search this blog

Tuesday, December 17, 2013

Near Eastern origin of Ashkenazi Levite R1a


Over at Nature Communications, Rootsi et al. report on a newly discovered Ashkenazi-specific subclade of R1a, defined by the M582 mutation. They argue that it's a marker of Near Eastern origin, and based on the comprehensive data in their paper, I'd say they're correct. However, it's important to note that this doesn't preclude an ultimate Eastern European or Central Asian source of M582 in the Near East. For instance, an R1a mutation ancestral to M582 might have been introduced by the proto-Iranians from the steppe into what is now Iran during the early Indo-European dispersals. Indeed, that's actually what Figure 1a from the study suggests (phylogenetic tree of R1a below). The paper is open access, but here are a few quotes anyway:

Haplogroup R1a-M582 was only sporadically observed in Europe, the Diaspora residence of Ashkenazi Jews. Notably, it was not identified among 2,149 samples (including 922 R1a-M198) of non-Jews from East Europe, where the Ashkenazi Jewish community flourished in recent centuries (Table 1).

...

Within 1,068 West/North European samples (106 R1a-M198), M582 was observed in just one German sample, and among 3,756 Central/South European samples (710 R1a-M198), it was found only in one Hungarian and one Slovakian sample (Table 1).

...

Among 3,739 Near Eastern samples (303 R1a-M198), R1a-M582 was identified in various populations, with the highest frequency occurring within Iranians collected from the southeastern Kerman population who self-identified as Persians, northwestern Iranian Azeri and in Cilician Anatolian Kurds, at 2.86%, 2.50% and 2.83%, respectively (Table 1). In contrast, among 2,164 samples from the Caucasus (211 R1a-M198), R1a-M582 was found in just one Nogay sample (Table 1).

...

Considering the historical records of Ashkenazi Jews, three potential geographic sources should be considered: the Near East, which was the geographic location for the ancient Hebrews; Europe, which was the residence of the Ashkenazi Jewish Diaspora and the region in which they evolved for nearly two millennia; and the region overlapping with the no longer extant mid-11th Century Khazarian Khaganate, whose ruling class has been suggested to have converted to Judaism [18]. Our data render the latter source highly unlikely since the Khazarian Khaganate overlapped with the Northern Pontic-Caspian steppe and the North Caucasus region, in which just one Nogay sample carried the R1a-M582 haplogroup (Table 1). Furthermore, the Nogays, formerly a powerful Kipchak Turkic-speaking nomadic confederation, are relatively recent inhabitants of the Caucasus, and the STR haplotype of the sole R1a-M582 Nogay sample lies outside of the Levite cluster. Had the Caucasus region been the source for the Ashkenazi modal lineage, we likely would have found R1a-M582 Y-chromosomes in some of its 20 local populations examined in our sample of more than 2,000 Y-chromosomes (Table 1).

...

Near Eastern populations are the only populations in which haplogroup R1a-M582 was found at significant frequencies (Table 1). Moreover, the representative samples displayed substantial diversity even within this geographic region (Fig. 1b). Higher frequencies and diversities often suggest lineage autochthony.


Citation...

Rootsi, S. et al. Phylogenetic applications of whole Y-chromosome sequences and the Near Eastern origin of Ashkenazi Levites. Nat. Commun. 4:2928 doi: 10.1038/ncomms3928 (2013).

See also...

The Poltavka outlier

74 comments:

sds said...

David, were they able to identify a time frame for this mutation?

Davidski said...

The lower bound estimate using the two (probably very closely related) Ashkenazi sequences was 1200 years. However, the upper bound estimate using the two Ashkenazi sequences and the Assyrian sequence was 4000 years. I think the latter estimate is more accurate for M582 overall, while the lower bound estimate is probably a reflection of a bottleneck within the Ashkenazi paternal gene pool. See page 5 in the study.

sds said...

So, what about the possibility that this marker originated at the time of the Babylonian exile, which was in the 7th century BCE?

Davidski said...

Yeah, maybe, or just before. But I don't know much about Jewish history, so it's hard for me to put that into any sort of historical context.

Anonymous said...

Davidski: "The lower bound estimate using the two (probably very closely related) Ashkenazi sequences was 1200 years. However, the upper bound estimate using the two Ashkenazi sequences and the Assyrian sequence was 4000 years."

A Z93? between 2000 BCE and 800 AD?

For the Record, not long after the Traces of Indo-aryan words and gods in the Mitanni Kingdom around 1,500 BCE, there is also mention around 1,300 BCE of a governor of Qiltu, a city near Jerusalem, having an indo-aryan name: svardata ("gift from the sun").

This powerful man, that maybe arrived with a larger group, might have had a sizeable number of descendants.
Some of them, maybe centuries later, might have entered the Jewish gene pool.

