June 10, 2014

The Mediterranean route into Europe (Paschou et al. 2014)

An interesting new (open access) paper in PNAS includes some new data from Crete, the Dodecanese, Cappadocia, and several other Greek (and a few non-Greek) populations, and proposes that the Neolithic followed an island-hopping migration into Europe. This is a study on modern populations that nicely complements the recent ancient mtDNA paper from PPNB which found an affinity to Neolithic Near Eastern populations among the modern inhabitants of Cyprus and Crete.

It is hard to imagine that there were ever any major impediments to gene flow between Anatolia and the Balkans as the Aegean islands and Hellespont are not formidable barriers to any culture with even rudimentary technology. Hopefully in the future it will become possible to look at ancient DNA from Greece and Anatolia and directly determine how the transfer of the Neolithic package into Europe took place and how much of the ancestry of modern populations stems from the Neolithic inhabitants vs. more recent shuffling of genes in either direction.

The authors also computed f3-statistics to see if populations were admixed, but found no significant evidence for it. If, for example, Dodecanesians were intermediate between mainland Greece and Anatolia they might have a negative f3(Dodecanesian; Cappadocia, Peloponnese) statistic. A negative statistic proves admixture but a positive one does not disprove it, but, in any case, there is no signal of admixture here so the results are compatible with the authors' model and probably incompatible with a recent admixture that would leave a significant negative signal (i.e., Dodecanesians/Cretans would have intermediate allele frequencies between Cappadocians and mainland Greeks).

PNAS doi: 10.1073/pnas.1320811111

Maritime route of colonization of Europe

Peristera Paschou et al.

The Neolithic populations, which colonized Europe approximately 9,000 y ago, presumably migrated from Near East to Anatolia and from there to Central Europe through Thrace and the Balkans. An alternative route would have been island hopping across the Southern European coast. To test this hypothesis, we analyzed genome-wide DNA polymorphisms on populations bordering the Mediterranean coast and from Anatolia and mainland Europe. We observe a striking structure correlating genes with geography around the Mediterranean Sea with characteristic east to west clines of gene flow. Using population network analysis, we also find that the gene flow from Anatolia to Europe was through Dodecanese, Crete, and the Southern European coast, compatible with the hypothesis that a maritime coastal route was mainly used for the migration of Neolithic farmers to Europe.

Link

24 comments:

Maju said...

"It is hard to imagine that there were ever any major impediments to gene flow between Anatolia and the Balkans as the Aegean islands and Hellespont are not formidable barriers to any culture with even rudimentary technology".

True in pure theory. But we do totally lack so far material information on any Neolithic culture in the Western half of Anatolia Peninsula that pre-dates Thessalian Neolithic, and no East Anatolian (Central-South Turkish) really approximates the cultural traits that we see in Thessaly either. So in any case the Thessalian Neolithic (and by extension all the mainline European Neolithic) is a peculiar culture with unclear roots.

Judging on genetics, there is clearly a West Asian component to it, with some NE African elements as well, but an aboriginal Balcanic element as well. The exact origin of the West Asian influences remains to be solved and a coastal route would explain many things.

Of course it is also possible that we just don't yet know the "missing link" in Anatolia but then again Cardium Pottery peoples were clearly good at sailing so a coastal and even open seas route by specialist sailor-farmers makes good sense.

Rob said...

About time ! Greece has been (surprisingly) underrepresented in studies

Santosh Rajan said...

Could the coastal route explain the predominance of R1b in western europe? Or was the timing too early for the spread of R1b?

bau said...

Again, no data shared. What a disappontment! Scientists will never learn!

eurologist said...

"It is hard to imagine that there were ever any major impediments to gene flow between Anatolia and the Balkans as the Aegean islands and Hellespont are not formidable barriers to any culture with even rudimentary technology."

I am glad to see that you agree with this notion, which I have promoted for some time now, based on geographic barriers and the lack thereof during a long time interval from before LGM until just before the beginning of the Neolithic. I date this to the initial impact of the Gravettian and the post-LGM Epi-Gravettian, which shares significant features in the wider region, but has marked differences with contemporaneous cultures in the remainder of Europe (mainly, the Magdalenian).

I'd also include the Bosporus as a non-barrier until it re-opened during relative recent times, after the spread of the Neolithic to Europe.

Unknown said...

Do you have the link to the article?

Unknown said...

