May 05, 2009

Supplement on "Geographical structure and differential natural selection amongst North European populations" (McEvoy et al. 2009)

From the supplemental material of a paper I covered in March, here are a couple of PCA plots of the first two principal components of the studied populations, with or without the Finns.


In the plot without the Finns, we see the expected British Isles -> Continental Europe differentiation in the order of Ireland, UK, Netherlands, along PC1. Swedes, and to a much lesser extent Danes deviate from this gradient in an orthogonal direction.
When Finns are included, PC1 now captures the major difference between them and the other Celto-Germanic populations which appear strikingly homogeneous along this component. The reason for the Swedes' divergence is now clear, as they are seemingly drawn towards the Finns, although the two clusters can be cleanly separated by a line at around PC1=-0.03.

It is fairly clear by now, that in northern Europe, there are two major distinctions (in that order): (i) between the Finns, and Finno-Ugrian influenced populations on the one hand, and the rest, and (b) a less important West-East gradient from Ireland to the Baltic.

The fact that factor (i) is the most important one pretty much vindicates the views of traditional physical anthropology since the time of Deniker at least. Despite the lack of data and statistical knowledge available at his time, Deniker, in the late 19th century, divided the light-pigmented northern European xanthochrooi of earlier classifications into two: the race nordique, associated primarily with the Germanic peoples, and the race orientale associated primarily with the eastern Slavs and Finns.

This classification scheme was continued by the better writers that followed, e.g., as razza nordica and razza baltica by Renato Biasutti, and as Атланто-балтийская раса (Atlanto-Baltic race) and Беломорско-балтийская раса (White Sea-Baltic race) in works written in Russian.

78 comments:

mikej2 said...

I want to see a picture including Russians, especially northern and northwestern Russians above Moskova. This surely will lead th Finns to shrink towards Sweden with same manner as Swedes shrink in the second picture by Finns. This play by PC-pictures without sensible undesrtanding about the cohesion of different historic tribes is simply ponderous :(

Polak said...

I can say right now that Poles won't ever cluster with any "Finnic" population except maybe the Estonians, who share a lot of ancestry via the R1a1 route.

Btw, Mike, the Finns won't be shrinking towards Sweden or anyone. They'll stay in their own space.

mikej2 said...

Polak, have you any arguments? I found my arguments to the creating method of the principal component pictures. It will be same and repeating with all data.

Polak said...

Yes, I have very strong arguments (ie. genetic distances) based on large datasets.

Sweden to Russia = 0.0025 Fst (Heath et al. 2008)

Sweden to Finland = 0.00437 Fst (the study featured in this blog).

Almost twice the distance, and that's because the larger Swedish and Russian populations preserved more of that North Euro diversity, which means their haplotype frequencies aren't extremely different. But because of founder effect and drift, the small Finnish population looks very different, as it only carries a small, and a bit skewed, part of the variaton seen in Swedes and Russians. So no, Finns won't ever cluster really close with anyone on PCAs, and their Fst scores are always gonna be huge.

Not that this means much in terms of Swedish admix in Western Finland, or old skool anthropology. That's why Dienekes' post is a bit strange IMO.

But anway, plz also refer to the PCAs, genetic distance tables and phylogenetic trees here...lots of info, and I'll be adding more.

http://eurogenes.blogspot.com

mikej2 said...

But you use now a self-evidencing proof, begging the question. When I say that Finns would be less genetic variation in PC-charts, just like Swedes in Dienekes' example, if the Northern Russians have been taken into the sample data, you offer me just this as proof!

Polak said...

Well I didn't write the report that found an O.0025 Fst difference between Swedes and Russians did I?

The same data goes into making the PCAs as the Fst distances.

mikej2 said...

I pointed to the another argument you used.

But you didn't mention wherefrom the Russian samples have been collected. You know, Russian is big country and all Russians are not Slavs.

Ponto said...

Xanthochroid was used by T.H Huxley, did not know Deniker used the term.

It would be extremely unusual if the Swedish population did not move genetically towards the Finns considering their geographic nearness, the presence of Saami, and the Swedish "colonisation" of Finland.

