search this blog

Wednesday, October 28, 2015

Prehistoric African admixture in Iberia

Open access at PLoS ONE:

Determining the timing, identity and direction of migrations in the Mediterranean Basin, the role of “migratory routes” in and among regions of Africa, Europe and Asia, and the effects of sex-specific behaviors of population movements have important implications for our understanding of the present human genetic diversity. A crucial component of the Mediterranean world is its westernmost region. Clear features of transcontinental ancient contacts between North African and Iberian populations surrounding the maritime region of Gibraltar Strait have been identified from archeological data. The attempt to discern origin and dates of migration between close geographically related regions has been a challenge in the field of uniparental-based population genetics. Mitochondrial DNA (mtDNA) studies have been focused on surveying the H1, H3 and V lineages when trying to ascertain north-south migrations, and U6 and L in the opposite direction, assuming that those lineages are good proxies for the ancestry of each side of the Mediterranean. To this end, in the present work we have screened entire mtDNA sequences belonging to U6, M1 and L haplogroups in Andalusians—from Huelva and Granada provinces—and Moroccan Berbers. We present here pioneer data and interpretations on the role of NW Africa and the Iberian Peninsula regarding the time of origin, number of founders and expansion directions of these specific markers. The estimated entrance of the North African U6 lineages into Iberia at 10 ky correlates well with other L African clades, indicating that U6 and some L lineages moved together from Africa to Iberia in the Early Holocene. Still, founder analysis highlights that the high sharing of lineages between North Africa and Iberia results from a complex process continued through time, impairing simplistic interpretations. In particular, our work supports the existence of an ancient, frequently denied, bridge connecting the Maghreb and Andalusia.

Hernández CL, Soares P, Dugoujon JM, Novelletto A, Rodríguez JN, Rito T, et al. (2015) Early Holocenic and Historic mtDNA African Signatures in the Iberian Peninsula: The Andalusian Region as a Paradigm. PLoS ONE 10(10): e0139784. doi:10.1371/journal.pone.0139784


bellbeakerblogger said...

I believe it was Ontoni (?) & company that actually theorized the majority of Western Euro H1, H3 and V came from Holocene North Africa?

The papers on this are fun to read though. The theoretical battle lines are drawn.

In any case I'm counting the days that Saharan Steppe Herder genomes of the Middle Pastoral come out. Maybe people will start to get interested in prehistoric North Africa.

By the way, I'd like to know from the Yamna nerds where the lithics and pottery of Beakers come from? (Wink)

Davidski said...

Have you seen this?

bellbeakerblogger said...

Very smart men, thanks for linking.

I think this is a convenient narrative that is great on big things but has problems with the conflicting details; and is ignoring the signals of Iberia and North Africa. Broadly, Beaker has many cultural and genetic similarity with Pontic steppe, but the fact remains that very little to the left of Corded Ware looks like an earlier migration that later turned into Beaker.

Also, almost all Beakers were hybrids with Corded Ware in the N and S continental zone so the fact that they are shifted against Corded Ware should be an omen for things to come.

Again, a broader question concerns the ancestry and flow of Volga woodsmen. From Karelia to Samara I view them as immigrants, many from the southern zone. So it's likely a number of population are Yamna like, but had different paths.

bellbeakerblogger said...

There's one more aspectId like to add.
The overwhelming majority of my American ancestors came to America on average around 300 years ago, many in the early 1600's, some as late as the Revolutionary war. Despite having lived much of that time invery close proximity to others, I have a whopping 3% Amerindian admixture and effectively 0% African using various calculators ( including Eurogenes)

The point being that everyone assumes certain things about North Africa that are as of yet unfounded. 4000 years ago Algeria was a different place. These mtdna lineages tell us a little about that destroyed population. The entire circum-Saharan contains relicts of it.

Anonymous said...

This is what they propose apart from the more recents events (Roman era, Arab conquests): “Thus, the most parsimonious model for the oldest demographic events and migrations across the Mediterranean Basin, based on mtDNA evidence, is the following: after 20 ky, U6 lineages had an extensive population expansion in northwest Africa, associated with the emergence of the Iberomaurusian industry in the Maghreb; this pool was further enriched by sub-Saharan L lineages, especially L1b, which began to arrive in North Africa in the beginning of the African Humid Period (AHP, ~11–5.5 ky BP) [57]; U6 and L lineages were introduced from northwest Africa into Iberia in the post-glacial period, most probably by the time of the Younger Dryas/beginning of the Holocene.”