Davidski said...

I'd say Z93 is a couple thousand years older than M582 and Z2122. If so, that means it was definitely around during the early Indo-European movements from the west to east across the steppe, which makes sense.

sds said...

Hey David,Michał Milewski posted this link at English molgen: https://sites.google.com/site/levitedna/origins-of-r1a1a-ashkenazi-levites/theory-re-nethinim-origins. This might shed some light on my previous comment.

Davidski said...

Yeah, I saw that. It makes sense. M582 looks like a marker of Iranian origin.

About Time said...

I wonder if this Levite R1a could be Mitannian or Proto-Iranian related.

The Torah has a lot about the Levites/Levoyim or really Tribe of Levi. They were a kind of guardian and administrators for the inner sanctum (mishkan) and wore metal breastplates (maybe the original armed bodyguards of Moses, but that's just speculation).

In the later tribal land divisions, Levites didn't get any land and went everywhere as priests and one suspects enforcers among the various tribal communities.

Torah describes the Levites as "fierce," and the Torah places just as much emphasis on Gevurah (precise or exacting judgment) as on Chesed (loving kindness or caritas). One gets the the idea that the original Levites were not people to be trifled with.

I suspect the Levite R1a is an original ancient Jewish lineage (Israelite, technically - this is before the rebellion of the northern tribes). Maybe going back to Proto-Iranian related Mitannians or Hurrians that were not just priests but guardians with military skills.

About Time said...

The Torah account is that the Moses and Aaron were themselves Levites. This would mean the Levites existed as some kind of family group (or office?) by the time of the Late Bronze Age. They kept to themselves as a separate kin group (distinct from the numerous tribes) in the kingdoms of Israel and Judah (before the split, when they were officially disbanded in the northern kingdom but not the southern Judean kingdom).

Clans don't always exclusively share one actual paternal genetic lineage, by the way. Look at clans in Scotland/Ireland for examples of this. The putative ancestral Levites would not have been different in that regard.

The midrashic accounts suggest that many or most Levites left the Jewish fold during the Babylonian captivity. Where they went or whether they stuck together in some cases, who knows.

Would be good to get a handle on the sub-lineages of R1a represented in early Central Asians. It's important to remember that Turko-Mongols carried out a lot of targeted depopulations in Central Asia and yes, Anatolia even in very recent times. We lost a lot of people (and thus their lineages) in this way. Why Tamerlane targeted specifically the city of Ashur for instance, who knows. The area of Khorasan was also heavily depopulated for some reason.

It means we are dealing with a very incomplete record in this part of the world, especially for anything related to Alans, Tokharians, Arsi, or other people who lived around there and probably had some R1a lineages that are now lost.

Onur Dincer said...

Would be good to get a handle on the sub-lineages of R1a represented in early Central Asians. It's important to remember that Turko-Mongols carried out a lot of targeted depopulations in Central Asia and yes, Anatolia even in very recent times. We lost a lot of people (and thus their lineages) in this way. Why Tamerlane targeted specifically the city of Ashur for instance, who knows. The area of Khorasan was also heavily depopulated for some reason.

It means we are dealing with a very incomplete record in this part of the world, especially for anything related to Alans, Tokharians, Arsi, or other people who lived around there and probably had some R1a lineages that are now lost.


I completely disagree with these. According to the results of the paper, the R1a-M582 lineage peaks in frequency (excluding Jews) and has the highest diversity in Iran, and it is completely or almost completely absent in Europe, the Pontic-Caspian region, the Caucasus, Anatolia, the Armenian Highlands, the Levant, Mesopotamia, Central Asia, South Asia, south Siberia, Arabia and Egypt (probably also North Africa as a whole). These suggest that the R1a-M582 lineage has nothing to do with Proto-Iranics and much less with Proto-Indo-Iranians and is instead an autochthonous (pre-Iranic and pre-Elamic) lineage from what is now Iran. The fact that the R1a-M582 lineage is in very low frequencies (except specific Jewish groups, whose levels are surely due to founder effect and drift) even in the populations it peaks in explains its complete or almost complete absence in almost all of West, Central, North and South Eurasia.

BTW, you mention Turko-Mongols as depopulators of the regions they settle, but according to the results of the paper one of the non-Jewish populations with the highest frequency of the R1a-M582 lineage is Azeris of Iran, which is a Turkic-speaking population. You should draw conclusions based on data rather than your personal opinions.

About Time said...

These suggest that the R1a-M582 lineage has nothing to do with Proto-Iranics and much less with Proto-Indo-Iranians and is instead an autochthonous (pre-Iranic and pre-Elamic) lineage from what is now Iran.

Interesting. Jiroft civilization?

