March 27, 2013

Population structure in the Netherlands

The three PCs are color-coded in panels b,c,d.

European Journal of Human Genetics , (27 March 2013) | doi:10.1038/ejhg.2013.48

Population structure, migration, and diversifying selection in the Netherlands

Abdel Abdellaoui et al.

Genetic variation in a population can be summarized through principal component analysis (PCA) on genome-wide data. PCs derived from such analyses are valuable for genetic association studies, where they can correct for population stratification. We investigated how to capture the genetic population structure in a well-characterized sample from the Netherlands and in a worldwide data set and examined whether (1) removing long-range linkage disequilibrium (LD) regions and LD-based SNP pruning significantly improves correlations between PCs and geography and (2) whether genetic differentiation may have been influenced by migration and/or selection. In the Netherlands, three PCs showed significant correlations with geography, distinguishing between: (1) North and South; (2) East and West; and (3) the middle-band and the rest of the country. The third PC only emerged with minimized LD, which also significantly increased correlations with geography for the other two PCs. In addition to geography, the Dutch North–South PC showed correlations with genome-wide homozygosity (r=0.245), which may reflect a serial-founder effect due to northwards migration, and also with height (♂: r=0.142, ♀: r=0.153). The divergence between subpopulations identified by PCs is partly driven by selection pressures. The first three PCs showed significant signals for diversifying selection (545 SNPs - the majority within 184 genes). The strongest signal was observed between North and South for the functional SNP in HERC2 that determines human blue/brown eye color. Thus, this study demonstrates how to increase ancestry signals in a relatively homogeneous population and how those signals can reveal evolutionary history.

Link

6 comments:

eurologist said...

Nice study.

"PCs derived from such analyses are valuable for genetic association studies, where they can correct for population stratification.
The divergence between subpopulations identified by PCs is partly driven by selection pressures."

I have said as much (regarding properly executed PC analysis) in the recent NE European study when arguing that the second, Uralic component was "real" (i.e., with a very deeply-rooted local background) and not just the consequence of recent drift.

"In addition to geography, the Dutch North–South PC showed correlations with genome-wide homozygosity (r=0.245)"

This could be a secondary effect: phenotypes such as blue eyes and blond hair strongly signal "you are part of us" or "you are not." Once they are established at a significant level, they may be self-preserving, and with that, lead to higher homozygosity in general. Of course, religious differences during the past 500 years surely also play a role.

I don't buy the author's "serial-founder effect hypothesis" over what is effectively just a couple of hundred kilometers.

"It is estimated that the ancestors of ~75% of what we currently call the ‘native’ Dutch population (autochtonen in Dutch) have immigrated into the Netherlands during the past 20 centuries."

Yeah - but I bet a huge fraction of that is from the neighboring Grafschaft, Emsland, and Frisia - who weren't that different to start with, and still speak mutually intelligeable dialects across the boarder, today. In fact, this probably accounts for much of the PC1 and PC2 gradients.

And I bet that if you made a similar study of thode three regions, you would find the same PC1 and PC2. I am less certain of the significance and origin of PC3; it may reflect the highest concentration of 'autochthonous' Dutch during Roman times.

Wim Penninx said...

The research (differences in homozygosity) is interesting. The conclusions are interesting speculations.

Other reasons for the differences seem more likely. My country (The Netherlands) consists roughly out of three areas.

South of the rivers (Rhine, Maas) was the area (higher and more sand areas) to have an increase of population.
It is cultural and population close to Belgium, Flanders, the language is closely related. The first large scale DNA research was
done as the Hertogdom Brabant (http://www.brabant-dna.org) since the southern Netherlands were part of the Brabant area. It was
also the part that was par of the Roman Empire.

The language of North and East is Low Saxon (http://en.wikipedia.org/wiki/West_Low_German). The population grew later than in the South.

Holland (west of the Netherlands) has mainly people who orginated from many areas, and who came
to this area much later. It is more a melting pot.

A northwards migration is not very unlikely. A later population growth in the north seems an easier explanation.

Abdel Abdellaoui said...

Thanks guys, you raise some interesting points!

