Saturday, November 27, 2010

Next Post on Tuesday

I'm taking a three day break.  My next post will be on Tuesday.

Rivers, Lakes and Mountains

Khor Virap Monastery with Mount Ararat
and the Fertile Valley of the Aras River in the Background

Update (November 2012):  This post of mine is extremely popular.  I wrote it over two years ago before I was able to aggregate a more complex picture of European ice age refuges.  I still do think that Anatolia has been a refuge for European populations.  I also continue to think that Europe was partially repopulated from Anatolia at the end of the last Ice Age.  However, the picture is much more complex than one would be led to believe from thinking that the population of Europe is comprised mostly of "farmers from the Fertile crescent."  There are a number of other ice age refuges in Europe that contribute to the modern population diversity of Europe.  It is a quite complex picture over time and space.  Suffice it to say that this post is a good description of the dynamic with the Anatolian ice age refuge.  However, then are many other European refuges and even an interaction with North Africa.

Update (April 2012):  The recent paper "Neolithic patrilineal signals indicate that the Armenian plateau was repopulated by agriculturalists" [Herrera, et al, (link)] reveals that the origin of agriculturalists is likely the Tigris and Euphrates River System.  Therefore, the movement of peoples along the river systems described below is from the Tigris and Euphrates into the Aras and Kara River systems in one branch, and in another westward branch, into the Kizilirmak River system.

Original Post (November 2010):

Revisiting the Haak et al paper, I'm going to review a little geography in this post.  During the water scarce Younger Dryas, access to a steady water supply for humans, and for the animals they hunted, would have been of paramount importance.

With that in mind, the genetic distance maps of Figure 3 in the Haak et al paper take on new meaning:

Figure 3A and 3B
"Mapped genetic distances are illustrated between 55 modern Western Eurasian populations and the total of 42 Neolithic LBK samples (A) or the single graveyard of Derenburg (B). Black dots denote the location of modern-day populations used in the analysis. The coloring indicates the degree of similarity of the modern local population(s) with the Neolithic sample set: short distances (greatest similarity) are marked by dark green and long distances (greatest dissimilarity) by orange, with fainter colors in between the extremes. Note that green intervals are scaled by genetic distance values of 0.02, with increasingly larger intervals towards the “orange” end of the scale."

The two maps, A and B, can be viewed as maps of a survival strategy in a water scarce world.

With Map 3A, what accounts for the peculiar Southern Caucasus-Central Anatolia pattern?  After looking at a map of rivers in the region, it isn't difficult to posit that this particular population was following systems of rivers and lakes.  Perhaps they weren't intensionly following rivers, but the animals they were hunting certainly were.

It isn't clear if the direction of migration was from Anatolia to the Caucasus or the Caucasus to Anatolia.  Let's assume the direction is westward.  The Aras and Kura rivers flow along the Southern Caucasus Mountains and connect Georgia to Azerbaijan:

Kura and Aras Rivers

Note that the dark green island in Figure 3A hugs the Kara and Aras Rivers.  But what connects this area to the dark green island in Central Anatolia?  Other river systems, of course!

Map of Rivers and Lakes in Turkey

Heading west again, the jumping off point from the Aras River was none other than the Tigris and Euphrates River system. Testing this idea, it's not surprising that the archaeological site Hallan Cemi Tepesi is located on the Batman River, a tributary of the Upper Euphrates River. Çayönü lies near the Bogazcay, a tributary of the upper Tigris River. Göbekli Tepe is located near the Upper Euphrates River. 

Jumping off from the westernmost source of the Euphrates is the Kizilirmak River. On a tributary of this river, the Melendiz River, Aşıklı Höyük is located. 

The earliest settlement dates for these locations have been dated to the late the Younger Dryas, 10,000 years ago. That puts our proto-LBK Figure 3A population on the Kizilirmak River 10,000 years ago. 

The Sakarya River (Sangarius River), a valley away from the Kizilirmak, would have taken a westward bound Figure 3A hunter-gatherer-farmer to the Bosphorus, just in time for the Holocene.

Looking at Figure 3B from the Haak et al paper, it is clear that this population created a refuge for themselves in the Northern Zagros mountains. They also followed rivers and established themselves on the Upper Tigris River as is evident by the settlements at Nemrik and Qermez Dere, for example. By way of the Aras and Kizilirmak Rivers, this population also reached the Southern Caucasus Mountains and Europe during the Holocene expansion.

Employing this river strategy, both populations were poised to push onward into the upper Danube and Dnieper, as is attested to by the near genetic distances of modern populations in these areas to LBK populations.

Friday, November 26, 2010

Northern Fertile Crescent Hunter Gatherers

Archaeological Sites in Anatolia and the Fertile Crescent

During the Younger Dryas, that thousand year reversal of fortune at the end of the last ice age, weather in Anatolia and the Northern Fertile Crescent became colder and dryer than the preceeding period.  Perhaps for this reason, the development of agriculture had to wait for a warmer and wetter period.

As conditions for farming had not yet developed, Anatolians and Northern Fertile Crescent inhabitants stuck to their hunter-gatherer past.  Evidence for this can be found in the earliest settlements of the Late younger Dryas and early Holocene:

Hallan Çemi Tepesi

"Hallan Cemi, Pig Husbandry, and Post-Pleistocene Adaptations along the Taurus-Zagros Arc (Turkey)", M. Rosenberg, R. Nesbitt, R. W. Redding and P. L Peasnall  (Link)

Abstract:  "Recent work at Hallan Çemi Tepesi and other round house horizon sites in eastern Anatolia indicates that the Taurus-Zagros flanks were a second autochthonous center of neolithization in southwestern Asia.  Fully settled complex hunter-gatherer societies are in existence in this area by the late Younger Dryas.  These settled village societies were based on adaptations that did not involve cereal exploitation, presumably because cereals were absent in this area during the Younger Dryas.  Instead, these adaptations revolved around the exploitation of nuts and pulses, plus the hunting of ovicaprids and deer supplemented by early experiments with animal husbandry involving pigs . . ."

