August 17, 2009

Coastal-inland differences in Y chromosomes of the Levant

More on this after I get a hold of and digest the information in the paper.

Just a quick comment, based only on the abstract, that the Levantine populations should be studied in a European context as well, as they have been influenced by prehistoric populations from the Aegean, Greeks, Romans, medieval Crusaders, or Ottomans of various origins.

UPDATE: The paper has several supplementary figures and tables.

In Figure S1 we see the biallelic markers used in this study, and their representation in the various populations. It is a chronic problem with studies of this sort to undertype samples; there are phylogeographically informative markers within haplogroups G, L, E1b1b, and J2 for example, which would have added important information about the specific affinities of these haplogroups in the studied populations.


Inspit of these deficiencies, we may still make some useful observations. For example, IE-speaking Iranians have largely the same haplogroups as Arabs, but a much higher representation of haplogroup J2 compared to J1. The converse is true for all Arabs except the Lebanese. But, we do know, that even in Lebanon itself, Muslims have a higher J1/J2 ratio than Christians, and Islam was the main vehicle of Arabization in the region. The Christians are descended from the pre-Arab Byzantine Greco-Aramaic populations (with an addition of Western European Y-chromosomes in some Christian communities, which would not have substantially upset the J1/J2 balance).

It is fairly clear to me that in the Middle East, Greek and Iranian-settled regions have a higher J2/J1 ratio than regions with solid Semitic or NE Caucasian populations where J1 predominates.

UPDATE II (Aug 27):

The paper reports a near zero frequency of haplogroup J1 in Tunisia and Morocco, after an earlier study by the same authors. However, a different study (Onofri et al.) on Moroccan and Tunisian Y chromosomes report 20 and 35% respectively, which is in agreement with an earlier study on North African Y-chromosomes (Arredi et al.) The discrepancy in the J1 frequency seems too large to have arisen by chance given the sample sizes, and it would be interesting to see how it may have arisen.

Annals of Human Genetics doi:10.1111/j.1469-1809.2009.00538.x

Geographical Structure of the Y-chromosomal Genetic Landscape of the Levant: A coastal-inland contrast

Mirvat El-Sibai et al.

Abstract

We have examined the male-specific phylogeography of the Levant and its surroundings by analyzing Y-chromosomal haplogroup distributions using 5874 samples (885 new) from 23 countries. The diversity within some of these haplogroups was also examined. The Levantine populations showed clustering in SNP and STR analyses when considered against a broad Middle-East and North African background. However, we also found a coastal-inland, east-west pattern of diversity and frequency distribution in several haplogroups within the small region of the Levant. Since estimates of effective population size are similar in the two regions, this strong pattern is likely to have arisen mainly from differential migrations, with different lineages introduced from the east and west.

Link

54 comments:

Gioiello said...

From another thread:

Similar haplotypes there are on the paper posted by Dienekes above (Coastal-inland differences in Y chromosomes of the Levant): Jordan 10AM71, 1AM60, 8AM71 and Iran 6AQ106.
The paper seems to demonstrate that hg. R expanded from Europe to North Africa and Middle East, being present on the coasts and diminishing when we go inside.

terryt said...

"we also found a coastal-inland, east-west pattern of diversity and frequency distribution in several haplogroups within the small region of the Levant".

That is only to be expected of course. Especially relevant, as Gioiello points out, to the Y-chromosomes of Jewish priests post. The people in the region are all from a diverse background and presumably today closely related. But the bitterest arguments are always those between family members.

Aaron said...

It says this is new data but it looks strangely similar to the previous data, ratios and everything. Would someone mind pointing out the new stuff for this blind fellow? Thanks

Paul_Johnsen said...

A total of 885 new samples from five populations (Syria, Jordan, Iran, Egypt and Kuwait) were collected and analyzed for
this study. In addition, samples and genotyping data from 951
Lebanese, 200 Syrian and 101 Palestinian men (Akka) were already available (Zalloua et al., 2008a, 2008b).

Gioiello said...

Letting apart the Hgs. J1 and J2, about which Dienekes has said the essential thing (J1 is the haplogroup of the pastoralist inlanders and J2 of the agriculturalists from Asia Minor or South Europe, and we knew already this from King and others), if the problem for instance between me and Vizachero was the origin of R1b, above all after R-L150+, this paper seems to say that this haplogroup was born in Europe and then expanded to North Africa, Middle East till Mesopotamia and Iran. That’s all.

Maju said...

My impression is that J2 is more frequent in highland West Asia and J1 in lowland West Asia. The exception of Daghestan should not change this essential dichotomy too much (specially since Daghestan and the other high J1 areas of the Caucasus are in fact in Europe).

I can agree with J1 having a Semitic (in West Asia) and more broadly Afroasiatic (in North Africa) affinity but I cannot agree with J2 being related to any Indoeuropean process (no particular correlation with R1a). It would seem more related with the Neolithic area of Zagros-Taurus if anything (J1 in turn would be more related to that of the Levant).

Also it's noticeable to me that J1 and E1b1b are pretty much teamed up: even if it's obvious that E1b1b is dominant in Africa and J1 in Asia, where one declines the other does too, reinforcing the idea that E1b1b in the Balcans (and otherwise in Europe) was caused by some very accidental founder effect (as it's not meaningfully associated with J1 there).

I miss sampling of Iraq, Kurdistan and Turkey anyhow. But in any case it seems Syria/Lebanon is transitional between the J1/E1b1b area and the J2 one.

The distribution of R1b and G appears somewhat patchy. R1a would look the same but the Iranian focus makes the rest appear unimportant.

I've been checking the R1b haplotypes and most seem within the most common ones in R1b*-Ht35 and R1b1b2a1. There are exceptions but are numerically less important.

Dienekes said...

It would seem more related with the Neolithic area of Zagros-Taurus if anything

The J2/J1 ratio exhibits maxima in Europe, in Iranian and former Iranian speakers of Central Asia, and in South Asia. It is negligible in North Africa and the Arabian Peninsula and has intermediate values among non-Arabian Arabs from old Persian-Hellenistic-Roman territories.

