May 08, 2008

The Druze as a population genetic refugium of the Near East

Table S2 has the data.

PLoS ONE 3(5): e2105. doi:10.1371/journal.pone.0002105

The Druze: A Population Genetic Refugium of the Near East

Liran I. Shlush et al.

Abstract

Background


Phylogenetic mitochondrial DNA haplogroups are highly partitioned across global geographic regions. A unique exception is the X haplogroup, which has a widespread global distribution without major regions of distinct localization.

Principal Findings

We have examined mitochondrial DNA sequence variation together with Y-chromosome-based haplogroup structure among the Druze, a religious minority with a unique socio-demographic history residing in the Near East. We observed a striking overall pattern of heterogeneous parental origins, consistent with Druze oral tradition, together with both a high frequency and a high diversity of the mitochondrial DNA (mtDNA) X haplogroup within a confined regional subpopulation. Furthermore demographic modeling indicated low migration rates with nearby populations.

Conclusions

These findings were enabled through the use of a paternal kindred based sampling approach, and suggest that the Galilee Druze represent a population isolate, and that the combination of a high frequency and diversity of the mtDNA X haplogroup signifies a phylogenetic refugium, providing a sample snapshot of the genetic landscape of the Near East prior to the modern age.

Link

12 comments:

Unknown said...

Interesting data.
Word Genetic Refugium was previously referred to India.

Due to Indian Caste system and tribal inhabitation in remote areas some of the haplogroups survived in isolation.

Druze maintaining isolation in the middle east is remarkable.

More research need to be done for mt Haplo x, Y haplo K, L, M and new T

Ebizur said...

It's interesting that this study found haplogroup L Y-DNA at a greater-than-average frequency only among the Druzes of Daliyat al-Karmel and (to a lesser degree) Lebanon. The Druzes from the other areas either completely lacked haplogroup L or only had 2% of it, which is less than or equal to the average frequency of haplogroup L in the Middle East.

Also, most of the Y-chromosomes that the researchers have labeled as "K*-M9" are probably actually T-M70 (formerly K2). Haplogroup K*-M9(xL, NO, P, T) is practically nonexistent outside of Southeast Asia and Oceania.

Ebizur said...

I don't see any genetic connection between populations of Yemen and populations of Oceania or Australia. Even the haplogroup C Y-DNA that is occasionally found in parts of the Arabian Peninsula (but only at frequencies of about 3% or less) belongs to the South Asian subclade, C5, rather than to the Australian subclade, C4, the Austronesian/Oceanian subclade, C2, or the Southeast Asian paragroup, C*. This suggests that the presence of haplogroup C in Arabia is due to relatively recent migration from the Indian subcontinent, which would also explain the presence of haplogroup H Y-DNA at various frequencies in the Arabian Peninsula.

Maju said...

There is one mtDNA clade of Arabia that might correlate with Australia or New Guinea (see: Khaled K. Abu-Amero et al, 2008). It's not very clear anyhow if it's the same haplogroup M-14 or just something related nor how it got there.

Some southern Arabians (Mahra, Socotra) have anyhow looks that remind of Australoids or Veddoids, whichever their origins.

...

Regarding the Druzes, I think it's excessive to claim they are a genetic refugium when they are so intensely inbreeding. The logical conclussion is for me that smaller effective populations would tend to concentrate some clades by mere drift, while erasing others. Instead in the "ocean" of the rest of the region, it's more likely that all clades will survive without major changes, specialy as there hasn't been major demographic events in the region (with the exception of the well identified Zionist colonization) since "the time of the Dawa".

Unknown said...

"This suggests that the presence of haplogroup C in Arabia is due to relatively recent migration from the Indian subcontinent, which would also explain the presence of haplogroup H Y-DNA at various frequencies in the Arabian Peninsula."

H is present all the places surrounding India including Stan countries. It is not due to recent migrations. same is with C. Genetic samples are relatively screened carefully for ancestry.