Hallan Çemi Tepesi, Demirkoy, Qermez Dere and M’lefaat

"The role of wild grasses in subsistence and sendentism: new evidence from the Northern Fertile Crescent", Manon Savard, Mark Nesbitt and Martin K. Jones   (Link)

Abstract:  "Sedentism is usually regarded as a pre-condition for the development of crop husbandry in Southwest Asia and, consequently, sedentary pre-agrarian sites are an important focus of research on the origins of agriculture. It is often assumed that wild grasses were as important for huntergatherers as domesticated cereals were for early farmers, and that wild grass exploitation may therefore have had a critical role in enabling sedentism. Results from the analysis of archaeobotanical assemblages from Hallan Cemi, Demirkoy, Qermez Dere and M’lefaat, and comparison with those of other sedentary pre-agrarian sites in Southwest Asia, challenge the role often attributed to the exploitation of grasses at this time. Archaeobotanical and ethnographical evidence instead suggests that hunter-gatherers took an opportunistic approach to the resources available and their subsistence strategies were not necessarily centred on grasses and ‘wild cereals’."


The Hunters of Nemrik, Theya Molleson (Link)

"The faunal evidence indicates a hunting economy. An enormous amount of animal bone was recovered. It was all kitchen waste. Wild animals, especially antelope, predominated over domestic. The fauna comprises some sheep, cattle and pig but mainly antelope. Horse, deer, wild cattle, boar, beaver, badger, buffalo, jackal and panther are present. The people also hunted migrating geese, cranes, bustards as well as resident francolin, chukars and sandgrouse. The presence of birds of prey, notably the eagle owl, hint at falconry. Both bustards and crows, depicted among the sculptures, were well known to the inhabitants of the settlement (Lasota-Moskalewska 1994).

   "Agriculture as a subsistence seems to have been adopted during the late phase c. 9500–9000 BP and grinding stones, rubbers, mortars and pestles are then part of the lithic assemblage. The range of artifacts includes grindstone, polishing plates, awls, bone needles, a whetstone, pounders, shaft straighteners and bolas stones (Kozowski and Aurenche 2005). The subsistence technology can be, to some extent, inferred from the artifacts for pounding grain (mortars), rather than grinding. No pottery has been recorded. The technique of hunting is less clear. There are bolas stones, but few ‘spear’ straighteners. Arrow heads are exceptional and exotic, occurring only in some burials."


Göbekli Tepe

"Paleozoological and paleobotanical studies running parallel to the excavation indicate that the population whose achievements we see at Göbekli Tepe represented an economic stage of development still dependent upon wild prey. The economic motor of the Neolithic village, forerunner of the oriental city, still lay far beyond the horizon. Only a collection of hunters who assembled on the mountain as if to attend an "Olympic council" could have been responsible for the outlay of labor necessary to erect this architecture. "First came the temple, then the city" would seem descriptive of the phenomenon we see here. It remains the role of future excavation either to confirm or discredit this conclusion."

"The most recent building phase at Göbekli Tepe (Level II) has been dated both comparatively and absolutely (C14) to ca 8000 BC, with an earlier primary building phase (Level III) ending as early as 9000 BC. The age of the earliest occupation cannot yet be determined; the depth of the deposit, however, would suggest a period of several millennia, which signifies that the site had already existed in early Paleolithic times."

Aşıklı Höyük

"Project directors, M. Ozbasaran and G. Duru propose that the earliest layer could represent a seasonal settlement, in contrast to the later occupations of Layer 2, and that permanent settlement did not occur at the site until about 8000 cal BC."

"Subsistence:  According to the analyses, game from the hunt and vegetables and fruit collected by the inhabitants constituted the basis sources of diet at Aşıklı. The amount of cultivated einkorn, emmer and durum wheat eaten was minimal, as was also true of the barley and legumes planted by the population. Wild wheat and barley was also reaped, however, and brought into the village, where it was husked. Most popular among the wild fruits were the red hack-berry (Celtis tournefortii)."

"The most frequently consumed wild animals were sheep, goat, pig and cattle. Horse, deer, rabbit and different kinds of birds and fish were also among their consumption list. Animals such as the sheep and goats may have been at a stage of proto-domestication although there is no evidence of truly domestic animals at Aşıklı. Because farming had only recently come into practise, it was mainly wild grain that was being consumed."


These archeological sites demonstrate that hunters had settled in villages during the Younger Dryas. They exploited the abundant game of their surroundings as well as wild berries, nuts and cereals. They were not yet farmers, but they had a varied and substantial diet and they were poised to take advantage of the warmer, wetter weather of the Holocene.

Constellation of Orion, The Hunter
Photography credit: Mouser

Tuesday, November 23, 2010

The Fertile Crescent Transition and Its Implications for the European Neolithic Transition

I've been considering the implications of the three component Fertile Crescent diffusion process.  The process of diffusion and convergence for the three Fertile Crescent components, defined by West Asian, South European and Southwest Asian, implies that the genetic signature of farmers has gradually been changing.  This gradual process of change also has meant that farmers who migrated west into Europe also have had a genetic makeup that had shifted, depending on when they left the Fertile Crescent.

The recent paper "Ancient DNA from European Early Neolithic Farmers Reveals Their Near Eastern Affinities", Haak et al, suggests that early LBK farmers are most like populations of the Northern Fertile Crescent.  From their paper, Figures 3A and 3B present maps showing matrilineal distances for two LBK populations compared modern populations:

"Mapped genetic distances are illustrated between 55 modern Western Eurasian populations and the total of 42 Neolithic LBK samples (A) or the single graveyard of Derenburg (B). Black dots denote the location of modern-day populations used in the analysis. The coloring indicates the degree of similarity of the modern local population(s) with the Neolithic sample set: short distances (greatest similarity) are marked by dark green and long distances (greatest dissimilarity) by orange, with fainter colors in between the extremes. Note that green intervals are scaled by genetic distance values of 0.02, with increasingly larger intervals towards the “orange” end of the scale."
What is evident is that Figure 3B represents an earlier, more genetically distant Fertile Crescent population than Figure 3A. 

If you look carefully at the Figures, you will note that Cyprus is at a genetic distance of 0.4 from the "green island" of Figure B.  By contrast, Cyprus is genetically near the general LBK "green island" of Figure A(genetic distance 0.24). 

Moreover, Sardinia is at a genetic distance of 0.4 from the population at the Derenburg site(Figure B), but only 0.32 from the broader LBK population (Figure A).

The diffusion graphs of the autosomal South European Component indicate a coalescence point in Anatolia and Cyprus and a radial diffusion from that point into the Fertile Crescent.  The West Asian Component peaks in the Caucasus and gradually diffuses southwestward.  The variance of the distributions indicate that the West Asian component is older and more diffuse than the South European Component.

Based on the diffusion results for the K10 Dodecad populations, in combination with the maps generated by Haak et al, I'm going to assert that Early Neolithic Migrators (Haak, Figure3B) from the Fertile Crescent were a people with a primarily West Asian signature.  Seven or eight thousand years ago, the first wave of expansion of Southwest Asian peoples would not have been felt in the Fertile Crescent.  The expansion of South Europeans would also not have been significant.  The Derenburg site likely represents these early migrators.  That is the population mapped in Figure 3B.