There are no slavic speaking European people in either of the two diagrams posted by Dienekes in his blog. Let us see where the Poles, Russians and other northern slavic speakers are located on the same diagrams, otherwise we are comparing apples with pears.

The big question is where did the Northern peoples come from, if not from those refuges every geneticist waffles on about? Mesolithic Europeans did not look like most modern Europeans.

Polak said...

Russians always seem to pull us Poles to the right, especially when there are no Finns.

But when Poles and Finns are on the plot minus Russians, then Poles are close to Swedes and Finns are in woop woop (ie. far away).

See second PCA from top...

http://eurogenes.blogspot.com/2009/03/another-pca-plot-of-europe.html

Anyway, like I said, all results thus far place the Finns far away from everyone else...see here for yet another genetic distances table. This is extreme founder effect and drift.

http://dienekes.50webs.com/arp/articles/greekadna/tehvalu.jpg

mikej2 said...

Polak, don't be confused with terms, Finn, Finnish, Finno-Ugrian, otherwise this is continuous misunderstanding...

Polak said...

What does the data I linked to say? Does it say Finno-Ugric, or Finland?

Read this report carefully...

https://oa.doria.fi/handle/10024/44624

mikej2 said...

Polak, I have already read both those texts and there is no contradiction. I have always tried to tell that Finns were in ancient times several tribes and this is the reason for large genetic differenses between East Finns, North Finns and West Finns. Earlier many bloggers claimed that Finns are very homogeneuous. Wrong I said. Finns are not homogeneous, but have small diversity. Now these bloggers seems to say same.

This has nothing to do with our earlier discussion about how pricipal component figures are generated. The question was about method. Even after we have added the Northern Russians in the PC-image, Finns are very different kind themselves, because of their history, but they are closer Swedes (because PC-mapping works so).

I know that Poles and Swedes are quite close each other. I am not sure, but it can be result from much travelling Eastern Germans in ancient times.

Polak said...

Poles and North Germanic nations share common ancestry via the Corded Ware cultures of the Bronze Age. After that we kind of drifted apart.

We don't have very much Germanic admixture per se. Russians have more Scandinavian influence and Czechs a lot more German. Many Russians also have a lot more Finnic admixture.

Otherwise Poles are almost identical to most Russians and Czechs.

Western Finns do have a lot of Scandinavian influence, and many do look very Nordic, as I'm sure you kow. That's why Dieneks' post here is way off.

But that doesn't change the fact that drift has skewed the frequencies of these genome-wide SNPs that you will never cluster close to anyone on any PCA plot no mater who's on it. You'll always find your own space.

mikej2 said...

Polak "But that doesn't change the fact that drift has skewed the frequencies of these genome-wide SNPs that you will never cluster close to anyone on any PCA plot no mater who's on it. You'll always find your own space."

You are right, but there is still much to do by geneticists and many of us now making serious conclusions have in future to eat own hat :)

Polak said...

I won't be dining on any hats, because I know what I'm talking about.

Crimson Guard said...

Yes, Deniker used those terms as he came up with the basic and true classification system which stood the test of time and all other seemed to have imitated with various degrees of results:

"The Xanthichroid group contains the two fair races of Europe, Northern and the Eastern."

http://s1.zetaboards.com/anthroscape/topic/1197926/1/#new

http://s1.zetaboards.com/anthroscape/topic/945429/1/#new

Polak said...

Deniker's system has stood the test of time in Crimson's bedroom, but nowhere else.

Going by Dienekes' theory here, correlating these PCAs to a Nordic vs. Baltic race theory obviously means that the Irish are the most Nordic of the non-Finns, while the Swedes the least Nordic and most Baltic.

In that case I think we have a problem, because I would say that classic Baltic traits are as common in the Irish as in Swedes.

Also, clearly, part of the genetic variation in Northern Europe that contributes to phenotyic differentation is low level extra-European admixture way up north (Uralic admix). The PCA plots seen here don't deal with that at all.

mikej2 said...

Polak

"Also, clearly, part of the genetic variation in Northern Europe that contributes to phenotyic differentation is low level extra-European admixture way up north (Uralic admix). The PCA plots seen here don't deal with that at all."

