Autosomal map : Mediterranean admixture (from Dodecad)

Maciamo

Veteran member
Admin
Messages
9,950
Reaction score
3,230
Points
113
Location
Lothier
Ethnic group
Italo-celto-germanic
I have made distribution maps for most of the major haplogroups. I thought it would be interesting to compare Y-DNA distributions with autosomal admixtures. I used the data for the Mediterranean admixture from the Dodecad Project to make this map. It's far from perfect because many key populations were missing (Croatians, Bosnians, Serbs, Czechoslovaks, Belarusians, Ukrainians, most of the Caucasus beside Georgia and Armenia). I used the data from individual members to fill the gaps.

Mediterranean-admixture.gif


In my opinion, this Mediterranean admixture has a strong E1b1b (E-M78 and subclades, though not E-M81) component to it, as well as a likely I2a1 influence in places like Northeast Spain and Sardinia. I believe that R1b in Western Europe is hiding a lot of anterior lineages, which could have been principally I2a and E-M78.

Autosomal genes are naturally more evenly spread out inside a population sharing the same borders or language. That is because women are more likely to be "imported" as wives from other regions, and also because autosomal genes mix very quickly in the gene pool. A third reason the autosomal map looks more evenly distributed is that I didn't have as much data for regional populations (mostly country averages).

There is no correspondence with G2a, J2 or J1 because this Mediterranean admixture is lowest in the Mediterranean region where G2a, J1 and J2 are the highest.

Haplogroup-E1b1b.jpg


I doubt that it is a coincidence that this Mediterranean admixture matches quite well the so-called Mediterranean Race described by anthropologist Madison Grant. One of the main characteristics of Mediterranean people is the dolichocephalic skull shape (long head), which they have in common with the Nordics.

Passing_of_the_Great_Race_-_Map_1.jpg
 
I agree that the distribution definitely matches E1b1b + I2a1 the best. It's also funny how that old anthropologist was actually up to something.

The only place where that strikes me as non-matching is northern Europe, however: how does one explain the large concentrations in Britain and Scandinavia?
 
I agree that the distribution definitely matches E1b1b + I2a1 the best. It's also funny how that old anthropologist was actually up to something.

The only place where that strikes me as non-matching is northern Europe, however: how does one explain the large concentrations in Britain and Scandinavia?

That's because these are high-R1b countries. I suspect that the Indo-Europeans might have killed a big percentages of the local men when they arrived. If that is the case, regions with a lot of R1b (Catalonia, Basque country, Western France, pre-Anglo-Saxon British Isles) probably inherited their Mediterranean genes from the local women taken as wives/concubines by the R1b invaders.

As for Scandinavia, the Norwegian Dodecad members have a little bit more Mediterranean admixture than the Swedish ones, which makes sense if Irish an Scottish people were imported to Norway in the Middle Ages (and they almost certainly were, as attested by R1b-L21). What we don't know if how many Goidelic women were taken to Scandinavia, but I assume that it is at least as much, and surely more than the percentage of R1b-L21. A study has already confirm the practice of Viking men taking foreign women as wives.

Archaeological and historical records suggest individuals from the British Isles and Scandinavia settled Iceland about 1100 years ago. Based on studies of today's Icelanders, researchers have suggested that less than 40 percent of Icelandic mtDNA — representing matrilinear ancestry — originated in Scandinavia, while as much as 75 to 80 percent of their paternal ancestry, gauged by Y-chromosome DNA, is Scandinavian.

That is consistent with the notion that Scandinavian men, namely Vikings, frequently settled the area with women from other regions.

In other words, 60% of Iceland's mtDNA lineages would be of Goidelic origin, and 20-25% of its Y-DNA too (almost all of it R1b-L21, which is why Iceland has over 40% of R1b, unlike anywhere in continental Scandinavia).

All this to say that if the Irish and British have about 22% of Mediterranean admixture, it is not surprising to find 15% in Norway. I expect it to be around 18-20% in Iceland.
 