In the ensuing millenia, a South European people expanded from a coalescent point in Northeast Anatolia and Cyprus.  We can tell this by the fact that the South European component continues to diffuse from this point today.  Together with the existing West Asian people, these people formed the second wave of Mideast Farmers to migrate northwestward across the Balkans and into the basins of the Danube and Dniester Rivers.  Their path is hinted at in Figure 3A.

This scenario would account for the difference seen in the larger LBK population with that of the Derenburg site.

This represents a refinement to the Demic Diffusion model.  It supports that idea that Near East Neolithic Farmers made a significant contribution to the genepool of Modern Europe.  However, it also suggests that the movement of peoples during the Neolithic was complex and was affected by changes that continued to unfold in the Fertile Crescent.

Monday, November 22, 2010

Fertile Crescent Components do the Wilkins Wakeley Model

I've decided to test how well the distributions for the Fertile Crescent components (West Asian, South European and Southwest Asian components [Dodecad K10 data]) adhere to the Wilkins Wakeley Model.

Essentially, can the geographic distribution of these three components be described with the Normal Distribution?

I've added some outlier populations.  These include the Georgians, Lezgins and Adygei at the northern extent, and the Saudis and Ethiopians at the southern extent.  Here's the stacked bar plot again, with these populations added on each end:

On the x axis: 1=Lezgins, 2=Adygei, 3=Georgians, 4=Armenians, 5=Turks, 6=Assyrians, 7=Cypriots, 8=Syrians, 9=Jordanians, 10=Egyptians, 12=Saudis and 14=Ethiopians.

In order to approximate some degree of geographic proportion, I've inserted two spaces around Saudi Arabia, since it is a large country in comparison to the others.  I now plot a separate graphs for each component and attempt to fit the data with the Normal Distribution:

On the x axis, 0 to 14 are as defined above.  The bottom graph plots the fitted distributions from the preceeding graphs.

What emerges is that a geographic point of coalescence can be discerned for the three components.  For the West Asian component, the point of coalescence appears to be in the Caucasus.  For this data set, which does not graph Sardinia, the Southern European coalescence point is in Cyprus.   The coalescence point for the Southwest Asian component is at a point adjacent to the Southern Red Sea or the Gulf of Aden.

It is also evident that the West Asian component is an older, more diffused component.

The distributions for the Southern European and Southwest Asian components can described with the same variance.  Both components are less diffused and younger than the West Asian component.

Friday, November 19, 2010

Estimating a First Arrival Date for Southwest Asian peoples in the Fertile Crescent

In my previous posts, I've been describing a diffusion of a Southwest Asian component northward over a time base of between five and ten thousand years.  You can infer that process by looking at the current distribution of the Southwest Asian component in Fertile Crescent populations:

The Southwest Asian component appears in light turquoise on the graph.  Not plotted are the Saudis, who have a 76% Southwest Asian component in K=10 Dodecad run.

We've established that it is plausible that Modern Assyrians are representative of inhabitants of the Fertile Cresent from the Middle Assyrian Period.  I'll be specific and say that they are representative of middle northern Fertile Crescent inhabitants from approximately 1500BC.   Using Assyrians as a "snapshot", we've inferred diffusion rates for populations from Saudi Arabia, Cyprus, Egypt and Babylonia.  We've examined the small contribution that far flung traders have made to the modern population of the middle northern Fertile Crescent (Modern Syria).  Finally, we've applied the First Order diffusion model, and have worked backward to a First Arrival Date in the Fertile Crescent for people from Southwest Asia. 
The First Order Model diffusion rates yield a Southwest Asian (SWA) First Arrival Date of approximately 4000BC.  The First Order Model implies that diffusion of SWA populations has been underway among Fertile Crescent inhabitants for approximately 2500 years before the Middle Assyrian Period.
It is important to keep in mind that these diffusion rates were estimated from a mid point diffusion in time and distance from coalescence.  In the graph, a best fit of the slope of the First Order Model, Normal Distribution Model, and the Unbounded Exponential Diffusion Model are obtained at the diffusion mid point.  From this fitting, we can estimate that the SWA First Arrival Date for the Normal Distribution is
1500BC + (2500BC)1.4 = 5000BC
where the factor of 1.4 has been estimated from the point where the SWA population density is less than 2%.
For the Unbounded Diffusion Model, the SWA First arrival date is estimated to be:

1500BC + (2500BC)2 = 6500BC
Summarizing Fertile Crescent SWA First Arrival Dates:
First Order Model:  4000BC
Normal Distribution Model:  5000BC
Unbounded Diffusion Model:  6500BC

The Coalescent in a Continuous, Finite, Linear Population

An interesting paper describes the process of gene flow that we see in the Fertile Crescent between the West Asian component and the Southwest Asian component:

The Coalescent in a Continuous, Finite, Linear Population, Jon F Wilkins and John Wakeley, Genetics Society of America, March 4, 2002.


From the Abstract:

"In this article we present a model for analyzing patterns of genetic diversity in a continuous, finite, linear habitat with restricted gene flow. The distribution of coalescent times and locations is derived for a pair of sequences sampled from arbitrary locations along the habitat. The results for mean time to coalescence are compared to simulated data. As expected, mean time to common ancestry increases with the distance separating the two sequences. Additionally, this mean time is greater near the center of the habitat than near the ends. In the distant past, lineages that have not undergone coalescence are more likely to have been at opposite ends of the population range, whereas coalescent events in the distant past are biased toward the center. All of these effects are more pronounced when gene flow is more limited. The pattern of pairwise nucleotide differences predicted by the model is compared to data collected from sardine populations. The sardine data are used to illustrate how demographic parameters can be estimated using the model."

The model postulates the normal distribution for the expansion of populations and also suggests that migration is conservative.

Thus, the location of a parent is normally distributed around the location of its offspring, with variance 2 sigma squared.

The boundaries of the habitat are reflecting, so a gamete that would otherwise land outside the habitat range is reflected back an equal distance within it. Each individual thus has the same expected number of offspring regardless of its location. This means that migration is conservative, so migration alone is sufficient to maintain the relative population densities at all locations in the habitat (Nagylaki 1980). Nonreflecting boundaries would correspond to the case where those gametes dispersing outside the habitat range are lost. In such a system, individuals near the edges of the habitat would have a reduced effective fecundity relative to those nearest the center.