But what is the connection between Iranians from Kerman, Cilician Kurds, and Iranian Azeris, if not Proto-Iranian? Medes maybe (if Medes were different from Proto-Iranians)?

Onur Dincer said...

Interesting. Jiroft civilization?

To make a safe guess about the identity of the earliest carriers of the R1a-M582 lineage we should first have a safe dating of that lineage, which I doubt we can have with the current dating methods.

But what is the connection between Iranians from Kerman, Cilician Kurds, and Iranian Azeris, if not Proto-Iranian? Medes maybe (if Medes were different from Proto-Iranians)?

Kurds of Cilicia are new comers to Cilicia, they arrived there during the last couple of generations from the Kurdish-majority eastern areas of Turkey and have not mixed much with the local population of Cilicia, which is composed of Turks, Arabs and an extremely small Armenian minority. So Kurds of Cilicia can only represent the genetics of the Kurdish-majority areas in Turkey (which are in the easternmost regions of the country) and not of Cilicia.

The connection between the relatively-high-R1a-M582-lineage-carrying non-Jewish populations is their common ancient connection to the territories corresponding largely to what is now Iran. If the R1a-M582 lineage was Proto-Iranic-related, we would probably see that lineage in Central Asia (including what is now Afghanistan), south Siberia and the Pontic Caspian region as well, but we do not see it in any of those regions, so it is very unlikely to be connected to Proto-Iranics.

Davidski said...

It's derived from Proto-Iranics. Look at the phylo tree:

Z93 > Z94 > Z2122 > M582

Onur Dincer said...

It's derived from Proto-Iranics. Look at the phylo tree:

Z93 > Z94 > Z2122 > M582


The Z93 lineage probably originated in an area streching from what is now Iran to the southern parts of Central Asia (including what is now Afghanistan) well before the Proto-Indo-Iranian times.

Davidski said...

Z93 shares the Z645 mutation with the European Z282 (which includes the Scandinavian Z284, Central European M458 and Balto-Slavic Z280). But the Northwest European L664 doesn't share this mutation.

Also, Z93* is found in Europe. So Z93 probably came from Europe, or the Volga-Ural region on the border between Europe and Central Asia.

Its first main expansion was probably with the proto-Indo-Iranian Sintashta chariot complex during the Bronze Age.

Krefter said...

Onur, I agree with Davidski there is ancient DNA evidence for what he is saying. Y DNA from suspected Indo Iranians and Tocharian's(tarim mummies) in Asia during the bronze and iron age 16 out of 17 had R1a1. They also had pale skin and mainly light hair and eyes pretty much prove they were European. They had a high amount(around 35%) of U5(all under were U5a), U4, and U2(all under were U2e) similar to Mesolithic hunter gatherers from Karelia, Russia and Siberia(who had mainly east Asian mtDNA though).

Their cultures are suppose to descend from copper age Yamna in far eastern Europe. Just like Corded ware which is probably responsibly for spreading R1a1a1b1 Z283 the brotherclade of R1a1a1b2 Z93. There are also wto R1a1's from Corded ware culture.

That is pretty good evidence copper age Yamna culture is the source of most modern R1a. Also Indo Iranian, Tocharian, and Balto Slavic languages.

Onur Dincer said...

Onur, I agree with Davidski there is ancient DNA evidence for what he is saying. Y DNA from suspected Indo Iranians and Tocharian's(tarim mummies) in Asia during the bronze and iron age 16 out of 17 had R1a1. They also had pale skin and mainly light hair and eyes pretty much prove they were European. They had a high amount(around 35%) of U5(all under were U5a), U4, and U2(all under were U2e) similar to Mesolithic hunter gatherers from Karelia, Russia and Siberia(who had mainly east Asian mtDNA though).

Their cultures are suppose to descend from copper age Yamna in far eastern Europe. Just like Corded ware which is probably responsibly for spreading R1a1a1b1 Z283 the brotherclade of R1a1a1b2 Z93. There are also wto R1a1's from Corded ware culture.

That is pretty good evidence copper age Yamna culture is the source of most modern R1a. Also Indo Iranian, Tocharian, and Balto Slavic languages.


We do not know to which clades of R1a1 or R1a1a the ancient sequences belong. So there is room for a lot of speculation. David may be right about the place of origin of Z93, but his datings are less credible. Anyway, accoring to the results of the above paper, M582 does not show any sign of having originated among Proto-Indo-Iranians or Proto-Iranics and instead it seems to have originated in pre-Iranic (probably also pre-Elamite) Iran.

Krefter said...

"We do not know to which clades of R1a1 or R1a1a the ancient sequences belong. So there is room for a lot of speculation. David may be right about the place of origin of Z93, but his datings are less credible. Anyway, accoring to the results of the above paper, M582 does not show any sign of having originated among Proto-Indo-Iranians or Proto-Iranics and instead it seems to have originated in pre-Iranic (probably also pre-Elamite) Iran."