This also complements Dienekes' post of 5th December 2013, in which he stated:

"When, and by what route, did farming first reach Europe? . . . New dates from Franchthi Cave in southern Greece . . . suggests that farming spread to south-eastern
Europe by a number of different routes, including potentially a maritime, island-
hopping connection across the Aegean Sea. . . ."

I commented then:

That farming spread by "island-hopping" should have been predictable. Few researchers in this field remark on the extreme vulnerability of necessarily-largely-isolated farming households to predation by wandering bands of thieves and brigands. Livestock and stored crops can easily be stolen or destroyed. The first requirement for farming is continuing peace. Who would want to sow crops today in the contested areas of Syria? Such peace is much more readily found on a remote island, rather than on any mainland site. This logic can reasonably be extended to larger islands, such as Cyprus, Crete and possibly Sardinia -- on which local warlords could have held sway with relatively little prospect of disturbance from mainland aggressors. It is no surprise that we see the first signs of advanced civilisation on such islands. Island-hopping (for the spread of farming) should be considered elsewhere, e.g. in the Far East. Also, some kind of 'island' must have provided the security, over generations, that enabled farming to initially develop.

Dean said...

"About time ! Greece has been(surprisingly) underrepresented in studies"

It's good to finally see central Peloponnese (Tripoli) get represented in a study. From what I can tell, this region clusters with other Greek sites in the study. The region may have "marginal" admixture (according to supplemental data), but its closest genetic neighbors are Greeks, other south Balkanites and Italians.

andrew said...

@ Santosh Rajan

"Could the coastal route explain the predominance of R1b in western europe? Or was the timing too early for the spread of R1b?"

Short answer: the maritime migration discussed in this blog post and in the Pre Pottery Neolithic B paper (ca. 8000 BCE) that came out a few days ago, is too early for the spread of R1b, and the route by which R1b arrived in Europe is not well established.

R1b has so far been absent in first wave Neolithic sites in Europe from both the LBK Neolithic and the Cardium Pottery Neolithic, where the modal Y-DNA haplogroup is G. And, all available ancient Y-DNA from the European Mesolithic (aka Epipaleolithic) is haplogroup I, although admittedly this is patchy data set that could miss regional variation that largely omits the areas from which R1b expanded (e.g. SW Iberia). There is Upper Paleolithic ancient Y-DNA R* near the Altai (Ma'alta boy ca. 24,000 years ago). Y-DNA mutation dating, while inexact and ill calibrated, points to a split between R1a and R1b around 24,000 years ago. There are several dozen Copper and Bronze Age examples across multiple archaeological cultures of ancient R1a from Central Europe to the Tarim Basin - including in formerly LBK Neolithic areas where R1a was previously absent. A recent Southern Siberian sample also reveals ancient Y-DNA haplogroup Q in the same region in the Bronze Age. The oldest published ancient Y-DNA R1b is from Germany in a Bell Beaker individual and dates to ca. 2100-2800 BCE, although I have heard a rumor of an unpublished find by Russian archaeologists of R1b in Central Asia in an Androsevo culture individual.

R1b's expansion, sometime after the earliest European Neolithic ca. 4500 BCE, and sometime before Bronze Age collapse ca. 1200 BCE, was very rapid. One recent study estimates that the main explosion in this haplogroup (by a factor of 1000 or more) took place over just 60-300 years from a very small effective founding population. It could be that R1b arrived in Europe during the Mesolithic but was regionally limited and did not hitch a ride with a dramatic population expansion until much later sometime in the Middle Neolithic to early Bronze Age. Or, it could be the R1b arrived with a subsequent wave Neolithic/Copper Age/early Bronze Age migration and expanded immediately upon arrival (realistically, the more likely scenario).

The only archaeological cultures that make sense for the R1b expansion are the Atlantic Megalithic culture and the Bell Beaker culture whose territories heavily overlap. Population growth was greater and more consistent for the entire R1b area of Europe in the earlier Atlantic Megalithic culture (from which we have essentially no ancient Y-DNA), but often we think of the Atlantic Megalithic culture as a first wave Neolithic culture (albeit distinct from the truly first wave LBK and Cardium Pottery) which we know from ancient DNA to be unlikely source of R1b expansion. Bell Beaker is a better fit from a timing perspective, and has an early ancient Y-DNA match to R1b, but had not been thought by archaeologists to have been an archaeological culture with a demic impact intense and dramatic enough to account for R1b's predominance in Western Europe today.