As Felsenstein (1975) pointed out, most continuous space models in population genetics assume a uniform population density that would not actually be maintained by the proposed reproductive scheme. A normal distribution of gametes without severe density regulation generates a population that is clumped together at certain locations and sparsely populated at others. With its absolute density regulation at all locations, the model of reproduction proposed here will immediately generate and maintain a population that is uniformly distributed across its habitat range.

What's striking about this paper is that it exactly describes the diffusion process that has been occuring in the Fertile Crescent.

In the previous post, I've looked at two other diffusion models, in addition to the first order approximation. Here, we add the Normal Distribution Model described by Wilson and Wakeley:
We can see that the First Order Approximation is in good agreement with the Normal Distribution Model for populations that are uniformly distributed.

Demic Diffusion Model: Bounded and Unbounded

I've discussed in a previous post that the solution to the diffusion equation produces a decaying exponential.  So far, in calculating demic diffusion rates for Syria, I've used a first order approximation to this equation.

How good is the first order approximation?    What does the data tell us about the shape of demic expansion?

In looking at the Dodecad K10 data, there are signs that more than one "shape" of demic expansion is at work.  In some cases, you see hard boundaries between peoples living several hundred miles from each other.  The separation between the Armenians and Georgians comes to mind, with each confined to their respective nearby valleys.  In other cases, you see an entirely different diffusion shape, with a very soft and gradual diffusion of peoples, such as that seen between Tuscans and Northern Italians or between Syrians, Jordanians and Egyptians.

The three models of diffusion can be graphed:
The First Order Model(FOM) is a straight line approximation to the decaying exponential.  The gradual diffusion model can be described by an unbounded exponential decay.  I'll call that the Unbounded Diffusion Model(UDM).  The third model is bounded, with a leading edge that lies somewhere between the Unbounded Diffusion Model and the First Order Model.  That's the Bounded Diffusion Model(BDM). 

For purposes of estimating dates for when the leading edge of a population component reached a particular area, I'll give the FOM date and the UDM date.  The UDM edge will be set at the 2% proportion level so that the UDM date will be twice that of the FOM date, as pictured in the above graph.  That's arbitrary, but makes for a nice rule of thumb. 

Wednesday, November 17, 2010

Evidence of the Ancient Spice Trade in Syria

Another revealing day working with the Syrian and Assyrian Dodecad Data.

Adhering to the idea that demic diffusion should be inferred from nearby real populations and not synthetically created ones, ie pure sub-components, I've interpolated backward from the Middle Assyrian Period (1500BC) to the Neolithic, using the diffusion rates for countries adjacent to Syria:

On the left of the above graph is the bar for Assyrians plus the "remainder" term inferred in the previous post.   There are two "seven bar" diffusion sets, one for the diffusion from the south and east, represented by Saudis, and a second term for diffusion from the north and west, represented by Cypriots.  The diffusion from these two processes occur contemporaneously, but can be represented separately because of the linearity of the diffusion equation.   Each bar describes 350 years of diffusion.  The 7x350=2450 years of diffusion take us from the Middle Assyrian Period in 1500BC to 3950BC. 

As the Egyptian Invasion of Syria occured in the 14th Century BC, I didn't include Egyptians in the diffusion process for this period.  I also didn't include a separate diffusion process for Babylonia.  Prior to 1500BC, Assyrians focused their trade westward.  Prior to 1500BC, there are "South Asian/Northwest Asian/East Asian" components in the Modern Saudis (that diffused from Babylonia) that model the more limited diffusion from Babylonia in the Bronze Age.

At a very low level during the Bronze and Iron Ages, another demic process was at work which contributed to the West African, East African and Northwest African components that are not accounted for by any other diffusion process. 

I propose that it is the Spice Trade, and to a limited extent, the trade in slaves that brought people from far flung corners of the Ancient World to the Fertile Crescent.  The spice and slave routes of the ancient world are described on the following maps.

Picking from populations along these routes, it is possible to account for some of the minor components in Syrians.  The results are plotted on the following graph:

On the x axis:

"1" represents the 4000BC Fertile Crescent population after diffusion from "Cyprus" and "Saudi Arabia", but not accounting for diffusion from other processes.

"2" represents the contribution to Syrians during the Bronze and Iron Ages from diffusion assumed to be due to the spice trade and slave trade.  Components are:
0.17% Northwest African
0.53% West African
0.49% East African

"3" represents the 4000BC Fertile Crescent population of "1" less the minor diffusion components ("2").

Some notes on the contruction of the slave and spice contribution: 
I decided to try to reconstruct the minor components in "2" from populations known to have traded with Syria, rather than raw components.  I was able to contruct the "trade" contributions from various combinations of the Dodecad Ethiopian, West African, the Maasai and Mozabite populations.  As all of these populations would have had access to spice and slave routes, and in fact were central to these forms of trade, it's plausible that Ethiopian, West African, Maasai or similar East African populations, and Berbers are the sources of these minor components in Syria today.

After accounting for all of these sources of diffusion, we arrive in the Halafian Fertile Crescent in approximately 4000BC ("3" in the above graph).  Normalizing this result of "3", we have:

The Dodecad K=10 components for this postulated Neolithic population are:

68.1% West Asian, 27.8% South European, 1.12% Northeast Asian, 0.65% North European, 2.33% South Asian and 0% on the other components.

Tuesday, November 16, 2010

Syria to Assyria: 3500 years of Demic Diffusion

Working from the simple demic diffusion model for Syria, I've adjusted the diffusion rates, based on the assumption that the Modern Assyrian population  is a snapshot of the central Fertile Crescent in the Middle Assyrian Period 3500 years ago.  Again, that date is chosen because it is before the Egyptian occupation of Syria and after the conquest Assyria by the Amorites (1840BC).  The assumptions behind that dating are further discussed in the Assyrian post.

The first bar in the above graph represents the modern population of Syria. The next four "five bar sets" account for four sources of demic diffusion into the Syrian population:  Egypt, Babylonia, Saudi Arabia and Cyprus.   Working backward through time, each bar represents 700 years of history.  The bar second from the right is the "snapshot" Assyrian population which we've assumed has been isolated for the last 3500 years.  The far right bar is the remaining population not accounted for by diffusion from Egypt, Babylonia, Saudi Arabia and Cyprus.  This combined remaining population accounts for less than 4% of Syrian population components.