The paper makes it pretty clear that M582 is most popular in Iranian speaking people besides Jews. M582 is descended from Z93 so if you go back 5,000 years its lineage was probably in the European part of Russia and is of Iranian origin. Even if these studies age estimates are to old for Z93 or one of its subclades to have an Iranian origin. Ancient DNA has great evidence it is. I agree with something Davidski said before that age estimates can be very inaccurate and what we really need ancient DNA.

Onur Dincer said...

Barak,

What we really need is ancient DNA, but those ancient DNA studies usually do not have enough resolution to provide detailed sub-lineage information, which renders them significantly less useful than the needed levels of usefulness.

M582 is relatively popular in Iran, but not in any other presently and/or historically Iranic-speaking or Indo-Iranian-speaking land, which indicates that it is an Iran-specific rather than an Iranic- or Indo-Iranian-specific lineage and originated in what is now Iran.

The place of origin of Z93 is less clear, so I do not have an as strong opinion for its place of origin as I have for M582.

Davidski said...

We have ancient DNA. This ancient DNA shows that European nomads migrated east across the steppe during the Bronze Age.

Their Y-DNA STRs indicate that they carried either Z93 or the closely related Z82. But Z282 is today missing from Iran and India.

The reason it's difficult to tell apart Z93 from Z282 with STRs is because they split only recently, and then expanded very rapidly.

Also, if you'd like to see Z93 origins deep in Asia, then you have to explain why Z93 is present in Europe, but the Central Asian L657, which is very common in many of the Central Asians stans, and just under Z93, totally missing from Europe.

Onur Dincer said...

David, what is your opinion regarding this comment from Palisto?

"I try to focus on individuals with STR111 data, only very few West, Central and South Asians have done that. Due to the high number of Europeans and low number of Asians in this data set the tree is shifting from Z93 towards Z283. This can be seen in the age estimates based on the tree (Z283: 5973 years; Z93: 5792 years) vs age estimates based on the actual observed STR variability (Z283: 6752 years; Z93: 6980 years).

Although the Asian side of R1a is heavily underrepresented in public data sets we already know the following:
From the SNP distribution of the "FTDNA R1a1a and subclades project" we know that the oldest branches of R1a were also found in West Asia.
http://i1127.photobucket.com/albums/l625/ft-d/1O-R1a.jpg
We also know that one of the oldest branches of the "European" Z283 branch were found in West Asia (including one individual with paternal Kurdish ancestry).
http://i1127.photobucket.com/albums/l625/ft-d/maps/z2832.jpg

I believe that rigorous SNP testing of West Asians will lead to many surprises in the R1a tree. Unfortunately, right now most Asian R1a tested individuals are from the Arabian peninsula that doesn't show much diversity."

http://kurdishdna.blogspot.com/2013/05/r1a-tree.html?showComment=1369332846297#c8820065339593069922

Onur Dincer said...

Also regarding this from Palisto?

"Just to clarify a few things. The R1a* haplogroup mentioned above does not belong to the "Asian" Z93 nor the "European" Z283. It is ancestral to both. So far, more detailed SNP information are only known for three R1a individuals with paternal Kurdish ancestry: one is R1a1a1b1 Z283+, one is R1a1a1b2 Z93+, and the latest one is R1a* based on ISOGG 2013 nomenclature. For all the remaining R1a Kurds it is only known that they are at least M17+ (R1a1a); no SNPs downstream of M17 were determined, so they could be Z283 or Z93 or ...
Given the current Kurdish data, nobody can determine, which subbranch of R1a is dominant in Kurds."

http://kurdishdna.blogspot.com/2013/10/r1a-discovered-among-yezidi-kurd.html

Davidski said...

He's hoping for a miracle. But the current data back up what I said above, and not what Palisto is hoping for.

Onur Dincer said...

What exactly is the "miracle" he is hoping for?

Davidski said...

He wants to see the split between the two main lineages of R1a, Z645 and CTS4385 (which I actually somewhat erroneously referred to above as L664), to have taken place in Asia rather than Europe. But CTS4385 is firmly placed in Northwestern Europe, with all of its substructure there, and not a single case of it in Asia. So what Palisto needs now to cast doubt on the European split between Z645 and CTS4385 is a parallel clade in Asia, which also isn't showing up anywhere.

There's just too much structure and age in European R1a to still hope that more samples from West Asia will change anything. Also, the project that he references to back up his argument that the oldest R1a is found in West Asia actually shows the oldest, most basal R1a in Poland, Belarus and Russia. Right at the top of the table here labeled "Old R1a"...

http://www.familytreedna.com/public/R1a,R1a/default.aspx?section=yresults

So yeah, anything's possible I suppose. But I doubt that a complete reshuffle of the R1a phylo tree is on the cards. If anything, we'll probably see some surprises from the R1a in Russia, especially from the ancient samples there.