The simplest explanation for the current distribution is that J2 was found at a high frequency in a belt encompassing Greco-Italic, Paleo-Balkan-Armenian, Anatolian, Indo-Iranian speakers (resulting in a high J2/J1 ratio), while in the Middle East itself, the arrival of multiple and well-attested waves of Semites from the south introduced J1 in higher quantities, resulting in lowered J2/J1 ratios in that region.

Vincent said...

I've been checking the R1b haplotypes and most seem within the most common ones in R1b*-Ht35 and R1b1b2a1. There are exceptions but are numerically less important.

Two are R-M73 (with DYS390=18 and 19), the first of this type we have seen in Egypt. Additionally, there are a handful of others (maybe 5-8) that seem to be R-P25*, though with the STR panel used here it is hard to say for sure.

VV

Maju said...

Two are R-M73 (with DYS390=18 and 19), the first of this type we have seen in Egypt.

That's interesting. Thanks.

Maju said...

The J2/J1 ratio exhibits maxima in Europe, in Iranian and former Iranian speakers of Central Asia, and in South Asia. It is negligible in North Africa and the Arabian Peninsula and has intermediate values among non-Arabian Arabs from old Persian-Hellenistic-Roman territories.

But that is an illusion, a tendentious interpretation.

In Iran J2 is clearly found together to R1a (and that would be the case in South Asia too). But that doesn't happen in Europe nor in the rest of West Asia.

The distribution of J2 is south of the Steppes, essentially between Italy and Western India (with J2a being dominant at the middle and J2b at the extremes). This distribution is what we would expect of a Neolithic lineage that expanded from a West Asian core, a highlander one to be precise. In Europe (but not South Asia) it is roughly similar to that of E1b1b, the other apparently Neolithic Y-DNA lineage.

Unless you want to equate Neolithic with Indoeuropean, like the discredited Renfrew, you can't go far through this line of thought.

J2/E1b1b (Neolithic) and R1a (Chalcolithic/Bronze/Iron IE) are different prehistorical processes.

Kepler said...

Is there a consensus of any sort in any group about what "Indoeuropean" is? Doesn't it have to do more with linguistics than ethnicity?

Maju said...

IE is above all a linguistic family, sure.

But as for the genesis and spread of this language family, the Kurgan theory is widely mainstream and the only one that gives satisfactory answer.

Linguistically speaking, anyhow, we know something from a lot of "Neolithic" languages from Iberia to India, most of them extinct now, and none of them looks IE.

Dienekes said...

In Iran J2 is clearly found together to R1a

The J2/R1a composition in Iran is similar to that of the Balkans. Indeed, in the belt I spoke of, we have a clear underrepresentation of R1a in Italy, an over-representation in South Asia, and similar proportions between Greece and Iran.

But, I really don't see what R1a has to do with anything in this discussion.

The distribution of J2 is south of the Steppes

J2 is well represented in the steppes, certainly as much as it is in South Asia or most of Europe outside the Mediterranean.

Unless you want to equate Neolithic with Indoeuropean, like the discredited Renfrew, you can't go far through this line of thought.

Renfrew has not been discredited, and he never equated Neolithic in general with Indo-European.

J2/E1b1b (Neolithic) and R1a (Chalcolithic/Bronze/Iron IE) are different prehistorical processes.

R1a is 1.5-2 times as diverse in Greece, the Balkans, and South India than it is in Russia and the Ukraine, so it does not by itself represent "Chalcolithic/Bronze/Iron IE", although yet-to-be discovered subsets thereof may be associated with steppe processes.

Gioiello said...

In a previous mail to Klyosov I said that probably R1a in Greece, Balkans, India (perhaps not only South)is a first and older dispersion of the Hg. That that of Russia/Ukraina is younger doesn't invalidate the Kurgan theory and the Indo-European one, which was carried perhaps more recently by the younger R1a. Interesting is the scarce presence in Italy, but you know I think it is because there was R1b, that I have always linked to a more ancient subset of Indo-European, i.e. the Rhaetian-Etruscan-Pelasgian.

Maju said...

Renfrew has not been discredited, and he never equated Neolithic in general with Indo-European.

From Wikipedia - Anatolian hypothesis:

"The Anatolian hypothesis is also called Renfrew's Neolithic Discontinuity Theory (NDT); it proposes that the dispersal (discontinuity) of Proto-Indo-Europeans originated in Neolithic Anatolia. The hypothesis suggests that the speakers of the Proto-Indo-European language (PIE) lived in Anatolia during the Neolithic era, and associates the distribution of historical Indo-European languages with the expansion during the Neolithic revolution during the seventh and sixth millennia BC".

(...)

"While the Anatolian theory enjoyed brief support when first proposed, the Indo-Europeanist community in general now rejects it, its majority clearly favouring the Kurgan hypothesis postulating a 4th millennium BC expansion from the Pontic steppe. While the spread of farming undisputedly constituted an important event, most see no case to connect it with Indo-Europeans in particular, seeing that terms for animal husbandry tend to have much better reconstructions than terms related to agriculture. The linguistic community further notes that linguistic evidence suggests a later date for Proto-Indo-European than the Anatolian theory predicts".

So, hmm, yes, J2 fits well with Neolithic spread as described in the Renfrew model but not with Indoeuropean expansion, which is a different and later process.

R1a is 1.5-2 times as diverse in Greece, the Balkans, and South India than it is in Russia and the Ukraine, so it does not by itself represent "Chalcolithic/Bronze/Iron IE", although yet-to-be discovered subsets thereof may be associated with steppe processes.

I am aware of the diversity issue in India but was not of the same happening in the Balcans. Never heard of it in fact.

IMO, the possible India origin for R1a1 does not contradict a main expansion within the Kurgan and post-Kurgan (Scythians, Slavs) migrations from the steppe: there is a mayor haplotype cluster, apparently much derived, that seems to have experienced most of the expansion attributable to the haplogroup. This cluster would fit perfectly with the Kurgan migrations.