´I don't buy the author's "serial-founder effect hypothesis" over what is effectively just a couple of hundred kilometers.`

I agree that it seems an unlikely effect to observe in such a small region, but as I say in the discussion:

´This effect does not necessarily have to reflect an upward migration that took place within the Netherlands; it may also be that, more recently, Southern Europeans migrated more to the South of the Netherlands, while Northern Europeans migrated more to the Northern parts of the country, maintaining the North–South distribution within the country.´

Periods where this could have happened is during the Roman empire and the wars with the Spanish for example.
It seems that we mainly capture the European North–South cline with PC1, considering the correlation with height, and selection pressures on eye colour and metabolism related traits (FTO & Lactase gene). These are just as unlikely to have happened within a couple of 100 kilometers. The serial-founder effect is just another effect that is also observed in the European North-South cline. And also don't forget Figure 2a...

I agree with you that religion may have played a substantial role in maintaining the North-South distributions. The map of the religious distribution in the Netherlands strongly resembles PC1 (see http://upload.wikimedia.org/wikipedia/commons/c/c9/Nederlandgodsdienst1849.PNG), and this country had a long history of religious segregation and assortment. I currently have a paper under review where I explore the role of religion on genetic variation in this country.

If you can't access the full text btw, the pdf can be found here:

http://www.tweelingenregister.org/nederlands/verslaggeving/NTR-publicaties_2013/Abdellaoui_EJHG_2013_epub.pdf

eurologist said...

A few more notes and observations on this. Frisians form an arc from the NE German "North Sea" to W along the coastline both in phenotype and language, to date. Frisian is actually close to W/SW Dutch, while low German is rather different and closer to eastern Dutch.

In the south of the Emsland and in the "Grafschaft" (Bentheim), people are a bit more metropolitan and show clear signs of admixture with people from the (partially Romanized) Lippe area (Westphalian vs. low German within Phalian languages --> Westerwoldinger Platt in the Netherlands). Linguistically, this is confirmed by partial appearance of the [X --> x] and [g -> ɣ] Westphalian substitutions, and novel, broadening diphthongs.

In phenotype, in the central Emsland you can find people of more Baltic features (a bit more stocky, more subcutaneous fat, and flat, broad noses with deeply-rooted wings curling up). Not quite born there, but think of a shorter Alexandra Popp with red-blond, slightly curly hair.

On the flip side, towards the southern Emsland and into the Grafschaft and NW Westphalia, you find a lot of rather tall, white-skinned gracile women with seemingly thin bones and a bit of an androgynous Elfen look; appearing nobel or wise, determined, at times fierce. A (typically more blond) Kerstin Garefrekes is in this direction, and you can find the same phenotype all the way to the Flemish and surrounding regions of Eastern Belgium.

eurologist said...

Here are some pictures from the German regions adjacent to the Netherlands:

Examples of typical, Baltic, Emsland faces:

http://www.sv-haren.de/media/fotos/2012-11-18%20Mehrkampf%20Haseluenne.jpg
http://www.sv-haren.de/media/fotos/2013-03-21-23%20TrainingslagerI.JPG
http://www.marianum-meppen.de/archiv/sportarchiv/sport2011b/sport2011b/sportnews2011_24_01.jpg
http://www.marianum-meppen.de/archiv/sportarchiv/sport2007/fbwk3maedchen.jpg


More "Elfen" features (especially, third and last from left):

http://www.marianum-meppen.de/archiv/sportarchiv/sport2008/2008/sportnews2008_02k.jpg

In the following, you can see the dichotomy of more "Elfen" vs. "Baltic" (or saxon, in parentheses) features:

e.g., top row, #4 vs #5 (top row, #2; second row, #3+6,)
http://www.emsdettenervolkszeitung.de/storage/pic/mdhl/automatischer-bildimport/mz-mlz-evz-gz/emsdetten/3168398_1_0522EMS-BORUSEN-MADELS.jpg?version=1337606318

mr. Knows When said...

The main reason there is there is such a big difference between north and south is quite simple. The Rhine was a barrier that prevent admixture with southern population for more than thousands of years. The first bridges across the Rhine were in the late 19th century. Besides the Meuse and the Rhine are both rivers that southern orientated hence, more brown eyed. While the northern shore remained more coastal populations of the Northern European descent.

I'm not sure if one can really make this clear distinction. However it is true in lower Saxony people are generally built more robust than in Holland or Friesland. The wrists and enkels are much broader, but also they are usually a bit shorter than the Dutch which are lanky.

But what I always find fascinating to see is that Belgians have a very different head shapes compared to the Dutch. Often they have ball shaped heads while the Dutch have more elongated heads.