From this, I calculate a diffusion rate for Syrians:

Egypt:             10%/3500 years = 0.002857% per year
Iraq(Burusho):   4%/3500 years = 0.001143% per year
Saudi Arabia:  18%/3500 years = 0.005143% per year
Cyprus:           13%/3500 years = 0.003714% per year


0.5% East African
0.5% West African
0.25% Northern European
0.25% Northeast Asian
1.5% Southern European
0.25% Northwest African

These diffusion rates can be used to interpolate further backward in time through the Bronze and Neolithic ages.  It is important to note that these diffusion rates represent the average rate of diffusion.  They are first order approximations of the more accurate decaying exponential predicted by the diffusion equation.  These rates are directionless and represent a "point" diffusion rate for the geographic region of Syria. 
Using these rates, it is possible to estimate a last possible date for the postulated departure from the Fertile Crescent to Europe.  That will be the topic of tomorrow's post.

Monday, November 15, 2010


"Ashur lived at the city of Nineve; and named his subjects Assyrians, who became the most fortunate nation, beyond others” (Antiquities, i, vi, 4): Flavius Josephus

Once again, with the recent Dodecad Ancestry Project results for Assyrians, we see a breathtaking concurrence between history and autosomal genetic data.

Having existed in various entities for so long, it is difficult to easily define Assyrians.  In my definition, I will reference, apologetically, the wiki page for Assyrians, as the hard work of defining the various Assyrian historical periods seems to have been done there.  Wikipedia also has a very good map of the Fertile Crescent during the Middle Assyrian period:

Prior to and during the Early Period (circa 1920 BC – 1840 BC), Assyrians were a people of the upper Tigris River who traded with people of the Anatolian plateau in what would be Wilusa and Assuwa on the above map.  Interestingly, the political structure of ancient Assyrian cities was one where the basis of authority rested with the city through an assembly of elders, a hereditary ruler and an eponym who was annually elected by lot to administer commercial interests. 

In about 1840BC, the city of Ashur was conquered by Shamshi-Adad I, King of the Amorite tribe to the southwest.  During the succeeding centuries of the Middle Assyrian period, the Assyrians allied themselves variously with the rising power of Babylon to the southeast and the Hanigalbats and Hittites to the north.  The above map represents that period.

Significantly, Assyrian trade and relationships with the Anatolian Plateau ceased during the Middle Assyrian period, as Babylonia became the focal point of Assyrian trade relations, rather than Anatolia.

With this history in mind, it is interesting to look at the genetic makeup of Modern Assyrians (2) against a background of Modern Syrians (1):

At first glance, it seems like these are a different people.  Thinking about demic diffusion, you can begin to sort out how Syrians and Assyrians, neighboring inhabitants of the Fertile Crescent, are the same and different.

I revert to the discussion of the simple demic diffusion model proposed for Syrians.  This model suggests that Syrians are comprised of an ancient Fertile Crescent mix of West Asian, Southern European and Southwest Asian components, as described in the Dodecad Ancestry Project.  Syrians also show evidence of the Egyptian occupation of Syria in the 14th century BC.   Likely by way of ancient Babylonia, they show a South Asian component.  The contributions from Egypt and South Asia are at 12% and 6%, respectively.

Assyrians, on the other hand, show very little influence as a result of the Egyptian occupation.  This may be due both to the geographic position of Assyrians on the Upper Tigris River, as opposed to further west, and also due to the religious isolation of Assyrians. Assyrians do show a 2% contribution from South Asians, perhaps indicating a very ancient but low level South Asian presence in Babylonia.

We now plot Assyrians in the "Fertile Crescent" picture, with the results normalized on the West Asian-South European-Southwest Asian components, and with results for Northern Europeans and South Asian components included:

In this picture, both Assyrians (3) and Syrians (5) show the familiar West Asian-Southern European- Southwest Asian mix.  Both present the Southern European component at about 22%.  Both show a small component from South Asia.  In fact, all populations except Egyptians show a small genetic contribution from South Asia.  However, what is most striking about these results is the shift toward the West Asian component for Assyrians compared to Syrians.  It is as if Assyrians, genetically speaking, are frozen in time, sometime in the Middle Assyrian Period:  they have a smaller Babylonian South Asian component (2% vs 6%),  lack the Egyptian component of the 14th century BC Egyptian occupation and have significantly less of the Southwest Asian component (23% vs. 36%) than Syrians. 

In combination with the high Southwest Asian Component results for Saudis, Egyptians and Ethiopians(Dodecad Ancestry Project), this results indicates that the Southwest Asian component has been diffusing northward since the Neolithic.  Moreover, the conquest in 1840BC of the Assyrians by Shamshi-Adad I, King of the Amorites, may have been a significant genetic event, bringing the Southwest Asian genetic component northward.

By comparison, the high West Asian components in Lezgins, Georgians, Modern Turks and Armenians indicates that the origin of the West Asian component is in the Southern Caucasus and has been diffusing southwestward since the Neolithic.

If we assume that Assyrians are a genetic "snapshot" of the Middle Fertile Crescent during the Middle Assyrian Period, we can then begin to guess at a rate for the northward diffusing Southwest Asian component and the southward diffusing West Asian component in this region.  I'll be developing this idea further in subsequent posts.

If this diffusion model is correct, prior to the Bronze age, Assyrians and their predecessors were a "West Asian"/ "South European" people.  That would make early northwestward migrating Assyrians an obvious candidate for a demic transmission of farming to the Western Black Sea and Balkans.

I'd like to thank Dienekes Pontikos and the Dodecad Ancestry Project for collecting and generating the ADMIXTURE results for Assyrians and also thank "handschar" on dna-forums for recommending that I look at the Dodecad Assyrian data.

Update January 24th, 2011: An update on the use of ADMIXTURE for analysing Middle Eastern Populations

Friday, November 12, 2010

A Simple Demic Diffusion Model for Syria

The Middle East

We explore a simple near neighbor linear diffusion model for Syria to estimate the source populations for this country.  From the solution of the diffusion equation, we know after their initial establishment, populations will spread out spatially in x as a function of a decaying exponential exp(-x).   Using a Taylor series expansion, we can approximate this function to first order as

Using this idea, we can begin to guess at the diffusion processes that have led to the subcomponent combinations that form the modern population of Syria.  Furthermore, because the diffusion equation is linear, we can use linear superposition to sum the contributions of the various "x" direction contributing populations to Syria.

We use the modern populations of adjacent countries for the estimates, even though it is understood that demic diffusion has been at work on these populations for millenia.   The subcomponent groupings, West Asian, Northwest African, etc. for these three countries are derived from an ADMIXTURE analysis of 40 living Eurasian populations.

To justify this simple approach, we also assume that the diffusion constant D, described in equation (41) of the linked paper, is small.  That is equivalent to saying that although the populations have been spreading out in space, the composition of the source populations hasn't changed very much during the time in question.