A Bell Beaker R1b expansion would probably imply a source immediately prior to arrival in Western Europe from Bohemia in the Czech Republic (although the trail gets fuzzier from there to points further East that are an ultimate source of European R1b) mostly via Southern Portugal which seems to ahve been the epicenter and source of the Bell Beaker culture's Western European dispersal.

Grey said...

"Bell Beaker is a better fit from a timing perspective, and has an early ancient Y-DNA match to R1b, but had not been thought by archaeologists to have been an archaeological culture with a demic impact intense and dramatic enough to account for R1b's predominance in Western Europe today."

If the Atlantic ecozone was relatively unsuitable for the neolithic farming package with the farmers largely restricted to the coasts and supplementing their food production with fish then a relatively small number of miner / metal smiths / traders looking for gold, silver and copper who had an advantage in that ecozone e.g. something related to cattle, could have had a very disproportionate impact on the inland demographics.

http://forwhattheywereweare.blogspot.co.uk/2014/02/neolithic-peoples-from-britain-and.html

"These new findings, together with archaeozoological and human skeletal collagen bulk stable carbon isotope proxies, unequivocally confirm rejection of marine resources by early farmers coinciding with the adoption of intensive dairy farming."

So maybe Bell Beaker brought lactose tolerance or better adapted cattle or something connected to cattle and this allowed them to expand dramatically into the hitherto unclaimed (by farmers) inland regions along the Atlantic coast.

So maybe in the LBK and Cardium zones where the neolithic package produced a larger population the Bell Beaker were a minority of traders and metal smiths but in the unclaimed Atlantic zone they expanded to become the bulk of the population?

Mark D said...

Andrew, there has in the past been some discussion of R1b expanding into Europe from southern Iberia, having transited across littoral North Africa (and leaving some remnants in remote areas of Algeria)"Celtic from the West" if you will. Your thoughts?

Grey said...

"It is hard to imagine that there were ever any major impediments to gene flow between Anatolia and the Balkans as the Aegean islands and Hellespont are not formidable barriers to any culture with even rudimentary technology".

I'm not sure this is true.

There are two kinds of barrier: geographical barriers and other people.

If the regions around the Bosphoros and Black Sea when the sea levels were lower used to be swamps with relatively high forager population density then you have a significant barrier.

If correct that would explain why contacts between the Balkans and Anatolia / Levant were by sea to avoid all the ragin' cajuns in between.

Alogo said...

Pretty cool to see an Eastern Rumelian and Cappadocian sample. I was worried that the older, unadmixed generations would have died out by the time people got around to collecting data (same for Macedonia to a degree).

Alogo said...

Pretty cool that they sampled Eastern Rumelians and Cappadocians. I was worried that the older, unadmixed generations would have died by the time they got around to testing them (same for Macedonia to a degree).

Are there any other data from those two Greek populations?

Rob said...

@ Andrew
"Bell Beaker is a better fit from a timing perspective, and has an early ancient Y-DNA match to R1b, but had not been thought by archaeologists to have been an archaeological culture with a demic impact intense and dramatic enough to account for R1b's predominance in Western Europe today. "

On the contrary, it might indeed have a great impact. The BB period has been described by some scholars as one of unprecedented and unconstrained genetic interchange amongst communities in W and C Europe. To be sure, this differs from simplistic scenarios of 'waves of invaders' from set focal points, much less 'ethnic cleansing'; obviously, for the better.

Maju said...

"The BB period has been described by some scholars [who?] as one of unprecedented and unconstrained genetic interchange amongst communities in W and C Europe".

The fact is that, so far, we know only very little not just about the BB period itself genetics-wise but critically about the pre-BB Chalcolithic (Megalithism and such) and even the post-BB Early and Middle Bronze Age.

Basically, even where there is more info such as in Northern Germany, we know of the First Neolithic, Late Chalcolithic and Late Bronze Age. With more than 1000 years between dates.

The Bell Beaker is interesting but mostly it is an extension of what preceded it, namely the Megalithic culture(s) and the Kurgan expansion in Central and Northern Europe. And of all these phenomena, it is clear that it was the Megalithic wave (associated to the Atlantic expansion of Neolithic and, in Northern Europe, partly to the Funnelbeaker phenomenon), which had the greatest demographic impact by far, according to recent high quality estimates of archaeological base.