That's because these are high-R1b countries. I suspect that the Indo-Europeans might have killed a big percentages of the local men when they arrived. If that is the case, regions with a lot of R1b (Catalonia, Basque country, Western France, pre-Anglo-Saxon British Isles) probably inherited their Mediterranean genes from the local women taken as wives/concubines by the R1b invaders.

As for Scandinavia, the Norwegian Dodecad members have a little bit more Mediterranean admixture than the Swedish ones, which makes sense if Irish an Scottish people were imported to Norway in the Middle Ages (and they almost certainly were, as attested by R1b-L21). What we don't know if how many Goidelic women were taken to Scandinavia, but I assume that it is at least as much, and surely more than the percentage of R1b-L21. A study has already confirm the practice of Viking men taking foreign women as wives.

This is a good point. The largest overall R1b concentrations are after all found in that "clean-swept arc" (seen in the distribution of Haplogroup E1b, but also in Haplogroup T, for istance) stretching along the Atlantic coast from the Basque Country to Brittany (and theoretically extending into the British Isles).

In other words, 60% of Iceland's mtDNA lineages would be of Goidelic origin, and 20-25% of its Y-DNA too (almost all of it R1b-L21, which is why Iceland has over 40% of R1b, unlike anywhere in continental Scandinavia).

All this to say that if the Irish and British have about 22% of Mediterranean admixture, it is not surprising to find 15% in Norway. I expect it to be around 18-20% in Iceland.

Wow. I had known before that the Vikings were big into slavery, but I had no idea that it was THIS drastic! :petrified:
 
Mediterranean is very well separated from the non European groups as It's easy to know while checking the distances, so I desagree of course. I posted the reasons in the other post.

And also, don't forget it's not the same West Med and East Med. Eastern populations carry also quite West Asian and Southwest Asian, so the influence in phenotype is not the same in others who simply don't have, or have very low of each admixtures (the really who can include the mentioned E-M78 influences together with others).
 
Where is the reference population for mediterranean admixture?

Sardinians?

i think part of the balkans should be more mediterranean... how can be Croatia less mediterranean than western Germany??

if this map is true, Friuli Venezia Giulia is genetically diverse from the rest of italy for this trait.. but i doubt..
 
It's possible, but there isn't data about Croacia. Probably this is due to more West Asian and Southwest Asian admixture. The same happens in Italy, less Mediterranean than Iberia (having Iberia very low West Asian and Southwest Asian levels).

Here you have the averages: https://docs.google.com/spreadsheet...owN3M3UWRyNnc&hl=en_US&authkey=COCa89AJ#gid=0

Sardinians are the top Mediterranean scorers followed by Iberians. However, the Iberian Mediterranean is more likely Southwestern/Basque in a large percent, and it isn't the same representative as others.
 
i think it's genes similarity to sardinians then.

Tuscans are:
Mediterranean 47,8%
West European 25,8%
West Asian 17,3
Southwest Asian 4,3%
Est European 3,7%

South Asian 0,5%
north-west african 0,3%
South east asian 0,2%
North East Asian 0,1%
East african 0,1%
Neo African 0%
paleo african 0%
 
I think this "mediterranean" autosomal corresponds to y-dna hap I and all it's subclades both I1 and I2. I think it is pure hap I and has nothing to do with hap E or any of its subclades at all. Look at y-dna, I (mediterranean) is going to be closer to J (middle eastern) as the both evolved from the same root IJ. E is so far away on the y-dna tree from I it is ridiculous. I think E will cluster with african haplogroups.
 
I think this "mediterranean" autosomal corresponds to y-dna hap I and all it's subclades both I1 and I2. I think it is pure hap I and has nothing to do with hap E or any of its subclades at all. Look at y-dna, I (mediterranean) is going to be closer to J (middle eastern) as the both evolved from the same root IJ. E is so far away on the y-dna tree from I it is ridiculous. I think E will cluster with african haplogroups.