We subtract by a fixed increments of "population" from "Syria", as shown in the graph below. 

Because it is known that Egypt occupied Syria from the 14th century BC, as a first choice, we subtract "Egyptians" from Syrians. 

Next, since the model doesn't present a population for Syria's eastern neighbor Iraq, we substract a proxy, Pakistani Burushos.  That's a big assumption, but we know that Syria traded with Mesopotamia.  We can see that there is a quite visible South Asian component in Syria and it's most likely source is via Mesopotamia.

Using this "1-x" approximation, where "x" are the modern populations of Egypt and a population from Pakistan, Burushos, we've now approximated a source of the East African and South Asian subcomponents in Syrians.

Next, we subtract Cypriots, an obvious and nearest neighbor trading partner throughout history.  The likely source for the Southern European in Syrians is through millenia old trade throughout the Mediterranean.

Finally, we subtract equally from Saudis and Georgians, near neighbors who are likely to be the sources for the Southwest Asian and West Asian components.  It is true that the path of these populations into Syria is by way of Turkey and Jordan, but for the sake of simplicity, we choose Saudi Arabia and Georgia.

The results, in the 7 "x directions", where x = x1, x2, x3, x4, x5, x6, and x7 yields:

Cyprus(x1) = 36%
Saudis or Arabian Peninsula(x2) = 24%
Georgians(x3) = 22%
Egyptians(x4) = 12%
Pakistani Burushos, via Iraq or Mesopotamia(x5) = 6%
East Africa(x6) = less than 0.5%
West Africa(x7) = less than 0.5%

All this "1-x" subtraction suggests that Cypriots, followed by people from the Arabian peninsula, followed by people from the Western Caucasus, are the dominant players in the composition of modern Syrians.

The model suggests that there is a contribution to Syria from Egyptians at about 12% and from India or Pakistan, by way of Mesopotamia, at about 6%.

The remaining Syrian components are a West African component at less than 0.5% and an East Africa component at less than 0.5%.

Using linear superposition, we can sum the contributions from these seven populations.  That's the modern population of Syrians.

Admittedly, the model is a little simple.  We can't separate time and space and therefore cannot clearly tell how long the diffusion processes have been going on or how large were the original contributions from each of the seven sources:  Egypt, India or Pakistan, Cyprus, the Arabian Peninsula, Georgia, West Africa and East Africa.

From what we know about ancient Mediterranean trading routes, we can guess that the contributing population from Cyprus was the largest contribution.  The contribution from Egypt may be due to the occupation of Syria 3500 years ago, but may also be due to a continuous diffusion process from Egypt, through Jordan and into Syria.   The contribution from India or Pakistan is interesting.  It's probably old, a over time scale similar to that from Cyprus, but from a smaller and more distant source population.

Finally, the contribution from Georgia and the Arabian peninsula may be very old.  Both of these are candidate populations for Syria prior to the develop of farming and trade in the Fertile Crescent and Mesopotamia.

Wednesday, November 10, 2010

Embracing the Hellenic Plural Identity

Doric Temple in Segesta, Sicily

From the early Bronze Age, the Minoans inhabited Crete and the Aegean Islands.  They were a wealthy trading power in the Mediterranean with commercial connections to Egypt and Mesopotamia.

The Myceneans occupied Southern Mainland Greece.  Other Greek speaking tribes such as the Aeolians and Dorians occupied positions to the north.

Starting in about 1150BC, the Dorians descended from the north and into southern Greece. 

Of course, there were other Greek peoples.  For simplicity, I confine the discussion to the Aegean peninsula.

Dienekes' genetic data for Greeks here and here does suggest that there is an echo of those ancient people who came together to form the Hellenes.

Looking at the recent Balkan groups that Dienekes' has presented, we have Bulgarians that are composed almost entirely of an Italo-Balkan component.  Serbians as well.  On the other end, we have Cypriots that bear a classic Mediterranean Fertile Crescent signature.  In Dienekes' Balkan-Italian-Caucasus plot, Cypriots get their own component. 

Here's where it gets interesting.  Greeks and Armenians seem to be sandwiched in the middle, with a little Georgian or West Asian thrown in for good measure.  That suggests that we're looking at the origin of Greeks:  Minoans (Fertile Crescent signature) intermeshed with Greeks who have been in place since the LGM and varying degrees of a Balkan genetic contribution.

It's too early to get a complete genetic picture.  Dienekes' Greek data is taken from only five samples; the Serbian and Bulgarian data from only a few people.

For the record, I plotted the Greek data from the Eurasian run and then tried to reconstruct it from Cypriots, as a proxy for Minoans.  Here's what I got:

        1=Greeks    2=Cypriots  3=Italo-Balkans

        Greeks = 0.48% Cypriot + 52% Italo-Balkan

The Italo-Balkan component has the familiar Proto Italo-Celt substructure:  North European-South European-West Asian in the amounts 24%-40%-36%.   It's a little different from the background populations I calculated for Italians.  The Greek background population (referred to here as Italo-Balkan) has more West Asian and less Southern European than Tuscans and Northern Italian background populations.

Given the isolation of many Greek regions, even into modern times, it will be interesting to see if there might still exist a North to South Italo-Balkan-Minoan cline in the Aegean peninsula. 

Time will tell.

Akrotiri Ship Procession (Minoan)

Tuesday, November 9, 2010

Fertile Crescent in Southwestern Europe: Revisiting Assumptions

A number of assumptions have been made while trying to work out the Fertile Crescent "who and how much" in the Southwestern European populations Tuscans, Northern Italians and Spaniards.  I'll state them explicitly here.

First, we have only chosen three populations thus far.  It is possible that other Fertile Crescent populations are the "who" in Southwestern Europe.  Notably, we've thus far left out Jews, Greeks, Cretans, Jordanians and Egyptians from possible sources for a westward migration. 

The reason I haven't looked at Jews is because I don't have an origin population for them.  Dienekes' looks at Ashkenazi and Sephardic Jews, but Iraqi Jews and many others were dropped from Dienekes' later K=10 run that I have been using.  It's clear that in their long diaspora in Europe, the European Jewry have absorbed some European genes.  In the Ashkenazim and Sephardim, you can see that in their Northern European components as well as in their elevated Southern European components, compared to Iran and Iraq Jews (see Dienekes' earlier k=10 run).  Without an origin population for Jews, it's difficult to infer their contribution to Southwestern European populations.  It's also difficult to nail down a position in the Fertile Crescent for Jews.  Iran and Iraq Jews look very Northern Fertile Crescent and Ashkenazy and Sephardic Jews look very Mediterranean Fertile Crescent. 