I am a participant in 23andme and I have about ten Finns and 150 other Europeans and European Americans on my comparison list. 23anme offers analyzing by MDS-mapping and a gene comparison tool, which compares 500k SNPs gene by gene. Some notices

- Northeastern European has often a tiny segment of Asian classified genes. This is true not only under Finns but under many Russians, Balts and Swedes. This means some bigger Asian influence.

- this Asian part of genes has nothing to do with the placement on MDS-map, it is too tiny to affect. It has also nothing to do with gene-per-gene comparison.

- Finns are the most Northeastern group, although there is an evident need for other Northeastern samples to have a comprehensive picture.


- in comparison gene-per-gene a typical top ten result for Finns are 6 Finns and 4 non-Finns. This is a typical result for all tested persons in all nationalities (6/10 stand for own nationality), but some other nationalities seems to have a bit lower mutual similarity values than Finns, which can be result from higher group diversity.

- the genetic distance evaluated by the MDS-map from continent Europeans to Southern Italians and Eastern Finns seems to be quite equal.

- as summary my opinion is that there is an actual need to study the genetic distances seen especially in different PCA-maps whether it is caused by a genetic drift, because this shown tiny Asian segments are not efficient and this leaves open the question where the Finns have come during prehistoric times.

Polak said...

Mike, you're confused.

Minor Asian and African admixture does have a direct effect on the placement of individuals on inter-continental PCAs. See second diagram here.

http://polishgenes.blogspot.com/2009/03/admixture-analysis-of-4448-central.html

And there are many more examples, like in the latest Auton study, where one Australian has a lot of South Asian admix, and is placed on a PCA betweeen Europeans and South Asians. See PCA here.

http://polishgenes.blogspot.com/2009/02/genomic-diversity-around-globe.html

It also has an effect on pairwise Fst comparisons between European and extra-European populations.

It doesn't have any effect on the placement of individuals on intra-European PCAs, like the ones presented in this blog entry.

As for 23and me, you need to have another look. Those with some Asian influence are closer to Asians on the MDS plot that shows the whole world. After that, on the intra-European and intra-regional MDS plots, that foreign admix becomes irrelevant.

mikej2 said...

Polak,

sorry, I gave consufing information. I meant Easte Asian blocks, CONTINUING SNP-sequences found in chromosomes. I wasn't talking about total amounts of some Asian genes. My apologies.

Dienekes said...

Going by Dienekes' theory here, correlating these PCAs to a Nordic vs. Baltic race theory obviously means that the Irish are the most Nordic of the non-Finns, while the Swedes the least Nordic and most Baltic.I really don't see how you can draw that conclusion. The PCA (and STRUCTURE) plots both have a first-order differentiation between Finns and the rest, corresponding to the Nordic and Eastern races in Deniker's terminology.

But, within the Nordic cluster there is a further cline on the one end on which are the Irish, on the other Danes/Swedes. Deniker understood this, which is why he assigned a nord occidentale minor race to Ireland and thereabouts.

In retrospect, and given the limitations of his time, Deniker was quite brilliant.

In that case I think we have a problem, because I would say that classic Baltic traits are as common in the Irish as in Swedes.Deniker never assigned any "Baltic" traits to the Irish. Indeed, I know of no researchers in the Deniker tradition who made that kind of observation.

Dienekes said...

Polak wrote: "Minor Asian and African admixture does have a direct effect on the placement of individuals on inter-continental PCAs. [...] It doesn't have any effect on the placement of individuals on intra-European PCAs, like the ones presented in this blog entry."

The above is factually wrong; the placement of individuals in any PCA plot depends on the genetic variation of these individuals and makes no distinction between "intra-European" or "extra-European" sources of variation.

The plots presented here actually demonstrate this quite nicely. In the first one, we see that Swedes deviate in one direction. We do not know a priori whether this is due to something specific to Swedes or due to their similarity/admixture with an unsampled population.

When we add the outgroup of the Finns (second plot) it is revealed that the deviation of the Swedes is due to them being "drawn" in a Finnish direction. But, again, we don't know whether this deviation is due to something specific to Finns, or to Finns being drawn towards an unsampled population.