There is nearly no haplogroup I in North Africa and the Middle East. Hg I is far more common in the Balkans, Scandinavia and Finland than the Mediterranean admixture. So I must absolutely rule out at least I1 and I2a2 from the Mediterranean admixture (unless Croatian and Bosnian samples, which I do not have yet, happen to have an exceptionally high degree of Mediterranean admixture, but I doubt so).

What you don't see beyond the map is that the East African have about 4% of Mediterranean admixtures, and the Ethiopians 2%.

Furthermore, Iranians have about 15%, while Afghan and Pakistani tribes range from 5 to 9%.

The Mediterranean admixture is almost certainly a component of Mediterranean E-M78 (especially E-V13), I2a1 and R1b (especially the L11- variety found around the Mediterranean and the Middle East, but also R1b-S28 and R1b-Z196).
 
North Africa doesn't need any haplogroup I to explain the Mediterranean admixture, it can fit in the maternal linages (H, V, etc.). Haplogroup I can perfectly fit in all European admixtures considering all branches, since most of them are enough old to do so.

By the way, 2-4% is not huge between East Africans, this can be easily inherited via the Middle East in thousands of years. Also, the Nepalese show more West + East European than Mediterranean...nothing rare in the percents given to Iranians, Afghan and Pakistanis. It doesn't mean anything.

About Bosnians, I don't know how homogeneous are they, but if the Balkan sample shows 30% Mediterranean, I don't expect most of them to be much different, and probably Croatians go in a similar way. The actual percent looks significant in my book.

It's not clear wich haplogroups include the Mediterranean admixture. It can include many different clades of I, even those two due to its antiquity. Also, some Eastern Mediterranean represented, in my opinion, by J2b and L11 too (that's possible), and probably there's a clade we are forgeting. But I don't see it in the case of E-M78 for the reasons I have already stated in other posts, this has been simply replaced by other autosomes long ago. Also, again, it's good not to underestimate maternal linages, since they can explain more admixtures than we usually think.
 
North Africa doesn't need any haplogroup I to explain the Mediterranean admixture, it can fit in the maternal linages (H, V, etc.). Haplogroup I can perfectly fit in all European admixtures considering all branches, since most of them are enough old to do so.

And what makes you think that H and V are not maternal equivalent of E1b1b1a ? The most common varieties of H in Iberia are H1 and H3, but these are found throughout Europe, North Africa and the Middle East. H3 and V are most common among the Basques and Sardinians, so they are probably linked to I2a1, but H1 peaks in the Cantabrians (known for their high percentage of E1b1b), West Andalusians and Maghrebans.

For the sake of the argument, assuming that H1, H3 and V were all related to hg I instead, and that all I2a men in North Africa were exterminated to the last one by E1b1b1a invaders, H and V only amount to some 30% of the mtDNA in North Africa. If autosomes were passed only by women, that the percentage of autosomal DNA from H and V would be only half, or 15%. In contrast, E-M78 has a frequency of about 30% in Morocco, ranging from 10-15% in the South to 40% in North Morocco. That's makes it a better candidate.


By the way, 2-4% is not huge between East Africans, this can be easily inherited via the Middle East in thousands of years. Also, the Nepalese show more West + East European than Mediterranean...nothing rare in the percents given to Iranians, Afghan and Pakistanis. It doesn't mean anything.

It means it can be related to R1a or R1b, but not to I1 or I2.
 
And what makes you think that H and V are not maternal equivalent of E1b1b1a ? The most common varieties of H in Iberia are H1 and H3, but these are found throughout Europe, North Africa and the Middle East. H3 and V are most common among the Basques and Sardinians, so they are probably linked to I2a1, but H1 peaks in the Cantabrians (known for their high percentage of E1b1b), West Andalusians and Maghrebans.

I wonder why Norwegian have not more Med admixture given the frequencies of H1 and H3 they have.
 
The amount of haplogroup H is much higher than what you say according to this study: http://www.biomedcentral.com/1471-2156/10/8

The percent goes to 55% without need to include V (here says 8.2%: http://www.healthanddna.com/EuropeanmtDNA.pdf). So of course it can fit. H1 and H3 are linked to Iberia, of course, but are also quite present in all Europe. And haplogroup E1b1b in Iberia is not extremely huge as whole, much less in several parts of Europe, so I think it's unlikely.