Ukraine, Poland or central Russia are in any case not the original homeland of Kurgan cultures but rather it is at the very eastern edge of European Russia. These areas were nevertheless early conquests for the IE horsemen and had (excepted Ukraine probably) a low population density prior to that conquest, what explains the high Y-DNA impact. Western IEs (apparently affecting specially to Central Europe and most of Russia) are, archaeologically speaking, a well defined subgroup of all IEs (Baalberge-Corded Ware cultural group). I would expect that area to have a subset of the whole R1a1 variation, not the bulk, even if we focu only in the haplotype cluster that is probably most directly related with Kurgan IEs.

But, I really don't see what R1a has to do with anything in this discussion.

You were the one associating J2 with IEs. It's only logical that R1a pops up in any IE-related discussion, right?

terryt said...

"The paper seems to demonstrate that hg. R expanded from Europe to North Africa and Middle East, being present on the coasts and diminishing when we go inside".

Perhaps providing support for the idea that the Danaan of Troy, the Denyen of the so-called Sea People and the Israelite tribe Dan are one and the same. The disruptions at the eastern end of the Mediterranean had reverberations in the west where people appear to have reached the Canary Islands at around the same time.

Maju said...

Ahem Terry: the Danaans were Greeks (same as Achaeans, Argives, it seems), not Trojans.

Anyhow, what I see is a patchy distribution of R1b, with highest concentrations in Kuwait, Jordan, Lebanon and Iran. But lowest concentrations among the other coastal people: Palestinians from Akka (Acre), who live more or less where the proto-historical Dan tribe lived.

And let's not forget that Upper Egypt and Sudan also have rather strong R1b (and then also Chadic speakers and Uyghurs and of course, Turks, Armenians and Europeans).

As for R1a (a different story) is basically concentrated in Iran, and then has also a patchy distribution that is almost the inverse of R1b.

I'd say that the focus of R1a in the area is in Iran while the R1b is less clear but could be Turkey perfectly:

R1b-DYS393=12 (should be mostly R1b1b2a1) is found mostly in Syria and Lebanon (near Turkey) and largely with haplotypes that appear commonly in Middle Europe (though there are several Lebanese haplotypes I could not find alikes).

R1b-DYS393=13 (Ht35) seem all derived from the "Antolian" modals.

And then you have some others: DYS393=14 is the most common (Lebanon/Syria). The DYS393=11 is also relevant because it includes most of the Kuwaiti R1b, though it's also found in Lebanon in another variant.

Does anybody know which STRs would predict R1b1c, which is reported so far only in Lebanon? Does anybody know if we can consider DYS393=14 derived from DYS393=13 and hence within (probably) R1b1b2a1? There's at least some Atlantic Europeans who also have that value and whose haplotypes would seem derived from other European ones.

Andrew Lancaster said...

The new paper by El-Sibai et al adds to work previously done by a similarly composed team of authors:
[list]
[li]http://www.haplozone.net/wiki/index.php?title=El-Sibai_et_al._%282009%29
http://www.haplozone.net/wiki/index.php?title=Zalloua_et_al._%282008b%29
http://www.haplozone.net/wiki/index.php?title=Zalloua_et_al._%282008a%29[/li]
[li][/li]
[/list]
When looking back over their rather complicated data tables (different data is from different places, and was apparently tested in all different ways), I discovered that I could extract this:

Zalloua et al. (2008b) found 26 E-M123 cases out of 164 Cypriots. This was apparently higher than the level of E-M78. Also see the same data set in El-Sibai et al. (2009) for clarity. The same Zalloua et al. study reports 27 Palestinians out of 291 tested, which the El-Sidai summary makes clear contained 100 M123 tests done on Palestinians from Acre, of whom only 6 were positive.

This data will be added to the collection here: http://www.haplozone.net/wiki/index.php?title=E-M123_data

The high levels in Cyprus are very interesting!

I do not think any paper reported this before. M123>M78 is a rare pattern also seen in parts of Turkey.

Best Regards
Andrew

Andrew Lancaster said...

Maju, why do you specify J1 being more broadly Afroasiatic than just Semitic in North Africa? Doesn't its distribution match Semitic languages better than broad Afroasiatic?

Maju said...

It's presence in North Africa seems too common and too old (structurally) to be of recent Semitic arrival (i.e. Medieval Arab basically). Actually, per Cruciani, the main sub-cluster of J1 seems rather centered in North Africa.

But the definitive evidence was provided by presence of loads of J1 among Canarian Aboriginals, who were never influenced by Semites but were wholly within the Berber linguistic/cultural area. It is the same as for mtDNA L(xM,N), whose presence in Guanche aDNA clearly indicates it's not only medieval/modern "slave trade" originated in North Africa.

So, as I see it, the process of genesis of Afroasiatic peoples happened at Upper Egypt or not too far away from there (Sudan?). And involved E1b1b1 as main original Y-DNA haplogroup but soon also incorporated "Asian" J1 and R1b, which in turn were carried to NW and Central Africa by the various Afroasiatic branches (Berber, Chadic). Similarly, E1b1b1 was brought to West Asia at the genesis of Semitic peoples and/or Meso-/Neolithic.

I can't be more precise than that because some of the archaeology involved is still confusing. But hopefully you get the idea, right?

Dienekes said...

I do not think any paper reported this before. M123>M78 is a rare pattern also seen in parts of Turkey.

Which populations (in Turkey or otherwise) have the highest M123 frequency?

Dienekes said...

But the definitive evidence was provided by presence of loads of J1 among Canarian Aboriginals, who were never influenced by Semites

The presence of J1 in the Canary Islands could be due to the Phoenicians, so the conclusion that it is non-Semitic is hasty.

Only if J1 is found in pre-Arab/Phoenician N Africans, then the simplest explanation will be that the Canary Islands were drawn from a N African J1/E1b1b population. Until then, the hypothesis that N African J1 is primarily Arab/Phoenician/Jewish and not indigenous is not really disproved.

Maju said...

The presence of J1 in the Canary Islands could be due to the Phoenicians, so the conclusion that it is non-Semitic is hasty.

Phoenicians never had any colony in the islands and J1 is way too high in Guanche aDNA to be a random Phoenician erratic. Instead Phoenicians had many and very old colonies and even the bulk of their late empire in Iberia and J1 is very rare there. The Iberian counterpoint can also be used in regard to Arab influence.