Similarly, when we include Europeans with an Asian outgroup, we realize that Finns do not simply randomly deviate from other Europeans but they deviate in an Asian direction.

Polak said...

You're suggesting that "east" on intra-European plots is the same as "east" on an inter-continental plots. But this is crap.

That's why East Europeans who are 100% European cluster east of Anglo-Americans with NA admix on intra-European plots, but west of these people on inter-continental plots.

So that kinda kills your whole theory right there.

Dienekes said...

You're suggesting that "east" on intra-European plots is the same as "east" on an inter-continental plots. But this is crap.In the intra-European PCA, Europeans occupy a narrow band along PC1. The main feature of PC1 is not an East-West cline across Europe, but differentiation between Finns and the rest. It is actually PC2 that captures the East-Wesst cline in northern Europe (from the Baltic to Ireland).

mikej2 said...

Dienekes

"Similarly, when we include Europeans with an Asian outgroup, we realize that Finns do not simply randomly deviate from other Europeans but they deviate in an Asian direction."

It can be easily seen how this affect to Swedish cluster.

But every new added population group affects of course to all individuals and the result depends on wherefrom we catch these new population groups. This difference is result of the individual level gene admix. If we have the new populations with an Asian outgroup somewhere near by Central Asia instead of North Siberia with same about of East Asian inheritance, we get a quite different result.

Polak said...

Dienekes,

Asian influence is basically totally overridden by within-European variation on intra-European plots. I can show you plenty of examples.

So Finns don't deviate in an Asian direction on intra-European plots, they deviate in a FINNISH direction...and indeed, some of these Finns on your graphs are 100% European, while some of the Swedes and Brits are not.

What this means is that you're totally off with your Deniker comparison, because you're equating the results of extreme Finnish genetic drift with some sort of Baltic sub-race. That's crazy.

Polak said...

Hey Mike, why don't you help us out here. You're sharing with plenty of people on 23andme, right?

Pick a Finn who's 100% European and a Brit who's 1% Asian. Then tell us where they cluster in respect to each other on:

1) The Global plot.

2) The European plot.

Dienekes said...

Yes, that mysterious "intra-European" variation that causes Sweden and Denmark to be closer to Ireland than to Finland along PC1.

you're equating the results of extreme Finnish genetic driftAs per my link, Finns deviate from other Europeans not in a random direction (as would be expected under drift), but in a specifically Asian one (as would be expected under admixture).

Dienekes said...

Your suggested experiment "Pick a Finn who's 100% European and a Brit who's 1% Asian. Then tell us where they cluster in respect to each other on:" is meaningless, as the average Brit has no Asian admixture, while the average Finn does.

Polak said...

What's mysterious about the fact that Swedes, Danes and the Irish come from mostly the same ancient population sources in North/Central Europe? Much of the Finnish genetic variation also does, but to a much lesser degree, and then of course there's the massive Finnish drift. Hence the differentation at PC2.

Btw, so all Finns deviate in an Asian direction even when they don't have any Asian admixture?

mikej2 said...

Polak said "Hey Mike, why don't you help us out here. You're sharing with plenty of people on 23andme, right?

Pick a Finn who's 100% European and a Brit who's 1% Asian. Then tell us where they cluster in respect to each other on:

1) The Global plot.

2) The European plot."

It will tell nothing, because 23andme reports only the amount of Asian "haploblocks", i.e. continuous SNP-blocks and these have in size class 0-2% no effect in this point of view.

23andme reports total similarities with Asian, Africam etc by percents of total genome (500k in this case). For me for example the Central Asian portion is 66,95 %. The North European proportion is 67,82%. These numbers and similarities are meaninful.

Dienekes said...

Polak wrote: "Btw, so all Finns deviate in an Asian direction even when they don't have any Asian admixture?"

I said: "the average Brit has no Asian admixture, while the average Finn does."

Polak said...

Mystery solved thanks to this structure plot...

http://img25.imageshack.us/img25/9633/northeuro.jpg

a) The small amounts of Asian admix (orange at K=3) make many Finns (but not all) and some Swedes drift away from the other Europeans on inter-continental plots. Like here...

http://img100.imageshack.us/img100/3987/finscb.jpg

b) The Finnish genes (purple at K=4) make the Finns drift away from other North Euros on intra-European plots in a FINNISH direction. Swedes also move that way somewhat because of some purple. Like on your plot on PC1.