BB seems more like an echo of this Megalithic influence, which, as such is not researched.

Also by European regions we have a very unequal amount of data. Germany, Hungary, Sweden and the Basque Country are rather well researched but France, Britain, Denmark, much of Iberia and many other places (IDK: Greece or Russia/Ukraine would be interesting)... much more poorly or nothing at all.

This is because most of the research has been done by certain universities with easy access to certain remains. Someone should ring the Sorbonne and Oxford, for example...

Rob said...

@ Maju , et al

Whatever the case, if what the majority of commentators, professionals and amateurs alike are correct in their current belief that R1b appears to have been a late or even post-Neolithic 'explosion' on most of western Europe, then the key to understanding it is the apparent population collapse (seen in many archaeological studies) which followed the initial Neolithic "boom". The ensuing low Ne, bottlenecking, and new lineage foundering must explain the precipitous rise of R1b in W.E.

Simon_W said...

I found it odd that the authors test the hypothesis of a Mediterranean coastal route vs. the route through the Balkans and Thrace, as if these were equivalent alternatives, when it's beyond doubt that the Neolithic spread along both routes, via Starcevo-Körös-Cris to the LBK on the one hand and with the Cardium pottery on the other hand. Archeologically there is no doubt that the origin of the LBK lies in the Danubian area and in the older Neolithic cultures in the Balkans.

Also I want to point out that their Italian sample is obviously North Italian, presumably the usual HGDP North Italians from Bergamo. In the map of their samples it's misleadingly placed near Rome, apparently suggesting it was a mixed Italian sample, but the analyses make it clear that it cannot be but North Italian. That's important to bear in mind because in the analyses the Tuscan sample is closer to Cappadocia than the Italian sample. Someone who believes that the Italian sample is overall Italian could draw the wrong conclusion that the Tuscans are more Anatolian than other Italians and unfortunately connect this with Herodotus' account of Etruscan origins.

@ Maju

You have to be careful with the designation West Asian. When used with a genetic meaning it usually refers to a component that is modal in the Caucasus and in Balochistan, but suspiciously absent in the early European farmers. The early European farmers deviate from Western European hunter-gatherers in the direction of Saudis and Bedouins - both of which are rather Southern Caucasoid than West Asian, in the above sense.

@ bau

Huh? I managed to download the paper for free.

@ the R1b topic

In my opinion, the spread of R1b in Europe was associated with the spread of some mutually related autosomal components: The Gedrosia component from Dodecad K12b, the Indo-Iranian component from MDLP World-22, and some ANE-admixture from Lazaridis et al. Hence it is very unlikely that R1b reached Europe via Northern Africa or that its western European expansion originated in the Iberian peninsula – because these autosomal components are not particularly strong in these places. I think the key is to avoid the treatment of the Bell Beaker culture as a simple, homogenous block that expanded from one single area. Both archeology (the existence of various local Bell Beaker groups using own unique pottery, the Corded Ware influences, the Vucedol influences, the spread of pottery styles not just along a southwest-northeast line, but also east-west etc.) and physical anthropology (tooth morphology, cranial measurements, isotopic analysis) suggest that it was more complex than a simple out of Iberia expansion. What expanded from Iberia was the original Bell Beaker package and the ideology behind it. It's quite possible that these Maritime Bell Beaker people from Portugal also had some R1b, maybe even some Indo-European language (related with Tartessian??) but they were hardly the main source.

Maju said...

Sorry Simon but West Asia is a quite clear region between the Caucasus and Yemen, between Anatolia and Iran (or if you wish Afghanistan but hardly Pakistan). I don't know where you get the idea that "West Asia" applies to that particular Caucaso-Baloch component, which is, as you say not representative of all the region.

In fact West Asia seems to have two main areas in terms of autosomal genetics: a northern or "highlander" one and a southern or "lowlander" one. They largely correspond to the distribution of Y-DNA haplogroups J2 and J1.

The Caucaso-Baloch component is also known as Ancient North Indian (ANI), even if it seems apparent by now that it is in essence a Neolithic inflow from the NE parts of West Asia.