Occam Razor says J1 in North Africa is not Semitic (though could well have been carried there by Afroasiatics - Capsian Epipaleolithic and/or Capsian Neolithic).

Only if J1 is found in pre-Arab/Phoenician N Africans, then the simplest explanation will be that the Canary Islands were drawn from a N African J1/E1b1b population.

That is what Guanche remains are: there's no Phoenician or otherwise material culture in pre-Spanish Canary islands: it's all typical North African.

Anyhow, sure: the more evidence, the better. But, sincerely, we have been waiting for many years for UP Europeans to be sampled and so far only one specimen has been tested for mtDNA alone. So I'm not in the mood of waiting another decade or two before we have that kind of evidence you demand.

Maju said...

Erratum: there's no Phoenician or otherwise material culture... should read there's no Phoenician or otherwise Semitic material culture...

Dienekes said...

J1 is not particularly high in the Aboriginal populations and is wholly consistent with occasional matings between Phoenician men and local women. A rich Phoenician man, carrying all sorts of goods for trade would be quite popular with the local women.

Similar patterns of European Y-chromosome introgression in various native peoples are observed throughout the world, e.g., among remote native Polynesians or Americans.

even the bulk of their late empire in Iberia and J1 is very rare there.

Iberia is a big chunk of land with solid agricultural populations. Phoenicians would have no substantial effect there, unlike small islands with backward populations.

Dienekes said...

There is no "material evidence" in many places that Phoenicians traded in, e.g., Cornwall. Nor is there any evidence for many of the hundreds or thousands of ancient Greek emporia (makeshift trading stations) throughout the Mediterranean, yet the ancient authors did not invent their existence.

Maju said...

16% J1 among Guanches, Dienekes, not just quite high but also in the right apportion in relation to E1b1b1 as found in the mainland.

Also if Phoenicians are Lebanese, what is correct grosso modo, then they should have brought more J2 than J1 to the area or at least much more variety. The quite strict difference in the presence of J1 and J2 at both sides of the Mediterranean, strongly suggests it's not related with the irregular Phoenician colonization, concentrated in some coastal spots both in North Africa (but not Egypt nor Cyrenaica), Iberia and the Italian islands.

It could all be a coincidence but neither I nor Occam believe in them too much.

Iberia is a big chunk of land with solid agricultural populations.

North Africa is larger than Iberia and has similarly old solid agricultural populations that were for the most part never conquered by Carthage (unlike Iberians), yet you're claiming that Phoenicians (and curiously Phoenicians with only Y-DNA J1 but not J2) altered their genetics up to as much as 25%. And you still have to explain some 25% J1 in Egypt... even more agricultural than either Iberia or North Africa and never colonized by Phoenicians in any way.

And Phoencian descendants: Lebanese are significatively more J2 (J2a mostly I think) than J1. While the amount of J1 found among ancient Guanches is similar to that of Morocco, West Sahara or Algeria, areas never too intensely colonized by Phoenicians. Excepting Algeria somewhat, these countries are also very low in other "Phoenician" (Lebanese) lineages like J2, G or R1b.

Even in the highly unlikely case that all the Phoenicians operating in the Morocco area came from Acre (comparable to the Palestinian sample of this paper), they should still carry half as much J2 as J1. But most likely they were form Tyre more than from anywhere else.

The overall relative lack of J2 in North Africa (and among Guanches in particular) strongly suggest that J1 came from an area without J2, like Egypt. And in order to have arrived from Egypt, it should be prehistoric (Caspian, Neolithic).

As I see it, Phoenician coastal commercial colonization should not have left in most places more mark than Portuguese colonization of Africa or the Eastern Indies. Maybe there are some exceptions but I can't find them.

There is no "material evidence" in many places that Phoenicians traded in, e.g., Cornwall. Nor is there any evidence for many of the hundreds or thousands of ancient Greek emporia (makeshift trading stations) throughout the Mediterranean, yet the ancient authors did not invent their existence.

Maybe but short-lived emporia like Hemeroskopeion don't left any mark, nor probably did Phoenicians buying tin at the Scilly Islands. But anyhow, the Canary Islands had little to offer to traders and were at the very edge of the Western Eurasian oecumene. They seem to have been left ignored for the most part.

Dienekes said...

Also if Phoenicians are Lebanese, what is correct grosso modo, then they should have brought more J2 than J1 to the area or at least much more variety.

The Lebanese are not Phoenicians, and there is no reason to think that J2 was represented among ancient Semites; it is not found at any substantial frequency in Arabians, Tunisians (settled by both Arabs and Phoenicians), or Ethiopians (setted by Ge'ez Semites).

And Phoencian descendants: Lebanese are significatively more J2 (J2a mostly I think) than J1.

They stopped speaking Phoenician in Lebanon 2,000 years ago, and there is no reason to think that modern Arabs from the region are particularly descended from the ancient Phoenicians. If Lebanon was some island which may have retained its gene pool despite a shift in language, perhaps one could argue for continuity, but as things stand, that is not possible.

Even in the highly unlikely case that all the Phoenicians operating in the Morocco area came from Acre (comparable to the Palestinian sample of this paper), they should still carry half as much J2 as J1. But most likely they were form Tyre more than from anywhere else.

Your assumption that Phoenicians carried a lot of J2 is unfounded. Indeed, the paucity of J2 compared to J1 in North Africans is one more reason to doubt that they did.

Dienekes said...

it is not found at any substantial frequency in Arabians, Tunisians (settled by both Arabs and Phoenicians), or Ethiopians (setted by Ge'ez Semites).


We can add the Soqotri to that list.

Kepler said...

Phoenicians were not Lebanese? Well, there was no Lebanon back then, but how do you get to that conclusion?
Were they all exterminated in that region? Where did this J2 come from? Saudi Arabia?
The Phoenician culture was made Aramaic, which is not such a big change as both are Semitic, that culture got some Hellenization for a while and then the Arabs arrived with another Semitic language, which would mix with the Phoenician-Aramaic stratum. Lebanese Arabic is said to be a mixture of these languages

Dienekes said...