I will say that many Jewish groups seem to bear the unmistakable marker of a small Mozabite component, but again, looking at Dienekes' earlier k=10 run, it is notably absent in Iraq Jews.  So we can't say for certain that the combination of a Fertile Crescent + Mozabite components is a universal marker for Jews.

For now, the best we can do is to say that the Jewish origin population fits somewhere in the Fertile Crescent triangle, but we don't know where.

I'll note that a close look at Dienekes' new K=10 run shows that there is a small Mozabite component across Tuscans, Northern Italians, Cypriots and certainly Spaniards.  That hints at a small contribution to Tuscans, Northern Italians, Cypriots and Spaniards from an Ashkenazi or Sephardic Jewish origin.

I haven't included an obvious point in the path of a westward migration:  that of Crete and other points of Greece.  Crete and Greece aren't in the Fertile Crescent, but based on Dienekes' latest runs with Greeks in the mix, Greeks are looking like (1) a Fertile Crescent/Northern European mashup or (2) a Fertile Crescent/Northern European/Southern European mashup. Greeks look a lot like Tuscans across four components.  I'll be adding Greeks to the analysis in the next few days.

I also haven't included Jordan or Egypt.

The Fertile Crescent Signature of Jordan looks very similar to that of Syria which is why I didn't include it.

It would be improbable that the point of origin of Fertile Crescent populations was from Egypt. The Egyptian signature is tilted heavily in favor of Southwest Asia and away from West Asia. As a contributor to Tuscans, Spaniards and Northern Italians, it yields a weird result where almost all of the West Asian in Southwestern Europe would have arrived by another non Mediterranean route.  Egypt is also an unlikely source because of the West African and Eastern African contribution in the Egyptians.  It's likely that due to ancient movements up and down the Nile, these West and East African components have been present in Egyptians at least since the Neolithic.  Therefore, if Egypt made a significant population contribution to Southwestern Europe since the Neolithic, these West And East African components would appear in Europe.  They do not.

A significant result of these linear additions is that Modern Turkish and Armenian Fertile Crescent populations don't work as a source for Southwestern Europe.  They've got too much West Asian compared to their Southwest Asian component to yield a solution for Tuscans.  It's likely that Turkey has not only absorbed Turkic populations in the last several thousand years, but has also continued to absorb more of a West Asian component.  Ancient Anatolia and certainly ancient Southern Anatolia likely had less of the West Asian component.  We don't know how much less.  What I've done here is set the limit on the greatest amount of the West Asian component, set against the Turkish Southwest Asian component, that can still yield an admixture solution for Tuscans.  I've called that mix "Anatolian".

We can improve upon inferred results from this model by including 1-sigma standard deviation results for each component in each population.  I haven't done that yet, but hope to in the future.

Finally, in terms of inferring background populations in Southwestern Europe, I've made two critical assumptions:

1.  The West Asian-Southern Europe-Southwest Asian component ratio in Fertile Crescent autosomal DNA is stable over thousands of years.  This key assumption allows isolating the contribution from the Fertile Crescent in Southwestern European populations on the Southwest Asia component. 

2.  I assume that the Southwest Asia component was not present in Southern European background populations. 

The method of inference for the background populations is most sensitive to these two assumptions.

That's an exhaustive discussion of assumptions.  I'll add to this and develop upon this in future posts.

Monday, November 8, 2010

Inferred Background Populations for Tuscany, Northern Italy and Spain

Working from the possibility of three different and possibly combined migrations from the Fertile Crescent (Cypriot, Syrian and Southern Anatolian) to Tuscany, Northern Italy and Spain, it is possible to infer background populations for these groups.

The proportions for the Fertile Crescent populations and their inferred Southern European background proportions for Tuscans, Northern Italians and Spaniards are shown in the following table:
Table of Southern European Admixture Components

It is clear that Tuscans have the highest contribution from the Fertile Crescent (24 to 42%).  The contribution from the Fertile Crescent to Northern Italians and Spaniards appears to be less (between 10 and 20%).

It is interesting to plot the inferred background populations in terms of their Western Asian vs. Northern European components and in terms of their Southern European vs. Northern European components.

In the graphs, the Tuscan background population appears in blue, the Northern Italian in yellow and Spanish in red.

The three points of each triangle are inferred by Syrian, Cypriot and Anatolian migrations, in the proportions defined in the table above.  These points are labeled respectively S, C and A.

In the Southern European vs. Northern European plot, Basques, plotted as a star, are added as a point of reference.

What can these graphs tell us?

First, the Northern Italian inferred background population appears to be very similar to the Tuscan background population.  The difference is in the proportion.  We can also see that while it is possible that the sole West Asian contributor to both Tuscans and Italians came only from a direct Fertile Crescent migration, it seems likely that a small portion of the West Asian component in both Northern Italians and Tuscans came by a Danubian route.  The Myres et. al. and Curdy papers would indicate that this occured with Bronze and Iron Age migrations.  If that's true, then it implies that Danubian Farmers absorbed some West Asian populations either in the Bronze or Iron Age.   That seems likely, as even Lithuanians have a small proportion (7.4%) of the West Asian Component.  We can't know for certain when this West Asian component was absorbed into the Danube and Dnieper-Don migratory paths, but whatever the timing, some of this West Asian component acts to pull the "possibility space" triangles for the background population of Northern Italians and Tuscans off the zero West Asian axis.  By comparison, the Spaniard background population hugs the zero West Asian axis, indicating that they are less influenced by the Danubian Bronze and Iron Age Migrations.

The plot of the Southern European vs Northern European possibility space also hints at the difference in the population of Spain in comparison with Italy.  The Spaniard background population matches the Northern European position of Basques, but has less of the Southern European component.

The Italian background populations, by way of reference, trend toward having both less Southern European and less Northern European than Basques (and correspondingly more West Asian).  At the tip of the Tuscan possibility space, assuming an Anatolian Fertile Crescent contribution, the background population for Tuscans matches closely with that of French Basques.

It should be noted that the "possibility space" as plotted does not include uncertainly.  The populations described in Dienekes' ADMIXTURE results have a variance, and were that variance known, it would be possible to add uncertainty into the graphical description of these background population possibility spaces.

Sunday, November 7, 2010

Tuscan Background Population: Who Are They?

Working with three guesses for the Fertile Crescent contributor to Tuscans, we can guess at the proportion of a Fertile Crescent contribution to Modern Tuscans.  That proportion varies from about 24% to 42%. 