I do agree that European early farmers seem to have a signature pointing to Palestine rather than, as was usually believed, to Anatolia. This "Palestinian" signature in turn points to some NE African minor but significant admixture, very possibly in the Mesolithic (Natufian, PPNA) and, together with a stronger affinity of Cypriots than Turks with EEFs, to a plausible coastal route from the Levant to Thessaly (to be confirmed however).

I also agree with you on the spread of Neolithic from Thessaly being equally important by both routes: the Northern Mediterranean and the Balcano-Danubian ones. It seems also apparent by now that Central Europe suffered demic replacement processes (from West and East Europe) that were not that active in other regions, hence that Southern and particularly SW Europeans are more similar to ancient farmers. But there was a time when Central and Northern Europeans were as well.

Simon_W said...

Maju, I was aware of the geographical meaning of West Asia, that it also includes the Arabian peninsula and hardly Pakistan. I was rather arguing in favour of a seperate, genetical meaning of the word. The bunch of related ADMIXTURE components I was thinking of were most often called West Asian in the Dodecad experiments. That was the case in Dodecad K7b, Globe10, Globe12 and Globe13, but also in earlier runs. A notable exception being the World9 analysis, where it was called Caucasus_Gedrosia. The more southern components were called Southwest Asian or Red Sea, depending on their distribution. Vadim Verenich from the MDLP also called the more northern component West Asian, in his World-22 analysis, the more southern one was called Near Eastern.

Of course names are debatable, and we can envision alternatives like Caucaso-Baloch as you suggested, or Highland West Asian, or more.

But in order to avoid misunderstandings and the generation of confusion an uniform terminology would be preferable.

Maju said...

I don't think you'll get uniformity from distorting the conventions of geography, rather the opposite. It's better to use more words and make sure what the other may mean.

Anyhow, there is no such thing as standardized autosomal components in my opinion and I personally disagree with Dienekes in his standardization or rather 'zombification' process. It can be a useful approach but it is a particular and unavoidably arbitrary dissection. I prefer to run ADMIXTURE freely, with careful sampling strategies (this is most important), and see which is the cross-validation optimal score for each comparison (other K values may be meaningless and misleading).

Simon_W said...

Note that I wrote of a whole bunch of components that I would call West Asian. The Dodecad K7b West Asian component isn't identical with the Globe10 West Asian component, etc. But there can be little doubt that they are related and similar. So there is no need to arbitrarily take one component as a standard. And even if we use ADMIXTURE always freely, there seem to appear similar components again and again (though not always at the same K), which seems to imply that there is some objective pattern in the genetic variation. But agreed, questions about the validity of a component, if it's a blend of other components, are important. Cross-validation is surely a useful technique, and there are also other means and hints.

Pneumatikon said...

Cretan Texan here. J2a Y-Chromosome.

I loved this. BUT THEY NEED TO STOP INSISTING WE ARE THE ETRUSCANS!! J2a spikes in Lazio, Umbria, and The Marches - the Papal States circa 1860.

We Anatolians were not the Etruscans. We were the ROMANS. The fact the Tuscany and Crete are so close as it is tells me we most certainly are.

Pneumatikon said...

Simon W's comment made me look deeper into the data:

"Also I want to point out that their Italian sample is obviously North Italian, presumably the usual HGDP North Italians from Bergamo. In the map of their samples it's misleadingly placed near Rome, apparently suggesting it was a mixed Italian sample, but the analyses make it clear that it cannot be but North Italian."

I've been convinced for quite a while that the Romans were from Crete - ultimately, Anatolia - and these "If the Anatolians are anything they are the Etruscans" ideas drive me up the wall. Yet here they are. Confirmed.

NOT.

The historical record is clear: Romans moved out of Rome into conquered territory. Slaves were moved into Rome and replaced them.

Florence is one of those cities the Romans moved to.

So Florence is actually a ROMAN sample, and Rome is actually a NORTH ITALIAN sample.

Oh my God! I was right! Check out that dark red dot representing Crete on the gene network samples. They really are the Romans!!! And they moved into Italy right before the Bronze Age Collapse in MUCH larger numbers than I thought. Samples from further south of Rome in the rest of Lazio, in Umbria, and in The Marches - all strong Italic territories - would look much more like Florence.

Oh. My. God. Look at the map of the Italic speakers! Look at them down in Sicily, too!

http://en.wikipedia.org/wiki/Italic_languages

You know what that means, don't you? If I'm right that means the Minoans spoke proto-Italic. Otherwise they would have been speaking Greek.