Phoenicians were not Lebanese? Well, there was no Lebanon back then, but how do you get to that conclusion?

As multiple studies have shown, the conclusion that modern gene pools reflect those of the ancient inhabitants is not generally true. So, until evidence to the contrary is presented, I see no reason to think that modern Lebanese Arabs have similar haplogroups frequencies as ancient Phoenicians.

Were they all exterminated in that region? Where did this J2 come from? Saudi Arabia?

Hard to come from Saudi Arabia, as it has almost none. J2 may have come from Italy, the Balkans, Anatolia, or Persia, all of which dominated at one time or another the region.

Kepler said...

It seems to me you apply the opposite to Ockham's principle here.

I don't think haplogroup distributions now are the same as before, but in some cases they give some idea, specially if you see the distribution map of J2. Iraq is chock a block with them. Are they also some Roman or Turkish influence?

There is no single nation that is the sole "owner" of any haplogroup, but it seems to me you prefer to believe Phoenicians simply disappeared from their original centre. Even if they did disappear politically from there, it is not that they were just keeping among themselves and then all decided to sail to Cartago.
Phoenician was extremely similar to Hebrew and Hebrew was a Canaanite language. I would imagine Phoenicians were not just secluded to their main cities

It seems strange to me that the inhabitants of a successful civilization which also was very related to others around would just vanish from their homeland...unless there is massive genocide or the like (which is what happened in many parts of Latin America and still there over 50% of women have A to D haplogroups.

Dienekes said...

It seems to me you apply the opposite to Ockham's principle here.

Not at all. Substantial genetic change seems to be the norm in almost every study that has compared ancient with modern populations from the same region. Therefore, the simplest explanation is to expect the modern Lebanese to be substantially different from the ancient inhabitants of that country, rather than the opposite.

Iraq is chock a block with them.

Semites are not native to Mesopotamia. J2 could very well precede the Semites there, and Iranian speakers could very well have added more later, given that they have a high J2/J1 ratio.

but it seems to me you prefer to believe Phoenicians simply disappeared from their original centre.

Not at all. Phoenicians were probably absorbed by other Semites of the region, and other non-Semitic peoples were added to the mix, just as the Phoenicians were added to pre-existing populations of the region. I don't see any justification for thinking that the Phoenician element is predominant in the genetic makeup of the current population compared to the pre- and post-Phoenician ones.

It seems strange to me that the inhabitants of a successful civilization which also was very related to others around would just vanish from their homeland

Well, Alexander and his successors killed a bunch of them and took over their trade. Soon enough, Phoenicians ceased to be reckoned as a distinct people or to speak a distinct language. As there were no natural barriers between them and the rest of the Levant, this would have led over time to an adulteration of their original gene pool, just as the Etruscan gene pool was adulterated when the Etruscan language and political autonomy ceased to exist.

Andrew Lancaster said...

Dienekes, M123 in Turkey is higher than M78 in Cinnoglus areas 3 (eastern black sea coast), 5 (border with Syria), 6 (south coast), 7 (centre incl Ankara), and 9 (Istanbul). Putting aside the metropolis this distribution looks like the area most long in contact with the Fertile Crescent? That would fit OK with Cyprus of course. I think M123 is also higher in other scattered areas, for example perhaps Oman, and maybe even NW Iberia.

As my review article explained I see reasons to believe E-M123 spread in a way which we associate today with Semitic peoples like Phoenicians and Jews, but it may have been an earlier wave of people before the history we know.

Maju said...

Obviously we are looking at this with very different lenses: for Dienekes, Lebanese can't be Phoenicians because they have too much J2, unlike Tunisians but like southern Iberians, who in turn virtually lack J1. They can't be Phoenicians because modern peoples have nothing or almost nothing to do with ancient ones, almost by definition.

For me (or Kepler) the opposite is true instead: Lebanese should be mostly descendants of ancient Phoenicians, regardless that some minor inputs may have also arrived in these two milennia - most of them Semitic, btw.

J2 may have come from Italy, the Balkans, Anatolia, or Persia, all of which dominated at one time or another the region.

Then why other lineages so important in these regions are not found or found at levels much lower than would be expected in modern Lebanon?

Iran for instance has 25% J2 but also 21% R1a but in Lebanon J2 is 28% and R1a only 3%. So it's not Iranian. Would it be the Macedon domination, we should find also R1a and I and E1b1b1-V13 in large amounts. Nothing of all that is there in fact. Same for Rome, mutatis mutandi (for instance Roman influence was much more intense in Tunisia and you see very little J2 there).

More ambiguous can be the Anatolian claim because certainly West Asian J2 seems to have an Anatolian affinity and I'd dare say centrality. But considering the weak influence the other suggested empires apparently had, I'd say that the Ottoman Empire was probably also non-influential or very weakly so. It's surely something older. Something that was there when the historical Phoenicians existed.

I can only go back to Tunisia and Andalusia and the other areas that belonged at some point to the Phoenician Empire: you see too wide differences, specially in the various J lineages: almost all is J1 in Tunisia and other parts of North Africa, regardless of their relation with the Phoenician, Persian, Macedonian, Roman or Ottoman empires, and almost all is J2 in Iberia also regardless of who dominated them historically. The difference runs along the Mediterranean from east to west and does not reflect any historical processes or at least mostly not.

Maju said...

(cont.)

Substantial genetic change seems to be the norm in almost every study that has compared ancient with modern populations from the same region.

I don't see that. Most aDNA studies show in the raw data strong continuity with modern populations, even if some of their authors have reached to biased conclusions. Chandler et al. found almost the same mtDNA in Epipaleolithic/Neolithic Portuguese as modern ones (they decided that there was no continuity somehow though. That other study on Medieval Basque DNA compared apples and oranges, and decided that modern Basques are not Medieval Basques because their aDNA sample showed no mtDNA V. But Basques are very low in mtDNA V except for Gipuzkoans and no Gipuzkoan site was sampled in that study (apples and oranges). Kefi et al. found that UP Oranians from Taforalt had about the same apportion of mtDNA as modern Riffians do, including some JT(xJ,T). The mtDNA found in a Gravettian from Italy seems to be H (could be U but most likely is H1). And so on. Almost every single aDNA study seems to ratify continuity, even from Paleolithic times. True that most are focused in mtDNA but still...