I don't really know how to explain the above in any more obvious way.

But more info here...

http://eurogenes.blogspot.com/2009/05/genetic-variation-in-baltic-sea-region.html

Polak said...

Mike,

Ancestry paintings correlate with plcement on the Global MDS plot (anyone with 1% or so of Asian influence is east of the rest).

Dienekes said...

Interestingly, Fst between Europe and East Asia is about 0.1. In this paper, Fst between Finland and CEU HapMap is 0.0637, i.e., Finland is moved away from Europeans by exactly as much as would be expected by the levels of East Eurasian admixture observed for them and Finno-Ugrian influenced Russians in other studies.

Polak said...

See now you're onto something.

Elevated affinities with East Asians (relatively lower pairwise Fst scores) do have an influence on anthropological traits. That's North Eurasian admix talking, and you can actually see it in far Northeast Europe.

But this has nothing to do with Deniker's Baltic race, or the Finns drifting in a Finnish direction on intra-European plots.

Scandinavians who normally cluster with the Dutch and English on intra-European plots often too have elevated East Asian affnity (lower pairwise Fst scores) and many do show Saami type facial characteristics.

Dienekes said...

Also consistent with the position of Finns in a global PCA after Jakkula et al.

Dienekes said...

the Finns drifting in a Finnish direction on intra-European plots.The "Finnish" direction is an East Eurasian direction. Finns are spread along this direction (while most other Eurasians are not) because Finns have a small (but variable) level of East Eurasian admixture.

Polak said...

Ok, explain to me why Norwegians have elevated affinity to East Asians in terms of Fst compared to Central Europeans, yet they always cluster further west than them on intra-European plots?

mikej2 said...

Polak

"Ancestry paintings correlate with plcement on the Global MDS plot (anyone with 1% or so of Asian influence is east of the rest)."

Of cource it correlates, but it is not any part of the MDS-plot. If the Ancestral painting shows enough Asian Haploblock, there is likely more Asian similarity. But Asian a tiny haploblock doesn't mean that who it has is placed to the east and without it is paced to the west.

Dienekes, also Orcanians have according to studies significant Asian inheritance.

Polak said...

Mike,

Do people with 1% Asian ancestry cluster east of those with 0% Asian ancestry on the global MDS plot, even if they're from western Europe?

Yes or no?

mikej2 said...

Dienekes "Interestingly, Fst between Europe and East Asia is about 0.1. In this paper, Fst between Finland and CEU HapMap is 0.0637, i.e., Finland is moved away from Europeans by exactly as much as would be expected by the levels of East Eurasian admixture observed for them and Finno-Ugrian influenced Russians in other studies."

The Geographical structure and differential natural selection amongst amongst North European Populations /2009 shows Population pairwise FST between CEU and Finns 0.00637

mikej2 said...

Polak

"Do people with 1% Asian ancestry cluster east of those with 0% Asian ancestry on the global MDS plot, even if they're from western Europe?

Yes or no?"

I cannot answer yes/no because 23andme doesn't give such kind of cluster values. They give only Asian Haploblocks, which is surely different thing. Secondly they give global similarity percents and they do have a correlation on their Mds-map.

You have a wrong question and I don't aim to give you wrong answer :) .

Polak said...

Why don't you see where a Western European with 1% Asian admix clusters compared to a 0% Asian Finns on the Global plot.

How hard can that be?

I just checked an Anglo-American with a bit of admix vs. a 100% European Russian, and the American was east of the Russian.

eurologist said...

This is all too complicated or convoluted to me. What ever happened to the idea that Finns are largely impacted by migrations from Eurasia just East of the Caspian Sea - an area that was not possible to cross until quite some time after LGM, rather than coming from "East Asia"?

Yet, this happened long before the age of agriculture and animal domestication, when main-stream Europeans started to dominate this region.

Polak said...

eurologist: "What ever happened to the idea that Finns are largely impacted by migrations from Eurasia just East of the Caspian Sea - an area that was not possible to cross until quite some time after LGM, rather than coming from "East Asia"?"