Alberto said...


The point being that everyone assumes certain things about North Africa that are as of yet unfounded. 4000 years ago Algeria was a different place.

Indeed, with archaeology the situation is similar as the one we have now with ancient DNA: there are too many blank spots (deserts are certainly a problem) and we're building our theories based on the available data without counting on the missing one. I understand it's difficult to count on missing data, but it's also wise to not forget about it.

Some serious investigation in the north of Africa would probably yield interesting results. In Iberia there is a rather mysterious first appearance of Neolithic culture (La Almagra) in the south (Andalusia), predating the arrival of Cardium Pottery in the NE by at least a few centuries.

Nirjhar007 said...

I remember i one suggested that BBC probably has N African origins, concerning also the apparent Afro-Asiatic type of Substratum in Celtic and Basque etc but we need some aDNA from there :)..

Maju said...

Interesting although I suspect that the actual age estimates are off by as much as 100%, i.e. the most likely time of genetic exchange between Iberia and NW Africa was at the Oranian genesis, c. 20 Ka BP.

@BBB: unlike what happens with U6, M1 and L(xM,N) lineages, which seem to have migrated (mostly) from NW Africa to Europe, mtDNA H clearly follows the opposite pattern as demonstrated by Cherni 2008. In general terms the diversity of macro-haplogroup H in Africa is limited, unlike what happens in Europe (and West Asia also at least to some extent). Hence hypothesizing as yo do with an African origin of European mtDNA H is, I must emphasize it, total nonsense: mtDNA H in NW Africa has a clear European origin in all meaningful cases (and this European origin is South Iberian except for H7 which may be French). What has been said about H is probably also applied to V. The frequency of European mtDNA in NW Africa (~30%) correlates well with the European autosomal ancestry in the region. The low frequencies of European Y-DNA probably respond to sex-biased Afroasiatic expansion beginning with Capsian culture (Epipaleolithic and Neolithic), however the Guanche mummies show greater frequencies of R1b and especially I, strongly suggesting that these were much more common in the past.

Romulus said...

Grimaldi man is obvious proof of North African admixture in prehistoric Europe, could even explan dark skin in WHG. mtDNA H1 & H3 I don't believe though, the signal doesn't show up till bell beaker, Neolithic farmer mtDNA contains a different balance.

Romulus said...

Also I should add, in a lot of the runs Genetiker has done on ancient samples have shown bits of African admixture. That was dismissed as noise, but, it could be accurate.

Chris Davies said...

Interesting paper. In relation to this, does anyone know the approximate age of E-M81 Y-DNA in Cantabrians?

bellbeakerblogger said...


I tend to get bugged-eyed on a few hot button issues. It drives me nuts that there is little interest to falsify, disqualify or seriously consider North Africa by mainstream archaeology when there are clear signals coming from that direction beginning in 3,600 B.C. and into the 3rd mil.

I'm just surprised by the ambivalence. If found this interesting, an idea of just how large the area is:



Let me correct myself, Ottoni et al (2010) saw H1 coming from Iberia to Africa 8,000 to 9,000 years ago followed by separation. This follow Achilli et al (2004) who suggested H1, H3 and V were Iberian refuge lineages that expanded in the Pleistocene into North Africa.

It was Ennafaa et al 2009 who contradicted this by saying that the age required exchange of H1, H3 and V, much more recent, but that there was a barrier only in the male lineages only (assuming that Berbers and Western Europeans accurately represent the y-chromosome landscape of prehistory, which they do assume)

I'll refer to Loogvali et al (2014) for the age and the structure of Haplogroup H. It's a very well writen paper in my opinion and I think it makes the directionality of Onotoni and Achilli nearly impossible.

But here's a more specific question. If Haplogroups H1, H3 came to North Africa from Iberia, what male haplogroups characterized that population?

I think this is a fair question because either Libyan Berbers are mixed, or Welshmen are mixed, or both, but one is nearly 100% R1b and the other 100% E-whatever.

Anonymous said...


In those same runs Neanderthal and Denisovan come out as almost completely African. It's actually not noise but shared ancestral and thus due to age, or the fact that some samples simply do not share recent drift or mutations.

On Grimaldi-man, we need old examples of Aurignacian Western Europe. More than just Oase 1, who is far more interesting than most people consider.

Grey said...