Onur Dincer said...

David, where and when exactly do you think the M582 lineage originated? Also, what do you think the R1a lineages of Proto-Indo-Iranians, Proto-Iranics, Proto-Indo-Aryans, Cimmerians, Scythians, Sarmatians, Sakas, Alans, Bactrians, Sogdians, Medes, ancient Persians, Parthians and Tocharians were?

Davidski said...

M582 originated in Iran, while the main paternal marker of all the Asian Indo-European groups south of the Eurasian steppe was always Z93. The Andronovo, Scythian and Sarmatian nomads of the Eurasian steppe probably carried both Z93 and Z282.

But the Tarim Basin mummies might have been the exception among the non-steppe Asian groups and carried some sort of M417 (Z93x), because they left Europe very early.

Krefter said...

"If anything, we'll probably see some surprises from the R1a in Russia, especially from the ancient samples there."

Have those people who got the 6 Yamna Kurgen Y DNA samples from Russia told you the results? Do you have any idea when Yamna-Catacomb DNA will come out? It has been 5 weeks since they said in a few weeks. What about the possibly R1a M417* from Palaeolithic west Eurasian like Siberian? I bet the stuff for La Brana-1 wont be out for a few months even though that Spanish article said 2 weeks.

Davidski said...

The Eurasian steppe paper will be out in the next issue of this journal in 2014...

http://archiv.ub.uni-heidelberg.de/ojs/index.php/arch-inf

But I suppose it MIGHT be out earlier than that here...

http://www.dguf.de/index.php?id=9

And in that Spanish article on La Brana 1 it says that the study will be out in a few weeks...but yeah, that probably means a couple of months.

Onur Dincer said...

M582 originated in Iran

At least we agree on that.

while the main paternal marker of all the Asian Indo-European groups south of the Eurasian steppe was always Z93. The Andronovo, Scythian and Sarmatian nomads of the Eurasian steppe probably carried both Z93 and Z282.

Sounds plausible, there are many alternative theories though. We'll see which one is true when more people from the Asian continent are tested for the downstream mutations and also when ancient DNA samples are tested for the downstream mutations.

But the Tarim Basin mummies might have been the exception among the non-steppe Asian groups and carried some sort of M417 (Z93x), because they left Europe very early.

Yeah, plausible.

Onur Dincer said...

BTW, David, you answered only half of my question regarding M582. You said where you think M582 originated but did not say when. I am expressing this especially because that I think our disagreements are mainly in the datings rather than locations.

Davidski said...

It had to have been after the Sintashta Culture (2100–1800 BCE). But well before the arrival of Jews in Ashkenaz (Germany), which happened around the 4th century. So let's say the M582 mutation appeared in the Near East around 3000 years ago, but certainly not much more before that.

Onur Dincer said...

So let's say the M582 mutation appeared in the Near East around 3000 years ago, but certainly not much more before that.

So, according to you, M582 arose among the Indo-Iranian-speaking colonizers of what is now Iran (3000 years ago they were new arrivals to what is now Iran from Central Asia and so were pretty much colonizers there).

Davidski said...

Yes, Z2122 and then M582 arose in the Indo-Iranians from the steppe who colonized what is now Iran and surrounds.

Onur Dincer said...

All the evidence indicates that the M582 lineage originated in what is now Iran or somewhere very close-by. What is open to dispute is when it originated. If it originated before the Indo-Iranian migrations to what is now Iran, then it probably originated in a non-Indo-European-speaking community of what is now Iran. If it originated after the Indo-Iranian migrations to what is now Iran, then chances are high that it originated among Indo-Iranian speakers of what is now Iran. In both scenarios the place of origin is what is now Iran or somewhere very close-by. I favor the first (non-IE) scenario since Indo-Iranian languages are relatively new comers to what is now Iran.

About Time said...

Question. If R1a Z93 (including M582, which is downstream) is not the Iranian lineage in Kurds, Iranians, and Azeris, then what is?

Davidski said...

Z645 shows a close correlation with Baltic, Slavic, North Germanic, Iranian and Indo-Aryan languages. It's not native to Iran.

Onur Dincer said...

Indo-European migrations did not occur in a vacuum, there were a lot of non-IE speakers to mix with in almost every region they migrated to. Z93 might have been a significant part of the genetic pool of Proto-Indo-Iranians, but it was probably already widespread among many non-IE groups who lived then, and some of those groups might have eventually been Indo-Iranized.

Languages and genes spread via very different processes, so you should not expect strong correlations between their spreads.

Davidski said...

Onur, all main subclades under Z93 show high correlations with Indo-Iranians.