It's not true.

Finns firmly cluster with other Europeans when compared with non-European populations, including Siberians and Central Asians.

But they deviate from North Europeans on intra-European plots because they have a heavy East European component, and because of extreme genetic drift due to isolation.

At the same time though, they're always in their own space because of that drift, and not close even to Estonians or other East Europeans. See here...

http://polishgenes.blogspot.com/2009/02/estonian-gene-project-estonians-more.html

mikej2 said...

Polak:

"I just checked an Anglo-American with a bit of admix vs. a 100% European Russian, and the American was east of the Russian."

I wonder how you got this admix, because 23andme doesn't tell such numbers. They tell only the global similarity and the combined sum of Asian habloblocks. They themselves have decided the size to be ancestral size.

But you'll have the answer :) One Finn have 1.2% Asian habloblocks and an another Finn has no such blocks (ancestfry 0%). The first one is just above the Poles near Swedes and the latter above Russians, far to east.

Dienekes said...

Ok, explain to me why Norwegians have elevated affinity to East Asians in terms of Fst compared to Central Europeans, yet they always cluster further west than them on intra-European plots?I suspect that you are referring to this study by Lao et al.

In that study, Southern European populations are included, so one of the axes of variation is latitudinal and northern European variation (i.e., both PC1 and PC2 of the current post) is projected on the other axis.

Thus, in that study, eigenvector2 is a composite of both East-West variation in northern Europe and Finnish specificity; the position of Norwegians on it is not determined by any sort of minor "Finnish" tendency which is unimportant for them.

The fact that PC1 here and eigenvector2 there are not expressions of the same entity can be further seen by noticing that Ireland-Sweden-Finland are collinear along eigenvector2 in Lao et al. there but not along PC1 in McEvoy et al.

Polak said...

You're looking on the wrong plot.

I said the world plot, not the European one.

mikej2 said...

Eurologist

"What ever happened to the idea that Finns are largely impacted by migrations from Eurasia just East of the Caspian Sea - an area that was not possible to cross until quite some time after LGM, rather than coming from "East Asia"?"

It was a valid theory, 100 years ago. Sorry, things goes complicated, but it is not my fault :) .

Polak said...

Dienekes,

Norwegians have greater Asian affinity than Poles.

Why are they west of Poles on every intra-European PCA plot ever released?

Polak said...

Mike,

Look on the goddamn global plot, not the European one.

Are you retarded or something?

mikej2 said...

Polak: "You're looking on the wrong plot.

I said the world plot, not the European one."

Same thing whatever level you look. Infact Finn with 0% aAian block is the most on the right on my map. Believe me, I know this issue :)

mikej2 said...

Polak, it seems that I am not able to explain you how these numbers work. Let's change the topic. You must at first make this clear to youself.

Dienekes said...

Why are they west of Poles on every intra-European PCA plot ever released?Now, go back, read and understand what I wrote, and perhaps you won't have to ask such an irrelevant question.

Polak said...

Ok fine.

How many Poles do you have results for? And what are their admix rates, if you don't mind saying?

Polak said...

Dienekes,

I understood what you wrote. It just means that Finns have been affected by strong random drift in their own direction.

Notice that in Heath et al. 2008 there are some Russians who deviate strongly towards Asians on the inter-continental plot, but then mingle with Poles on the intra-European one.

So the Asian factor is not what's making the Finns look funny, because it would also make the Russians look funny. The Fst difference in Asian affinity between Poland and Russia is huge, and that must be the result of the subset of Russians deviating to Asia, and these are probably just Finno-Ugrians or Tatars with some Slavic admix.

mikej2 said...

Polak "So the Asian factor is not what's making the Finns look funny, because it would also make the Russians look funny. The Fst difference in Asian affinity between Poland and Russia is huge, and that must be the result of the subset of Russians deviating to Asia, and these are probably just Finno-Ugrians or Tatars with some Slavic admix."

Russian have assimilated many ancient Asian tribes, don't forget the Golden Horde, which never came to Finland. Finnish eastern genes are very much older and evenly spread. But what is meaninfull, is that Russians have their own Eastern European genes, typical only for them.