Indo-Iranian contact zone is separated by just 50 miles and you can see clear difference in Y distribution. In Iran R1a is < 10% and across the border in Pakistan and Afghanistan it is 1/3rd. Mt M is close to 1/3rd in Pakistan and 5% in Iran. The lack of Y H in pakistan is due to Roma displacement etc. L also part of the same pattern.

Inbreeding is common to all the population just before 1000 years. Whatever survived survived.

Shlushli said...

Dear Maju
Thanks for your comments
Indeed the Druze are highly inbreed like many other Near Eastern population now and much more in antiquity. But if Inbreeding is the issue than you would expect low diversity among the Druze and specifically low diversity of haplogroup X, but as you can see the diversity was considerably high and localized to the Galilee region, so what other explanation besides refugium can you give.

The demographic changes of the near East goes far ancient from the recent Israeli inhabitants, I am talking about the Babylon the Assyrians, the Greeks the Romans, the Arabian invasion all of these nations shaped the demography of the near east in the last 3000 years and probably other factors also

Maju said...

Certainly haplogroup X is intriguing but still most of the Druze X haplotypes are within West Asian X2, of which one branch (X2*P5 and derivates) seems unaccounted for and probably represents a formerly existing clade that vanished elsewhere with drift (or, more likely, that is still around in small ammounts but has not been sampled - quite likely). More intriguing are the two X1 lineages but they rather tell me of a founder effect of some sort and maybe an ultmate North African origin.

In this sense, it's important to notice that the Dawa corresponds with 1017 CE, and since that time no major migrations seem to have happened in the area, excepting the 20th century Jewish colonization.

You mention population movements of Antiquity but these happened a lot before the Druze sect was created. Arabs conquered Palestine in the 7th century (and they probably only had a limited impact anyhow, specially in mtDNA, being nomads conquering a densely populated agricultural region). More important surely was the Roman genocide that followed the Great Jewish Rebellion and, previously, the Assyrian "demographic engineering" that deported so many Jews to Mesopotamia and created the Samaritans. But all these are pre-Dawa, long before the formation of the Druzes as separate group.

Furthermore, the authors have not even sampled Galilean Palestinians ("Israeli Arabs") as control. We don't know how frequent these clades are among other peoples traditionally living in the same area. I assume there must be some differences because of the strong founder effects of each Druze village but I cannot accept easily the conclussions that posit Druzes as a living fossil, precisely when they meet all conditions for extreme drift (in each village).

Maju said...

He! I just noticed that you may be the lead author of that study, Liran Shlush, right?

All what is said is with all due respect: you are the expert here and I am just an amateur. And I really appreciate the study anyhow, no matter I may not agree with some of the conclussions.

I also noticed that the second in the list of authors is D. Behar, who has produced some interesting stuff lately too. But who I suspect does have what I described not long ago as "an Israeli mindset". I think there's a (conscious or unconscious) bias among some Israeli geneticists in trying to prove a supposed pure blood bond to Palestine/Israel among Jews (and possibly disprove it for Palestinians), with political reasons behind. Behar particularly has called my attention in this regard more than once, even if I do value his raw data.

I fear that the conclussions of this study have the same kind of bias. It may be a largely unconscious bias but I believe it is there.

I am sure you can do a lot better if you just accept that a large part of Palestinian ancestry is with all likehood from Roman times (Jewish or colonist, I can't say) and that diaspora Jews have probably mixed with other peoples, not just in Northern Europe but all along from the times they were mostly living in Asia Minor.

Science should not be political.

Shlushli said...

Dear Maju.
Thanks alot for your thoutfull response and intriging arguments.

Indeed my name is Liran Shlush.

The article we wrote is scientific, (Introduction, Methods, Results, and most parts of the discussion).

Bias is also scientific and can be quantifies, there are specific kinds of bias that could have caused a false positive high frequency and high diversity of haplogroup X and haplotypews of NRY, and i think that the most importent of them is the sampling method that have not been used in the past.