One way or another, these travellers from the east integrated with non Fertile Crescent Tuscans.  One wonders who was in Tuscany before the arrival of the easterners.  It's clear that Southern Europeans were.  Their signature is everywhere in the Mediterranean.  The composition of nearby Modern Sardinians, who are composed of 96% South Europeans, leaves little doubt that South Europeans also were the dominant population in Tuscany before the Eastern arrival. 

There is also a North European component in Tuscans.  Depending on where in the Fertile Crescent the Easterners arrived from, there could be a West Asian contribution to Tuscans that did not arrive directly from the Fertile Crescent. 

That's one reason I've been calling the background population "Mystery Composite".  We don't know the proportions of "Sardinian" versus other non Fertile Crescent populations.  We'll call these Non Fertile Crescent populations the Tuscan Background Population.

For now, we can plot the Northern European verses Southwest Asian "possibility" space for the Mystery Composite Tuscan Background Population:

Next, using the same Fertile Crescent Cypriot-Syrian-Anatolian Triangle used for Tuscans, we'll look at the background Northern European vs Western Asian possibility space for Northern Italians. 

Saturday, November 6, 2010

Summary: Three solutions for Tuscans

% Three solutions for Tuscans in 10 vector format
% Ten vector format is [WA, SWA, SWA, NWA, SWA, EA, NE, WA, EA, SA]

%            0.338*cypriots(WA,SE,SWA normalized)+ 0.66*mystery1
tuscans1=0.338*[0.404 0 0.364 0 0.232 0 0 0 0 0] + 0.66*[0.15 0 0.475 0 0 0 0.375 0 0 0]

%             0.24*syrians(NW, SE, SWA normalized)  + 0.76*mystery2
tuscans2=0.24*[0.419 0 0.220 0 0.361 0 0 0 0 0]   + 0.76*[0.18 0 0.50   0 0 0 0.32  0 0 0]

%            0.42*anatolians(NW, SE, SWA normalized)+ 0.58*mystery3
tuscans3=0.42*[0.56  0 0.26   0 0.18   0 0 0 0 0]    + 0.58*[0     0 0.57   0 0 0 0.43  0 0 0]

Tuscans, Guess Three: Setting an upper bound on the Fertile Crescent West Asian Contribution

Let's review the component signature for the Fertile Crescent populations:

So far, we've looked at Cypriot and Syrian like populations as possible sources for the Tuscan Fertile Crescent component.

Repeating the same process with Modern Turk or Armenian samples to infer a "Mystery Composite" leads to an interesting result.  It turns out that there is no solution with these populations.  There is too much "West Asian".  Even if "Mystery Composite" has no "West Asian" at all, combining these Turkish or Armenian populations with a South Europe-North Europe Mystery Composite will yield a Tuscan population that is too rich in the "West Asian" component.

What this tells us is that there is an upper bound on the West Asian component in the Tuscan Fertile Crescent contributor.  Keeping the South European component the same as for Turks and trading off West Asian with Southwest Asian, the upper bound for the normalized Fertile Crescent signature is approximately 56% West Asian, 26% South European, and 18% Southwest Asian.  Plotting this West Asian upper bound signature (see stacked bar 3 in the graph) with the other Fertile Crescent signatures gives us some idea as to its geographic origin:

Using this 56%-26%-18% guess for the Fertile Cresent contributor to Tuscans, we get a 42% contribution from the Fertile Crescent and a 58% contribution from "Mystery Composite".  In this case, "Mystery Composite" is composed of 57% South European and 43% North European.  In this bounding case, there is no West Asian component in "Mystery Composite."

This solution would indicate a Southern Anatolian origin and would concur with genetic evidence.

Friday, November 5, 2010

Tuscans, Guess Two

An Ancient Phoenician Coin

In the previous post, we guessed at some Fertile Crescent contribution to modern Tuscans, based on the examination of Dienekes' k=10 ADMIXTURE data, which indicates that a Fertile Crescent signature is buried in the mix of population components that make up Modern Tuscans.

As a first guess, we took a stab at Cypriots as the source for the Fertile Crescent contribution to Tuscans.

Because the Fertile Crescent Signature is buried in other similar populations, it's not clear exactly which Fertile Crescent population contributed to the Tuscan gene pool.  So we'll take another guess at who the intrepid Mediterranean seafarers might have been.

This time, we'll guess Syrians.  That's a good guess, as Syrians are an intermediate population in the Fertile Crescent.  Their Fertile Crescent genetic signature is similar to neighboring Jordanians.  In addition, their proximity to Lebanon would hint that the Syrian Fertile Crescent composition might be a close match to those mythical seafarers, the Phoenicians.

We follow the same process as in the previous post.  First, we assume that all of the Southwest Asian component comes from the "Syrian" contribution.  Syrians have a 32.4% Southwest Asian component, while Tuscans have a 7.7% Southwest Asian Tuscan Component, so we estimate that the Syrian contribution to the Tuscan gene pool is 7.7/32.4 = 23.76%.

Again, we set up a series of guesses for the "Mystery Composite", the contribution to the Tuscan gene pool that is not from the Fertile Crescent.  These are shown in the table, below:

Table of Population Subcomponents 1 to 10,
Guesses for “Mystery Composite”
Adjusted for “Syrian” Fertile Crescent Contribution



We plot Tuscans against 0.2376 Syrians + 0.7623 Mystery Composites(six guesses from table above):
Again, Tuscans are represented on the left of each bar set.  The result of the 0.2376 Syrians combined with  guesses 1 through six for the Mystery Composite are shown on the right bar in the six bar sets.  It should be noted that I have slightly renormalized the Syrian result, based on the observation that there is no observable South Asian, East African or West African contribution to Tuscans.  You can see these minor components on the top of the right bars.  If Syrians or perhaps Phoenicians did contribute to the Tuscan gene pool, the group that did lacked these components.  Alternatively, it is possible that the samples chosen for Tuscans somehow miss these minor components.

Examining the table, it is guess two that most closely approximates Modern Tuscans.  With Syrians as the Fertile Crescent contributor, Mystery's best guess is described by a ten component vector, as in Dienekes' table for the k=10 run:

[0.18    0     0.5   0       0     0     0.32    0     0     0]

18% West Asian, 50% Southern European and 32% Northern European.

It's notable that this result isn't very different than the last result.  With a part Syrian origin for Tuscans, rather than a part Cypriot origin,  "Mystery" has only slightly more West Asian and slightly less Northern European, but the results are similar.

Next, we'll look at a part Anatolian origin for Tuscans.