Non-Indo-Europeans, like Turks and Jews, carry nothing except obvious Indo-Iranian lineages of Z93, but with much less diversity caused by severe bottlenecks.

Onur Dincer said...

The distribution of Z93 and its sub-lineages, as in the other R1a lineages, are largely region-specific rather than linguistic affiliation-specific. The European Jewish case is special since European Jews mainly descend from West Asian paternal lines and European maternal lines and they are a highly drifted population due to the ancient and medieval bottlenecks. I do not know what you mean by "Turks" since that word is ambiguous in meaning.

Davidski said...

Onur, you can't push shit up a hill. Here's the situation.

Z645 - Z2122 > Iranians

Z645 - Z2123 > Iranians and Indo-Aryans

Z645 - L657 > Eastern Iranians and Indo-Aryans

Z645 - M458 > Slavs and Germans

Z645 - Z280 > Slavs, Balts and Germans

Z645 - Z284 > Scandinavians

Now kindly please give some examples of non-Indo-European branches of Z645. I'm not aware of any, but obviously you are. So what are they?

Onur Dincer said...

David,

The Indo-European languages spread so far and wide in West and Central Eurasia during pre-historical and historical times that the linguistic legacy of non-IE speakers in those regions were largely erased. But their genetic legacy was much harder to erase so was much less affected by the spread of the IE languages.

The distributions of the R1a lineages you presented in your last comment are all region-specific rather than language group-specific, which supports my arguments.

Davidski said...

Why is it then that the regions they cover are all Indo-European regions? Why aren't these or any related markers common in regions that Indo-Europeans never colonized?

Are you suggesting that these markers were somehow attracted to regions and ethnic groups that spoke Indo-European languages?

Onur Dincer said...

Look at this world frequency map of R1a1a in indegenous populations.

http://en.wikipedia.org/wiki/File:R1a1a_distribution.png

It shows up in notable frequncies in some non-IE-speaking populations with no likelihood of IE origins such as northern Uralic groups, non-Brahmin Dravidian speakers and large parts of the Caucasus.

Davidski said...

All of these groups had significant contacts with Indo-Europeans at some stage. They were usually conquered by them, like the Dravidians by the Indo-Aryans and the Siberians by Russians, or they shared early contact zones with them for centuries, like the Volga-Finns.

None of these groups shows any unique subclades of R1a. They always carry bottlenecked versions of subclades typical of the Indo-Europeans they came into contact with.

Onur Dincer said...

What IE origin do the Chenchu tribals (Dravidian speakers from southern India) have? These people were exclusively hunter-gatherers until recently, yet they carry 26% of R1a1.

Davidski said...

All Dravidians belonging to R1a carry lineages under Z93.

http://www.sciencedirect.com/science/article/pii/S0378111913005714

You should really read this as well.

http://eurogenes.blogspot.com.au/2013/05/south-asian-r1a-in-1000-genomes-project_15.html

Onur Dincer said...

All Dravidians belonging to R1a carry lineages under Z93.

Apparently so do all Indo-Aryans belonging to R1a, which is again supportive of my arguments.

About Time said...

Would have been good to test other Middle Eastern R1a for M582. There is at least one person from Al Shaybi al Qurayshi who is R1a (old family in charge of Abrahamic holy site). If Quraysh don't count as legitimate Arabs, I don't know who would.

It's true that Iranian shows up pretty late in the Middle East. Unquestionable dates for Medes are later than I posted earlier, more like 600's BC.

R1a Z93 might not have originated exclusively with Indo-Iranians and does show up elsewhere. But M582 is much more specific and Proto-Iranian would not be a bad tentative label. As you say, Azeris are considered basically Caucasian Albanians and Iranians who now speak a form of Turkish.

If Levite R1a is not Iranian, it came in with some ethnos that was extremely closely related to Iranians. Gutians, Hurrians, Mannaeans. Not a lot if other options.

Onur Dincer said...

But M582 is much more specific and Proto-Iranian would not be a bad tentative label.

It is a pretty bad label since Proto-Iranic descendants are in no way restricted to Iran and environs and Jews and Jewish descendants while M582 seem to be so.

About Time said...

Fair enough. How about tentative "Levite-Iranian" for M582?

Onur Dincer said...

How about tentative "Levite-Iranian" for M582?

If you use the "Iranian" part in territorial sense rather than ethnic or cultural sense, why not?

Davidski said...

Onur, if you knew what you were talking about, you'd know that you don't have any arguments.

Dravidians carry Z93 because they were conquered by Indo-Iranians.

Onur Dincer said...

David, even hunter-gatherer Dravidians carry Z93. So apparently not all Z93 in Dravidians come from Indo-Iranians.

Davidski said...