Unlike what you say you can see that we have used palestinine data (Figure 2), and i can add that haplogroup X frequency among palestinians is 1\110 X1. Most of the samples came from other regions of Palestinians living in Israel and not from Palestininans living in the Galil, but as i understnd from you they should be the same people, they are one nation, though it might be intersting to sample more of the Galilli Palestinians in the same sampling method we used for the Druze.

As you can see in the article (page 2) "In
contrast to the overall high NRY haplogroup diversity for the
entire sample, village based pockets of low diversity were observed,
consistent with local paternal founder effects. For example, the
village of Beit -Jaan included 5 samples from haplogroup K, all of
which belonged to the same lineage (H^ = 0). Similar phenomena
were observed in the village Sajur, in which all 5 haplogroup K
samples belonged to the same lineage (H^ = 0). This pattern of
overall high diversity with local regions of low diversities was also
evident for other villages and other NRY haplogroups, as well
mtDNA haplogroups (data not shown)"

We are fully aware of the local founder effect which is actually even mhore supportive to our hypothesis, of Refugium, we are talking here about Isolates within isolates that preserved part of the regional diversity and decreased the village diversity which is quit differnt from what we see in other populations in the region.

One of the main conclusions of our article which is well supported by Druze oral trditions, is that the Druze as a group has been formed much before the Dawa, otherwise we can not explain by migration (and we have tested various models of migration) the high frequency and high diversity of haplogroup X in the mtDNA and Haplogroup K of the NRY among the Galili Druze.

Actually figure 2 can answer many of the questions you have raised regarding Palestinians. As you can see Palestinians have low migration rates with Druze and high migration rates with other Near eatern populations. This can be true or a false positive results due to the different sampling methodologies.

Yours
Liran

Maju said...

I didn't mean your paper is not scientific, Dr. Shlush, I just meant that there may be bias in your conclussions.

But anyhow, I specially raised the isue because I noticed the presence of D.M. Behar among the researchers and I didn't like at all his conclussions on the Jewish diaspora matrilineal ancestry. I actually flipped a little bit too much when he suggested that H1 among Sephardites of Iberian origin is a "founder effect", when we all know that H1 is a lot more frequent in Western Europe than in West Asia. This in contrast to the rare case of H0 that he acknowledged must be Iberian and the so called "inrogressions" (what an euphemism!) in the practical totality of Indian Jews. Basically he acknowledged no admixture in West Eurasia just because it's easy to be unsure about it, but instead of stating it this way (as uncertainty) he simply suggested mysterious and unlikely "founder effects".

In spite of that, I just mean to criticise his conclussions - not his data nor his methods. These are in any case very interesting and I really welcome that these studies are published in open access format.

This is also the case with your paper: the method is fine, the resulting data interesting but the conclussions do not seem to be so obvious to my critical eye, sorry.

Regarding comparison with Palestinians, I overlooked that. I'm not so sure if people from elsewhere should be an adequate representation. A sample from the Negev or Jerusalem could well have significative differences with respect to Galileans. In the first case it's much more likely to represent a different, more nomadic and "genetically Arabized" stock, in the second case it may represent a much more heterogenous population gathered along the centuries around the holy city. Certainly the best comparison would be native Galileans (Arabs, Jews or whatever), just in case there are regional differences - what considering the quite large mtDNA diversity of West Asia is likely to happen.

Anyhow, the only thing I really get from figure 2 is abstract and unlikely divergence times. You can't seriously believe that Druzes and Palestinians (who we know only began to diverge some 1000 years ago) have been diverging through 6,000 years. It just makes no sense, sorry. That can only be product of the abstractness of the method, the wrong choice of Palestinian outgroup or to if Druzes would be immigrants from elsewhere. It's hard to see how most Galileans and Galilean Druzes, whom apparently only began to diverge less than 1,000 years ago, have been diverging for so long. It goes agains common sense, sorry.

This may be because of the abstractness (inaccuracy) of the model... or can be because the outgroup is not representative of the very neighbours from where Druzes apparently sprang 1000 years ago. Or it may be because Druzes are the outgroup, the immigrants here.