Phoenicians were probably absorbed by other Semites of the region, and other non-Semitic peoples were added to the mix...

If only J1 and E1b1b can be considered "Semitic", then Lebanese are 65% "non-Semitic" (60% if we consider T as "Semitic" too), what is a total nonsense.

Even if you'd be right (you may be right in what regards to the original proto-Semites), it would still mean that the original J1-Semitic influence to the area (Amorites/Canaanites or whatever) was low and that the J2 was in Lebanon and Syria before that and therefore was part of the historical Phoenicians.

Have you even considered PPNB? PPNB replaced Natufian/PPNA in the Levant and is a North to South migration, that could perfectly explain alone the Anatolian affinities of Lebanese and Syrian Y-DNA. But guess it's just too old for you to even consider it.

Semites are not native to Mesopotamia. J2 could very well precede the Semites there...

If the mainstream models of Semitic expansion are correct, Semites were not native to Lebanon either but actually to the semidesert (circum-desertic pastoralist complex) with a possible ultimate origin in Southern Palestine (Harifian).

Ricardo Costa de Oliveira said...

J1 is far more complex than previous thought. Some of the most basal forms are found in the Northern areas just like the 388=13 and the J1b M365. J1 is found in the Saharawi people on the Atlantic edge of North Africa, in Portugal J1 is found in small but heterogeneous types meaning diverse waves of traders and conquerors at regular frequencies not that different of the frequencies of the I haplogroup, for instance. J1 is found in the Northern Mediterranean shores with a lot of diversity and we can find some exquisite J1 haplotypes even in Northern Germany and in the Magyar Plains, not to mention the regular and old frequencies of J1 haplotypes in places like Greece, Crete and the heavy concentrations in some points of the Caucasus, what is not found in other important European haplogroups nowhere else. In Turkey, Anatolia and in Iran J1 is well over 10% what is a significant proportion in areas with several haplogroups sharing the same space, and of course the J1 P58+ L147+ concentrated realm to the South in areas with sharp identities like the Arab identity of the Fertile Crescent, the Islamic Arabian Peninsula, the Bedouins and the Jewish Cohanim group, all full of extravagant identitarian passions spreaded far away and always interesting to be seen. J1 is also well represented in agricultural Egypt and in Eastern Africa with some diversity too. Pakistani J1 haplotypes and Arab haplotypes went to the Indic Ocean gone with the centuries.

Gioiello said...

Ricardo, if you look at Junusbaev Bayazit Bulatovitch, Populyatsionno-genetitcheskoe issledovanie narodov Dagestana po dannym o polimorfizme Y-khromosomy I alu-insertsii, you can see, at page 19, that the centers of dispersion of J1 are: one in the middle left shore of the Caspian Sea, i.e. Dagestan, and a second, less important, in the Arabian Peninsula. As the centers of dispersion of J2 are, from West, North Italy, Bosphorus, Central Turkey, North West Iran/South West Caucasus, South East Iran, and since J1 is linked to J2 and both with I, we can hypothesize that also J1 is from the regions around the Black Sea and not from South, where it migrated in a second moment. Then you can consider yourself the true J1, that linked to the original.

Dienekes said...

Then why other lineages so important in these regions are not found or found at levels much lower than would be expected in modern Lebanon?

Once again, you are making the dubious assumption that modern populations are like ancient ones. You are assuming e.g., that modern inhabitants of Iran are like ancient Persians or Medes or the other Iranian-speakers that were part of the Persian empire.

The same type of "logic" has led some Indian researchers to conclude that there was no Central Asian input into South Asia, because India lacks the various C, Q, O subclades that are found frequently in Central Asia today, ignoring the fact that these were added later in the population.

PS: Your numbers for Iran are wrong. R1a1 is found at a frequency of 3% in north Iran and 15% in the south ("Iran: Tricontinental Nexus for
Y-Chromosome Driven Migration").

Maju said...

Once again, you are making the dubious assumption that modern populations are like ancient ones. You are assuming e.g., that modern inhabitants of Iran are like ancient Persians or Medes or the other Iranian-speakers that were part of the Persian empire.

I understand that IE/Kurgan migrations pre-date the historical Persian and Macedonian empires, hence it is a most reasonable assumption of mine that ancient Iranians (Persians, Medes or whatever else within the geography or modern Iran) were already 21% R1a. Maybe you think it's something brought by the Mongols? Or maybe R1a and I among Northern Greeks is something borught by the Slavs only?

PS: Your numbers for Iran are wrong. R1a1 is found at a frequency of 3% in north Iran and 15% in the south ("Iran: Tricontinental Nexus for
Y-Chromosome Driven Migration")


I am using the data of this paper: 19/92=0.2065, rounded to 21%. Whatever the case, Iranians still have less J2 and much more R1a than Lebanese.

The only affinity I see clear would be with Anatolians.

Ponto said...

I am finding this pseudo scientific argument of who is European/White and Who is Asian/Brown somewhat irritating.

All haplogroups in Europe are from Asia: the other side of the conceptual division of Europe (derived from an Asian princess Europa). If you know Old Grecian legends Andromeda was Ethiopian. Asia is the other side of the Aegean i.e Turkey eastwards.

I have said this before: Wells and Zalloua cannot be trusted as they have Whitewashing tendencies regarding Lebanese people. Wells is a publicity seeking dilettante.

The fact J1 is predominate in Semitic language speakers, and J2 predominate in a minority of Europeans (the black haired brown skins of the South) is not significant. Founder effects in both populations, the Brown skins of Europe and Asia. Genetically both type of dark Whites have more in common with each other than the pasty skins of Northern Europe. In any case as Maju said you cannot extrapolate backwards in time. Europe thousands of years back was a totally different place with totally different inhabitants some of whom are ancestral to present day Europeans, many did not leave descendants or have been out bred.