Dienekes said...

Notice that in Heath et al. 2008 there are some Russians who deviate strongly towards Asians on the inter-continental plot, but then mingle with Poles on the intra-European one.Russians also "mingle" with Poles in the global PCA. Not very surprising, as Russians and Poles are very close relatives, with the former having picked up some Finno-Ugrian elements as well.

"Interestingly, the first two components of the European only
PCA are almost identical to components 3 and 4 from
the PCA, including all four HapMap cohorts"

Polak said...

There's a subset of Russians on the global PCA who are further from Poles than the Spaniards.

That's who I'm referring to. You didn't notice them?

Polak said...

Btw, what's your point exactly abot components 3 and 4?

1 and 2 are the main components, and if we looked at all the components, Poles and Russians would still not be in the same spot in relation to Asian samples (see Fst data).

Polak said...

Ok, I'm getting bored so here are the knockout punches to your litle theory. Fst values from Heath et al. 2008

Russians vs. Chinese 0.1036
Swedes vs. Chinese 0.1073
Poles vs. Chinese 0.1086

..and False positives from Tehva Lu.

Swedes vs. Finns 0.078
Poles vs. Finns 0.096

My god, nothing fits? :(

Dienekes said...

My point is that the global and European PCAs capture different aspects of variation in this case. This is proven by the fact that the first two principal components in the European-only PCA correspond to components 3 and 4 in the global PCA.

It is not at all surprising then that Russians and Poles mingle in the European-only PCA and are separated in the global PCA: the "East Eurasian" element in the global PCA is captured in the first two principal components, and the European variation is captured in the 3rd and 4th one, in which Poles mingle with Russians due to their common genetic background as Eastern European Slavs.

Incidentally, the fact that in components 1 and 2 Russians and Poles do not mingle, and in the lower-order components 3 and 4 they do, further reinforces the fact that eastern Eurasian influence is the major cause of variation in this region, which is also supported by the Fst table.

I don't know what relevance you see for Finns, as Finns deviate from the European cluster in the first 2 PCs for both intra-European and inter-European plots. If Finns were "mingling" with Europeans in the European-only plot, but not in the global plot, then it could be argued that the European-only plot captured only intra-European sources of variation. But that if is false.

Dienekes said...

Ok, I'm getting bored so here are the knockout punches to your litle theory. Fst values from Heath et al. 2008Which "little theory" of mine is invalidated by the fact that Russians are closer than Poles to the Chinese by an amount that is 17 times the Russian-Polish Fst and in the order of the largest observed intra-European Fst?

mikej2 said...

I have following Fsts for Finns:

Denmark 0.00551
Ireland 0.00721
Netherland 0.00627
Sweden 0.00437
UK 0.00642

And from an another study

Finn East
Sweden 0.0072
FinW 0.0032
GER 0.0084
BRI 0.0094

Finn West
Sweden 0.0030
FinnE 0.0032
GER 0.0033
BRI 0.0042

What next?

Polak said...

Dienekes, the factors influencing the frequencies of anthropolgical traits within Europe go way beyond the first few components of between-population genetic variation. All that stuff about individual best matches has to be taken into account too, and then also the environment and living conditions. Surely you know that.

Deniker didn't do anything remarkeable, apart from making a few broad observations, some of which have stayed true to today. But I personally don't see much of a correlation between these particular PCAs and his theories.

You're just encouraging stupidity (fe. from Crimson) by making posts like this.

pconroy said...

OT, but if anyone want to share 23andme data - just contact me, same name as here. I'm happy to share at either Basic or Full.

Cheers!

argiedude said...

Somewhat off topic.

On page 48 of the recent study
"Human Genetic Variation in the Baltic Sea Region: Features of Population History and Natural Selection", by Tuuli Lappalainen, it talks about criticisms to Zhivotovsky's effective mutation rate and refers specifically to the work by Pontikos (2008).

Алексей said...

to mikej2and everybody who wants detailed information about Russians.

Here's the acticle "Two Sources of the Russian Patrilineal Heritage in the Eurasian context" by O. Balanovsky at al. (in English).

