 19th, 2008 at 7:50 am
Is it a scientific work or it’s a student MSc work ?
Why the whole Yu are the same? Where are the Slovakian?
And at last why there are nothing about Bulgarian, they have been here in Europe as a country more then 1300 years ?
August 19th, 2008 at 9:50 am
Until quite lately one fourth of the Finns spoke Swedish.
Exaggeration.
We thought they had common ancestors with the Swedes but apparently they didn’t after all. From the 1930s to 70s some 500 000 Finns emmigrated to Sweden but apparently they didn’t stay.
Those “Finns” who moved to Sweden were largely those Swedes who didn’t wish to live in independent Finland and identify as Finns. That migration was actually a big reversal of genetic links between Finland and Sweden, not the other way around.
August 20th, 2008 at 5:56 pm
Exaggeration.
You’re right about that, but this is also an exaggeration:
Those “Finns” who moved to Sweden were largely those Swedes who didn’t wish to live in independent Finland and identify as Finns.
There are still hundreds of thousands of Finnish-speaking immigrants and their descendants in Sweden. Besides, the immigration was strongest in the ’70s, so why would the “Swedes” wait that long after independence to move out…
And, there were earlier waves of immigration. It would be interesting to see genetic study of the areas where the Forest Finns lived.
August 24th, 2008 at 6:37 pm
I recall that during the Migration Period both what is now Galicia and Northern Portugal and Northern Italy were settled by the Germanic tribe Suebi. Maybe that partly explains the similarities between Italians and Portuguese.
I wonder whether the sample somewhat isolated from the rest of the Romanians are the Hungarians of Transylvania? At least they seem to overlap with the Hungarians.
As a Finn, I now feel very exotic. Though it would be interesting to see, how the people to South and East from us would place in that map. And it would also be interesting to see the placement of the Sami people separately.