IIRC Maju has been pointing out for a while that when people talk about African admixture in Iberia they miss a distinction between the bulk of Iberia which has one pattern and the Atlantic coast having a different pattern.

The pattern along the Atlantic coast might be from the megalithic culture time or maybe much earlier.

This might add another sedimentary layer to some populations.

(Although megalith culture must be default option it's always seemed likely to me that even if the main flow OoA was via Arabia / Mid East there could have been smaller flows up the Atlantic coast from NW Africa.)

Krefter said...

"Also, almost all Beakers were hybrids with Corded Ware in the N and S continental zone so the fact that they are shifted against Corded Ware should be an omen for things to come."

There's no prove BBC got their ANE from Corded Ware. They could get it from a separate source.

Maju said...

@BBB: R1b and I2 plausibly, both documented along with other worse defined Eurasian lineages among the Guanche mummies. However E1b-M81 was almost certainly already present (its presence in Iberia is tightly associated to North African mtDNA) and this may introduce some distortion, which I don't feel able to quantify.

I won't take seriously any claim based only on age guesstimates. Structure is what matters and for age estimates we must use unquestionable references from other disciplines such as archaeology (or also ancient DNA where available) and not "molecular clock" pseudoscience. (In any case I can't find any 2014 paper led by Loogväli, are you sure you are not referring us to a 2004 one that does exist and deals with mtDNA H?)

As for the map size comparison, sure: Africa is huge but NW Africa, once you remove the desert, is about the size of France.

@Grey: "IIRC Maju has been pointing out for a while that when people talk about African admixture in Iberia they miss a distinction between the bulk of Iberia which has one pattern and the Atlantic coast having a different pattern".

Actually there's very little African admixture except in the Western third (not just the coasts, also inland regions like Extremadura, León...). It's not two patterns: it is a region where it is clearly important (Western third) and another where it is too small to matter, sometimes even zero (Central and Eastern thirds). The presence in Asturias particularly, along with the likely presence of such admixture in La Braña, strongly suggest a Paleolithic origin, namely in the Solutrean period, which is also the period when the Iberomaurusian or Oranian culture formed in NW Africa, probably via Iberian influence.

So IMO there was a sizable Iberian immigration to NW Africa c. 20 Ka BP (still apparent in autosomal and mtDNA but less so in Y-DNA) and an associated backflow to the Western parts of the peninsula (Portugal initially), which also affected Asturias (whose Solutrean facies is related to Portugal and not to Aquitaine, as happens with the rest of the Cantabrian strip). Later there are no opportunities for this kind of exchange before the Neolithic.

bellbeakerblogger said...


"There's no prove BBC got their ANE from Corded Ware. They could get it from a separate source"

Here's what I'm saying which I've been trying to communicate to no avail; that is the nature of Beaker culture in much of the continent appears hybridized, not with generic Steppe pastoralists, but with Corded Ware specifically.

So while we don't know to what degree they mixed, we do know that German Beakers have uniformly dissimilar paternity from German CW and also being less steppe-like.
This leaves one natural conclusion, the original Maritime Beakers(probably R1b) who mixed with CWC were a lot less like Corded Ware than is currently assumed.

Not saying they had zero ANE, but just a much different profile.


It would be nice to test that theory with high resolution y-snps of modern populations, especially those that are common to Europe. No ancient DNA required for that.

Anonymous said...


"This leaves one natural conclusion, the original Maritime Beakers(probably R1b) who mixed with CWC were a lot less like Corded Ware than is currently assumed."

As far as I know the Iberian origin of BB is not considered settled science yet.

Maju said...

@Epoch: C-14, C-14, C-14! That "paper" (dissertation) has not a single date, hence it can only be considered ranting, open speculation without any sort of scientific basis. Also what happened to the Portuguese "copos" as possible origin of the BB design?

Even if we'd accept (why?) that Vucedol was at the origin, it'd be impossible to argue that the spread took place via Central Europe and we'd have to look at North Italy or South France instead (or Iberia too, why not?) as the key spots of transformation of the precursor into true BB, which is what really matters. The chronology is certainly 300 years older in the South of Europe (mostly Iberia but not only).

Anonymous said...


Yes. C-14. But I heard there may be issues there. The C-14 calibration might be off as a number of the oldest sites may suffer from the so called "reservoir effects" (Dixit Alan from anthrogenica). Furthermore, the older dates are recalibrations. If you look at the original Müller/van Willingen (2001) paper they need some math to come to these dates (i.e. not all items on a site dated similarly).