The detailed information about research of Russian gene pool by O.Balanovsky and E.Balanovska with more maps you may find in extracts from their monograph "Russian gene pool on East-European plain" (in Russian).

mikej2 said...

Thanks Aleksei,

I can see that in this study we have YDNA mappings over the European part of Russia. There is some problem to compare this study to what we have been discussing, because the Ydna has no recombinationa and it has therefor longer invariability.

But some things can still be seen. On page 242 you can see that the northern part of Russia, to north and northwest from Moscow, has remarkable quantity of to haplogroups, I1 (formely I1a) a and N1-tatc (formely N3-tat). Basically this is the area which I want to see also in autosomal MDS-maps. The problem is that Russia is a multinational country and all look it from own point of view, as we have seen before.

drfaust-spb said...

mikej2,

not a problem with autosomal DNA. The research of russian gene pool by Balanovsky and Balanovska include their own expedition data on anthropology, dermatoglyphics and genetics (autosomal, Y and mtDNA). In extractions from their monograph you can see syntetic maps for every of these fields of knowledge.

What about syntetic autosomal maps, here they are (link to monograph extractions upwards):
- figure 8.2.3. - the map of gene pool divesity in East Europe (first principal component);
- figure 8.2.4. - the map of gene pool divesity in East Europe (second principal component).

Here's also the map of genetical distances from average russian population in Eastern Europe (autosomal DNA).

P.S. My another one login.

drfaust-spb said...

Note, to get better quality of images you should download the document from Scribd.

mikej2 said...

drfaust-spb, I don't see any autosomal maps, only a MDS-map of Y-chromosomal variation. If we are talking about the same Balanovsky study...

mikej2 said...

Yes, I se now what you mean. There is a couple of problems; I don't understand Russian and these maps have not European data. I hope to see PCA- and MDS-maps covering Russian, Eastern Europe and Northern Europe, because in real PCA and MDS everything corralate everything.

drfaust-spb said...

The refered monograph couldn't help you about Northern and Western Europe. There's much information on Y and mtDNA for Europe and Eurasia, but complex autosomal data is too large-scaled: Europe, Northern Eurasia, Indostan etc.

Much more detailed and wide information there is for populations of ex-USSR and Russians. So if you need something about them - you're welcome.

mikej2 said...

Thank you for all information! I got help in Russian and have now basic knowlwdge to understand this study. The last picture seems very interesting, the locations of Finns, Lithuanian, Northern Russian and Karelians are in my interest and this seem to be something I figured earlier. I am very thankful :) I believe I could have much new information from Russian studies.

drfaust-spb said...

Our interest in specified people concides. :) If you'll need a deeper information in direction of East-European gene pool (the extracts are just part of results) or help with Russian, you may find contacts in my profiles.

antoine1706 said...

I have following Fsts for Finns from "Genome-wide Insights into the Patterns and Determinants
of Fine-Scale Population Structure in Humans", Shameek Biswas, 2009 :

Finns (Helsinki) :

Austria 0.0060
Bulgaria 0.0090
Czech Republic 0.0060
Estonia 0.0040
France 0.0080
Northern Germany 0.0060
Southern Germany 0.0060
Hungary 0.0060
Northern Italy 0.0130
Southern Italy 0.0160
Latvia 0.0070
Lithuania 0.0070
Poland 0.0060
Russia 0.0060
Spain 0.0110
Sweden 0.0050
Switzerland 0.0090
CEU 0.0060

Finns (Kuusamo) :

Austria 0.0130
Bulgaria 0.0150
Czech Republic 0.0120
Estonia 0.0090
France 0.0150
Northern Germany 0.0120
Southern Germany 0.0130
Hungary 0.0130
Northern Italy 0.0200
Southern Italy 0.0230
Latvia 0.0130
Lithuania 0.0130
Poland 0.0120
Russia 0.0120
Spain 0.0170
Sweden 0.0110
Switzerland 0.0150
CEU 0.0130

As we can see there is less genetic distance, for example, between Northern Italians and Palestinians (fst 0.0108) than between Northern Italians and Finns (even from Helsinki fst 0.0130) ! Bad news for these stupid euro-nationalists that would like to match "political" europe borders with genetic ones ...