They don't just carry Z93. They carry subclades of Z93 way down the R1a phylo tree which are most commonly found in Iranians and Indo-Aryans.

The Z93 lineages that you want to be in India (ie. those not characteristic of Indo-Iranians) are actually found in Europe.

So you still have no argument, just your own opinion not based on any facts.

Onur Dincer said...

They don't just carry Z93. They carry subclades of Z93 way down the R1a phylo tree which are most commonly found in Iranians and Indo-Aryans.

The Z93 lineages that you want to be in India (ie. those not characteristic of Indo-Iranians) are actually found in Europe.

So you still have no argument, just your own opinion not based on any facts.


I am not a supporter of South Asian origin of Z93. I am just questioning your claim that it first arrived in South Asia with Indo-Iranians, and my arguments are based on facts.

Davidski said...

What facts are these?

Onur Dincer said...

What facts are these?

I have already stated them multiple times in this thread. I hate repeating myself over and over, so won't repeat again.

About Time said...

Looking more closely at the Near East pooled** big data set from the paper, M582 really looks specific to what later were Iranian speakers.

That said, Levites have highest frequencies so are primary bearers of M582. It bears repeating, but Israel/Judah predate all Arabs and Iranians in documented sources.

Some Samarians/Northern kingdom populations were deported to the cities of the Medes according to Assyrian inscriptions

This means M583 could well be first Levite Jewish and then only later became Iranian. Either some Levites mixed into later Iranian populations (Medo-Persia arose after Israel/Judah were well in decline), or Levites and Iranians share partial descent from the same population (Mitannians? or Gutians or Mannaeans, as Onur suggests).

Davidski said...

The Levite M582 is the result of a severe founder effect and a subset of Iranian M582. Look at the STR tree in the paper.

Moreover, M582 is downstream of Z93 and Z94, which are both Indo-Iranian markers, and downstream of Z645, which is an Indo-European marker.

So no, M582 could not have traveled in time and parachuted into a Near East without any Indo-Iranians.

Onur's arguments are based on personal incredulity, and you're heading that way too.

About Time said...

Could be that M582 started out somewhere in N Mideast (Mitanni is my guess). Most stayed and ended up in Iranian speakers right at the source (Azeris, Kurds, Karman). Some became part of the Proto-Jewish or Proto-Levite population further south. Could be Surdata of Qiltu type local elites, who would make sense as a cultures elite in an otherwise humble Proto-Judeans context.

Other possibility is that somehow actual Iranians were absorbed specifically by Levites at some later time. This is a bit of a stretch though, because Judeans and Levites especially were highly highly endogamous on religious grounds (Pinchas) and even made special efforts to ensure this later on (reforms of Ezra upon return from return from Babylonian exile).

At least on the Jewish side of things, we have written sources that can also be checked against Assyrian sources for some of these events, so it's not just pure guesswork.

Davidski said...

Introgression can happen in many ways, but might be covered up, especially within a group that is supposed to be highly endogamous on religious grounds.

Fine-scale SNP phylogeography and STR variance are more informative when looking for the origins of Y-DNA markers, and these clearly show that Levite and indeed Jewish M582 is a subset of Iranian M582, which is in turn a subset of Indo-Iranian and Indo-European R1a.

So at some point in time, probably about 2000 years ago, M582 entered the Jewish gene pool due to limited introgression from nearby Iranians. About 1000 years later the Levite M582 experienced a severe founder effect.

About Time said...

Why 2000 years ago and not at Levite ethnogenesis c 3000 years ago?

What caused the founder effect?

Could this have something to do with Alanians or similar instead of Proto Iranians?

Also wondering about the midrashic notion that most Levites were lost during Babylonian exile. Hard to imagine people separating then returning, but again these were an endogamous and mobile priestly group with a highly defined ethnic identity within Judaism.

About Time said...

Sorry to double post, but just to hash out what I'm saying. How do data contradict:

1. M582 origins in Proto Iranians or someone extremely close to that.

2. Subset of M582 is absorbed in first Proto Levites (Mitanni or similar, Surdata of Qiltu type elites near Judah). 3500-3000 years ago.

3. Bottleneck of Levite M582. Could have happened in revolt of northern kingdom (late 900s BC) when Levites were outlawed as a group in the larger Samarian/Omride northern kingdom. Some stayed in Judah (including a few immigrants from northern Levites).

You mention later dates, but are these dates out if the ballpark?

Davidski said...

Based on the estimates in the paper, M582 is 1000-4000 years old. That's not exactly a precise timeframe, but it kind of agrees with the most logical assumption, which is that the M582 mutation appeared in what is now Iran some time after the early Iranians got there.