It appears that all the studies focusing on Middle Eastern and North African populations have been ignored. The majority of J1 in North Africans is of Neolithic age due to movements of people in the Holocene period. South Arabian J1 is derivative of Northern J1 and lacks diversity when compared to Northern J1. J2 is as Middle Eastern as J1. Haplogroup I originated in the same locale as J1 and J2. European and Anatolian J1 is small in percentage but is more genetically diverse than any in Arabia, or North Africa. Tofanelli showed that fact. Genetic diversity means older. Anatolian J1, and the J1 found in Italy and Iberia is older than that found in Arabia. Ethiopian J1 is older than Arabian J1. Lebanese J1 is older than Arabian J1. Caucasian J1 cannot be considered due to extensive inbreeding in those mountains. Take the time to read these reports.

R1b the common European haplogroup is Asian in origin and less than 18 ky old. It most likely entered Europe along with the IE speakers which is why it ended up hemmed in the western littoral of Europe. R1a entered later with proto Slavic speakers. Haplogroup I is the sandwich between the two R groups but has a SE to NE pattern of spread from the Balkans to Scania.

It is about time that studies trying to prove who is whiter than whom or less Asian originated than whom should stop. All Europeans are descended on the male side from men who entered Europe after the Holocene started and after agriculture was invented in Asia.

The female side of Europeans is similarly Asian. Haplogroup H and U are Asian in origin barring some derivatives like H1, H3, V and Ub5.

Get off your high horses and start reading and understanding what you read.

Maju said...

Genetic diversity means older.

Not necesarily. Brazilian R1b, for instance, must be more diverse than Portuguese R1b and can't be older.

As a rule of thumb is a good idea to think that way because of statistical likelihood. But this is just a statistical effect and certain factors can alter it.

In this case of J1, if West Asian J1 is more diverse than North African and Europe has got its J1 from West Asia (only or mostly), not in a founder effect event but in a gradual chaotic arrival in the last few milennia, European J1 can perfectly be close in diversity to that of West Asia in spite of being recent.

Instead, if most North African J1 corresponds to a founder effect in the Mesolithic (and therefore excludes a significative part of the original West Asian diversity) and has not received much other inputs since then (or inputs with limited diversity as well), the lineage will be both less diverse but still older.

In a sense we would be comparing apples and oranges, because the processes involved would be too different to compare that way.

So word of caution on this.

Maju said...

Also:

Haplogroup I originated in the same locale as J1 and J2.

How do you know. The divergence of IJ into pre-I and pre-J almost necesarily happened "in the same locale" but from that to the formation of I and J as such (their diversification) passed some time and probably space too.

In fact I understand that each branching of a lineage out almost necesarily implies a branching out of the carrier population; otherwise drift would cause normally all but one derived lineages to be eliminated quickly under low population conditions. Only in two different groups could two different lineages survive normally.

It is very probable that the formation (consolidation, diversification) of I happened already in Europe. And IMO at about the same time J diversified in West Asia.

Maju said...

And also:

R1b the common European haplogroup is Asian in origin and less than 18 ky old.

Why?

It most likely entered Europe along with the IE speakers which is why it ended up hemmed in the western littoral of Europe.

This is an oxymoron. The populations with highest R1b are the ones that got indoeuropeanized the latest. There were no Celts or any other IEs in the Britain or Ireland before 300 BCE, nor in most of Iberia before 700 BCE. Not to mention Basques, of course.

R1a entered later with proto Slavic speakers.

From India? O_O

All Europeans are descended on the male side from men who entered Europe after the Holocene started and after agriculture was invented in Asia.

Impossible. You'd have to prove that with meaningful aDNA sampling before I can swallow it.

Haplogroup H and U are Asian in origin...

I am not persuaded in regard to H (nor V for the case). R0 is surely West Asian but H and its derivate lineages are essentially found in Europe, with the only exception of H8 (Central Asia). Furthermore, aDNA evidence strongly points to its antiquity.

Gioiello said...

Neither R0 is fully Asian. I made a relative of mine test at SMGf, he is a Tuscan, and his R0a is very different from all the others. Then there was a Tuscan (Etruscan) R0a that exists in Italy at least from 10,000 years or more. This haplotype lacks also in Turkey, where someone thinks Etruscan came.
The wrong thing is to think that and haplogroup is exclusive of a region of the world or of an ethnic group.
As I have said many times in the past, if R0a is in Italy from 10,000 years or J1 from 6,000 years they are Italian like mtDNA H found 24,000 YBP.

Maju said...

I did not mean R0 is exclusively Asian but originally West Asian, as it's the case of HV probably too. But IMO not of H or V. What I mean is the place of origin or coalescence of the haplogroup.

terryt said...

"In fact I understand that each branching of a lineage out almost necesarily implies a branching out of the carrier population; otherwise drift would cause normally all but one derived lineages to be eliminated quickly under low population conditions. Only in two different groups could two different lineages survive normally".

Thanks for that. It's taken a long time.

"The populations with highest R1b are the ones that got indoeuropeanized the latest".

Exactly. So that puts R1b in Western Europe quite early. Gravettian? Or Danubian?

Maju said...

Danubian does not correspond geographically in any way. There's no Danubian South or West of Paris nor North of Bremen. Instead there are Danubian cultures in Poland, Western Urkaine or even Bulgaria. It's a Central European culture.

terryt said...

Gravettian then?

Maju said...

At this stage of the research I still don't know for sure but Paleolithic in any case, Epipaleolithic the latest.

Ponto said...

What closed and prejudiced minds some of you have. In fact you think like Semitic language speakers think: Fanatically and dogmatically. Muhammad would be proud of you.

Y chromosome haplogroups are older than any language groups. Tofanelli et al showed that the peoples who are the most genetically diverse in J1 are not from Arabia or European Jews or speakers of Semitic languages. They are Kurds, Iranians, Anatolians, Europeans like Portuguese and Italians, and Ethiopians. The later group is Semitic language speaking, probably an accident of history, like Finns and Hungarians speaking essentially Asian languages more in common with Japanese and Korean. The evidence based on genetic studies points to a Northern Southwest Asian origin for J1 at a time when there were no Semitic speakers or Jews. Semitic language presence in the Middle East does not go that far back in time. Of the Arabian language speakers the Lebanese, Omanis and Tunisians have a good diversity in their J1. The Lebanese outdoing the Omanis. Despite what Arabian folklore says, Arabia was seeded after the Neolithic and from the North.