And I'm sorry, but it doesn't mean anything a small percent between 2-4%. This is linked absolutely with nothing, it's just the result of an admixture core in the Middle East wich was inherited in small quantities by East Africans in progressive migrations coming from there (more than 20% Southwest Asian supports what I'm saying). Not of course in one week, thousands of years are required for this. It's just residual, the Nepalese example shows more or less the same.

And if you are interested, I can check a Bosnian girl at 23andme, who at least shows the major similarity with Southern Europe in the public profile. So I'm afraid it's very likely she has substantial Mediterranean percent. I could try if she wants to join Dodecad project too. I'll invite her, and will see...
 
Last edited:
I wonder why Norwegian have not more Med admixture given the frequencies of H1 and H3 they have.
Keep in mind this haplogroups are very old in Europe. It's very difficult to tell exactly what they represent in terms of admixture. It's good always to check both Y-DNA and MtDNA to have a better idea, although it's never enough.
 
I agree with Knovas. The mediterranean component peaks in Sardinians at 55%, North-Italy 49% and Iberia 48%, the E1b1b in these regions is, respectively, 10-11-7, so it doesn't seem very correlated. What's more, Catalans show almost 50% med, and the E1b1b in Catalonia is very low 2-3%. So, the connection with E1b1b or J is discarded. It has to be associated with I2, even I1, R1b-S116 ? and maternally with H1/H3/something else. Seems like a paleolithic european component, associated with hap I1/I2
pd: If it was a neolithic near-east component it would peak in the South-East (Balkans, Greece,etc) and not in the West of Europe
 
The Mediterranean cluster includes Southwest and Southeast Europe, sure (Greeks have very high percent too). So I don't discard haplogroup J, but we should focus in concrete clades like for example J2b or perhaps another one, as well as other concrete subclades of other haplogroups wich could be representative of the Eastern Mediterranean side. And it's important not to forget MtDNA.

What I absolutely discard, but it's just my opinion, is the E-M78 connection. This simply diluted and broght West Asian and Southwest Asian due to the migration through the Middle East and Anatolia, and diluted again with the indigenous Southern Europeans while entering Greece, Macedonia and the Balkans. So quite of the different I subclades must be linked, some R1b's, clades like J2b (you are free to think in another one similar), and MtDNA haplogroups in my opinion. The Catalan example is very obvious, because there isn't probably any measurable E-M78 between them, and the other E's appear in the Northwest/East African admixture. Very difficult to match this with the other explanation.

Also, I have checked the Balkan_D portrait (30% Mediterranean average), and several individuals from unknown regions show more than 35% Mediterranean, and there is one with more than 40%. I'm willing to contact the Bosnian girl, since I'm pretty sure she has a very similar profile comparing with the last one.

PD: DOD747 is the Balkan with more Mediterranean (44.1%). Still thinking I1 and I2a2 have nothing to do here considering the haplogroup frequencies? I highly doubt it.
 
i think that this mediterranean admixture can correspond to south west mediterranean genes, given the fact that Sardinia the peak, and secondly the area around Catalunia and Valencia region has the highest part of this genes, but then spain should have the same quantity as catalunia no?.. and why Turkey and south est europe are somewhat high in this genes?
 
Because as, I said, it includes Southwest and Southeast Europe, it's not only Southwestern. For this reason Greece has a significant amount.

And I can assure you Sardinia is not the most Southwestern population. These are probably ethnic Catalans, although there are surely Navarrans, Aragonese...even Gascons or other Southern French who can get into the category. Sardinians probably have significant Eastern Mediterranean in comparison with Iberians, but not at the same level as Greece, even probably less than most Italians.
 
i agree it could be a west plus east southern european genes
but
those it in the "south west part" of the map are sardineans, sardinieans are as western as spaniards, but more on the "southern part" of the map
Europegenetics.jpg
 

This thread has been viewed 104701 times.

Back
Top