It may not be that straight forward.

Maju said...

@Epoch: not persuaded (sounds like clinging to burning nails without nothing specific) but anyhow... something that truly frustrates me is that, for all those pan-European academic programs like Erasmus, the availability of pan-European prehistorical syntheses and databases is today almost as bad as 10 or 15 years ago, at the beginning of the Internet era. Universities and national academias are still pretty much autistic, so it is really difficult to understand these periods unless you are a dedicated specialist (and even then probably).

Rob said...

The reservoir effect can be catered for, as it has been for Yamnaya dates now all below 3000 BC

This doesn't change that proto-Beaker package originated in Iberia

Simon_W said...

@ Krefter

BR1 from Hungary probably gives good clues about the autosomal structure of the Hungarian Yamnaya. The reason is this: BR1, although dated to c. 2100 BC, is from the Mako culture, which started about 2700 BC, i.e. very shortly after the immigration of Yamnaya. Also, the Mako culture was located on both sides of the Tisza, effectively incorporating the Yamnaya area. Furthermore, there is no evidence that there was a later immigration from somewhere else that gave rise to Mako. So Mako must be some mix between Hungarian Yamnaya and local farmers. We don't know the mixing proportions, but we know what the local farmers were like, hence we can infer what the admixing newcomers must have been like. Actually RISE 479 from central Hungary (Vatya culture), who lived about 400 years after BR1, was even a little less farmer-like than BR1, so may give the better clues.

And what were they like? In Eurogenes K15 they had a rather strong Baltic component (32% in RISE 479), consistent with quite a strong WHG admixture, and no West_Asian component (0%). In Dodecad K12b they had no Gedrosia (0%). The latter two observations are consistent with the unsupervised ADMIXTURE run in Allentoft et al., where they have very low teal, almost none.

Needless to say that this autosomal structure is very different from the Corded Ware structure, which had much stronger North_Sea and East_Euro components and less Baltic. And the Corded people had extremely strong Gedrosia and also considerable West_Asian. The Bell Beaker people from Germany were less steppic than the Corded people, but they did have a fair share of West_Asian and Gedrosia, a rather limited Baltic component, and considerable North_Sea and Atlantic components. And of course modern northwest Europeans also have plenty of North_Sea and Atlantic, and also more Gedrosia than any other Europeans.

The consequence can only be: Hungarian Yamnaya, if it was in any meaningful way related with the immediately subsequent Hungarian cultures, what seems very likely, cannot have been the main source of ANE for German Bell Beakers and modern Northwest Europeans.

And there is no other, more plausible candidate than the Corded Ware.

Simon_W said...

That Harrison & Heyd paper linked by David is great stuff. A shame I didn't know it, although it's already been available since 2007. It should be compulsory reading for anyone interested in R1b-M269 and the Bell Beaker phenomenon.

It took me several hours to read it in detail, but it's worth the effort. Although strenuous with its obsession with archaeological details it's useful to read the whole paper, and not just the discussion at the end.

This convinced me that the R1b in the eastern Bell Beakers is probably from the Hungarian Yamnaya and not from the Corded Ware. It's just a matter of adding up 1 + 1.

Simon_W said...

Having said this, I still believe that a lot of their autosomal DNA was from Corded people. This isn't hard to grasp: No matter what the Corded Ware mtDNA was like, their women were not appreciably more EEF-like than the Corded Ware males. No wonder, because of the recombination. Shove aside the males, replace them with R1b bearers, and you have your eastern Bell Beaker people (after having added some more farmer admixture).

Simon_W said...

It's nice to see the proto-Beaker package that expanded from Portugal concisely summed up. And it's not hard to imagine that this was most of all an attractive fashion; no need to postulate the expansion of a Bell Beaker folk. And much less a linguistic conformity across the entire complex.

As for the Taurid brachycephalic Bell Beaker cranial type, I'm not sure if that can be tied to the Yamnaya migration as well. I might be wrong, but as far as I know this type was less common in Bohemia and Hungary, i.e. where the Yamnaya autosomal impact was probably strongest. And it was very common in the Rhine group, of which we don't know if it was R1b dominated as well. Also, I've never heard that such a type was common in Yamnaya graves in Hungary.

Simon_W said...

@ bellbeakerblogger

The paper is full of details and in no way ignores signals from Iberia!