We are fully aware of the local founder effect which is actually even mhore supportive to our hypothesis, of Refugium, we are talking here about Isolates within isolates that preserved part of the regional diversity and decreased the village diversity which is quit differnt from what we see in other populations in the region.

I agree that is different and that Druzes can show some odd lineages but I cannot agree with Druzes being more representative of the past actual genetic diversity than their neighbours. If anything it should be the other way around because drift is much less noticeable (tending to null) whithin large effective populations.

One of the main conclusions of our article which is well supported by Druze oral trditions, is that the Druze as a group has been formed much before the Dawa, otherwise we can not explain by migration (and we have tested various models of migration) the high frequency and high diversity of haplogroup X in the mtDNA and Haplogroup K of the NRY among the Galili Druze.

That's relying quite a bit in oral tradition. Anyhow, assuming it's true, how older could these groups be? Nothing like them is mentioned in history. If they were formed (and is a lot to assume) in the religious chaos of the Late Roman Empire (a gnostic sect?) that would account for less than 2000 years, not more. If they are as old and different, they might even have arrived after the Roman genocide (I presume there was some resettling of the half-empty province) and in that case they would not even be representative of any local population. The more we speculate, the less clear is that Druzes represent anything related with Palestine. Their founder was from Egypt, by the way.

Actually figure 2 can answer many of the questions you have raised regarding Palestinians. As you can see Palestinians have low migration rates with Druze and high migration rates with other Near eatern populations. This can be true or a false positive results due to the different sampling methodologies.

Yes, figure 2 (that I overlooked before) is interesting.

Following that figure the smallest divergence rate of Galilean Druzes is with Greeks and Adygeis (less than 2000 years). If the model is correct, then Druzes are derived from Caucasian Greeks (or something like that). The next divergence figure is with Egyptians (c. 2000 years). So what does that tell us about Druzes? If anything that they do not look Levantine at all.

Of course it may be all just distortion because founder effects and extreme drift, I can't say from my amateur viewpoint.

Thanks for your very kind reply. Believe me that I feel honored with your attention. :)

It's a good paper but, as you said, bias happens. In this case, concluding that Druzes represent the past local population better than the rest seems strongly self-contradicting, really - specially when they are closest to SE Europeans (and then to Egyptians) than to anyone else.

dave in boca said...

My Lebanese uncle was a Druze from Allei above Beirut and we had long talks about the Arab world---I studied Arabic in Beirut & served in the State Dept. His oral tradition, for what it's worth [he didn't practice taqqiya with me as far as I know], was that the Druze didn't consider themselves as "Arabs" though they spoke Arabic. In my study of the Druze [very informal while I lived in the Middle East], I found that some of their doctrine derived from the Fatimid Shi'a dynasty in Egypt, and that incarnation was among the beliefs which differentiated them from mainstream Islam. [Also, the Alawites of Syria hold incarnation, it is said, as a secret teaching.]

For what it's worth, my aunt's grandson just got a full scholarship to Northwestern U. journalism school, one of the best in the USA. The whole family is either very smart or, on the female side, very naive.

Finally, I was lucky enough to visit Yemen many times while living in Jeddah, and even made a land journey to Sana. The Yemenis had a sea-faring tradition, since somewhat degraded, long ago which proselytized Indonesia [Aceh in Sumatra as early as 1000AD] and may also have hit Australia & picked up some Australoids as slaves. Yemenis were slavers from Zanzibar south to Kuwait & the Indian sub-continent. Serendib was the island of Sri Lanka discovered by Yemeni sailors, tradition has it.

FWIW Just some anecdotal stuff from my family & travels. And I have read a couple of Carleton Coon's writings on the region which l found very useful for the Arabian Peninsula.

Unknown said...

That contradicts our results based on the marginal consensus and parental signatures of what is not actual Genetics! Why do Geneticist allow software programs to include Historical references apart from actual Genetics? Especially when there isn't evidence of SNP haplogroup X to suspect any Druze in our results! I guess its a 'no brainer.'