I think you lot with your R1b obsession, it being European in origin and with your language fixation or your combining of haplogroup with ancient cultures are wrong in a big way. Face the facts. R1b, R1a are not European in origin just common in Europe. Haplogroup I, the brother haplogroup of J, is almost wholly restricted to Europe but it is not of European origin either. Unless you can find those Paleolithic skeletal remains that are R1b or R1a or mtDNA H I suggest you stop making things up. If Tofanelli was wrong about J1 then prove it so. Prove Paleolithic European were R1b, prove the makers of certain stone tools were R1b. The Cro Magnons looked Caucasoid but their skin color was most probably dark brown, and their faces have more of an Amerindian look than European. The Grimaldi folk had their own different look again. Not so long ago in European history, the population was devastated and reduced to a third of what if formerly was by a disease from Asia called the Black Death. Is it not possible that the present day distribution of European haplogroups has more to do with disease like the Plague than any fanciful Paleolithic or Neolithic continuation. Think about it.

Maju there is a saying in English which is not your native tongue: A little knowledge is a dangerous thing.

Maju said...

Unless you can find those Paleolithic skeletal remains that are R1b or R1a or mtDNA H I suggest you stop making things up.

Would I have the means, I'd be making such tests right now. In fact I would have already made them years ago...

Anyhow, mtDNA H has been found quite apparently in three separate Paleolithic cases:

1. The CRS HVS sequence of Gravettian Apulia (southern Italy) that should be by modern standards H1 most probably (though could be either H or U).

2. Most of Taforalt (Oranian culture) remains of Northern Morocco (Kèfi 2005).

3. 5/9 Portuguese Epipapaleolithic remains (Chandler 2005).

So I understand that mtDNA H is clearly confirmed to have been in Iberia and North Africa since at least 12,000 years ago (date of the Taforalt remains) and was probably around since the Gravettian period or before.

As for the rest, I do not claim that Y-DNA R1b is of European origin (though I'd do for I) but I do claim that R1b1b2a is of European origin quite clearly.

I do not claim that J1 is Semitic or Afroasiatic strictu sensu but I do claim that it spread (along with E1b1b1) together with Afroasiatic languages in many instances. They do have a correlation even if they can't be related in all cases.

I agree that haplogroups, at least at the level that we're talking here, are older than any language family we know of. But that means they are really old: Paleolithic.

Not so long ago in European history, the population was devastated and reduced to a third of what if formerly was by a disease from Asia called the Black Death.

That's simply false. A brutal exaggeration. The black death may have caused such casualties in some very localized cases but for the most part it was more a matter of a mere sudden and short-lived brutal increase of mortality. Some areas (Belgium, Carpathians, NW Pyrenees) were almost fully spared.

From the few known cases, Millau (Auvergne) lost 1/3 of people to the Black Death, some Italian cities (Volterra, San Grimigano) show a brutal decrease of maybe as much as 3/4, Toulouse is believed to have lost 1/3. But in Hungary and Poland the population grew dramatically in that same period and doesn't seem that the plague had any meaningful impact. The loss of population, where known, can't also be only attributed to the plague, as it was a quite dark period full of conflicts and when, maybe partly because of the increased mortality, people gained much mobility and emigrated ofte. In fact it was often believed that cities' "foul air" was the cause of the plague and hence all who could fled to the country, so much of the documented population decreease in urban areas may not have been because of deaths but because of people running away.

I'd estimate that the real population decrease may have been at most of 1/3 and probably quite less (10-20%), though it did affect specially to the young and hence affected fertility for some generations. As "bottleneck" the Black Death was at most a very mild one.

In fact I do not believe in "bottlenecks" and when I find that term it's often applied to founder effects and drift, not true bottlenecks (i.e. death of at least 90% of the people).

Gulu said...

J2 in Iraq-Lebanon

The records fo Pre-Semitic populations show the people of Mesopotamia as Agriculturalists (J2 Sumerians) struggling with migration problems by the Pastrolists (J1 Semites)

So J2 in in Pre-Semitic Mesopotmia must have been higher & continued as an important genetic element even after Semitization into Semitic speaking Akkadians who still had a high % of J2 amongst them in addition to the new J1 input, a Century later those Akkadians invaded Lebanon adding more J2 to the earlier arrived J2 from Anatolia.

J1 % kept increasing with the arrival of more Nomadic Semites out of Arabia not just single wave after Islam with occassional big migrations (continued migration from 3000 BC to 1000 AD). Naturally Lebanon & Iraq held on to higher % of J2 because they were fertile enough to support a heavily populated civilized population & later re-enforced by more J2s who were able to prosper in such envt(Greco-Roman in Lebanon & Persians in Iraq)


E1b1b1c-M123

IMO is the Proto-Semitic lineage, its found mainly with Western Semites & its split from the Coastal Dwelling M78 by the Sarawat mountains that extend from SW Arabia to Southern Anatolia, Cyprus is directly located infornt of the only coastal preach the highlands make into the Med Sea (Byblos) which could mean the spread of M123 into Cyprus was much older

Gulu said...

Once again, you are making the dubious assumption that modern populations are like ancient ones. You are assuming e.g., that modern inhabitants of Iran are like ancient Persians or Medes or the other Iranian-speakers that were part of the Persian empire.

J2 people are usually civilized-settled people, who want to hold on to their lands rather than move into another land so usually they quickly assimilate to the invaders (In the Fertile Crescent it was mainly J1 Semites or R1a1a Indo Iranian Speakers)

If you go deeper into the homeland of the Semites & tribal population you notice J1 increases, same with R1a1a the more you move East into Indo-Iranian lands the higher the % of R1a1a reaching big % in Pashtuns

So J2 people are the ones who got caught in between (IE & Semites), you can make a case for J2 having influence on both the Semitic & IE languages