I suppose the introgression of M582 into the Jewish gene pool might have happened right after M482 appeared, but I doubt it. So I think my estimate of 2000 years ago, more or less, is reasonable. And there was definitely a founder effect within the M582 Levite pedigree, because that's what the STR data show, and probably why the age estimate of M582 based on the two Ashkenazi M582 lineages is only around 1000 years.

About Time said...

David, my problem with this idea of "hidden mixture"' is that ancient Judea and esp the priesthood was one of the most highly literate and endogamous societies in the ancient world. When they mixed, it was obvious to all and openly criticized. Eg, when the northern kingdom mixed with Cimmerians the street prophets openly lampooned them and said it would lead to a downfall.

Read wiki about the nethinim and ancient attitudes towards mamzer status. Same with some of the Judean kings who mixed with Edomited (like Herod's family). This was openly criticized and well known to all,

The notion that the priesthood (which was highly socially visible as the elite of a basically religious society) could just sneak in a bunch of Persian lineages at a late date stretches credibility in that cultural setting. Even in the Middle Ages, someone would have written or said something (if only to challenge Leviticus authority).

This theory ignores all the basic factors (endogamy and visibility of the powerful hereditary priesthood) of how Judah/Israel actually functioned.

Davidski said...

It had to happen somehow, like maybe an non-paternity event after an Iranian raid? I don't know, but the M582 is there, and it's not a marker native to the Levant. The upstream Z2122, Z94 and Z93 also aren't native to the Levant, or even the Near East.

The Near East in general has a thin sprinkling of a wide variety of R1a types, usually under Z93, without any clear affinities between these subclades and the native populations there. So all of them look like admixtures from the wide variety of Indo-Iranian groups just to the east, who show very clear patterns in terms of the R1a that they carry.

About Time said...

"Native to the Levant" can be a tricky concept. Arguably, Mitanni and Hittites were not native to the Levant, but they were around 3500 years ago.

All the old Jewish sources are pretty explicit: don't marry Canaanites. "Go to your uncle Laban to marry one of his daughters instead."

The Levite M582 looks like a derived endogamous lineage, exactly what you'd expect for actual ancient Levites. Again, they did have a known bottleneck when the large Omride kingdom broke away and banned the Levites. Some Levites remained in the much more isolated and poor southern kingdom Judah.

For M582 to come in later, it would be a specifically Iranian source, not a steppe lineage as far as I can see. That means some possibilities like Alans or Central Asian Iranians who fled Khorasan in the Turko-Mongol conquests look unlikely (wrong sub lineages of R1a).

Onur Dincer said...

This means M583 could well be first Levite Jewish and then only later became Iranian. Either some Levites mixed into later Iranian populations (Medo-Persia arose after Israel/Judah were well in decline), or Levites and Iranians share partial descent from the same population (Mitannians? or Gutians or Mannaeans, as Onur suggests).

About Time, I did not make a specific ethnic group suggestion for the origin of R1a-M582.

Onur's arguments are based on personal incredulity, and you're heading that way too.

David, I have enough reasons to doubt your hypothesis of Indo-European origin of R1a-Z645 and all of its subclades. Your hypothesis is so radical that it needs much more evidence than you provide. But of course, only more detailed and extensive study of ancient and modern DNAs can settle the dispute. I do not have any agenda against your hypothesis. I would happily accept your hypothesis if there was enough evidence for it, as its simplicity makes it look, at least to me, beautiful and I like Indo-European languages in general.

Davidski said...

This hypothesis isn't radical at all. It's been around for a while and is backed up by ancient DNA from Kurgans.

And now here's more evidence that Y-DNA R1a, or indeed R1, is an Indo-European marker.

http://eurogenes.blogspot.com.au/2013/12/ancient-human-genomes-suggest-three.html

Amit said...

Guys, in all your discussions and debates about Indo European or Indo Iranians, you are forgetting the real source and origin of all these, the Indians or South Asians from where it all originated!

Palisto said...

Davidski 2013: "He wants to see the split between the two main lineages of R1a, Z645 and CTS4385 (which I actually somewhat erroneously referred to above as L664), to have taken place in Asia rather than Europe. But CTS4385 is firmly placed in Northwestern Europe, with all of its substructure there, and not a single case of it in Asia. So what Palisto needs now to cast doubt on the European split between Z645 and CTS4385 is a parallel clade in Asia, which also isn't showing up anywhere."

In 2013 I responded elsewhere:
"L664 is not a very old branch, which can be seen by the low STR diversity; CTS4385+L664- is only known for a single individual from UK, so it is a bit premature to make big conclusions based on this one individual."
It is 2015 now and still there is only a single individual (259861) from UK with CTS4385+L664-. I would not be surprised if his results have some errors.

Davidski said...

So what, now we have R1a from Mesolithic Europe, and I can tell you now that Anatolia and surrounds was nothing like it is today during the Mesolithic and Neolithic.