April 04, 2011

New paper on Y chromosome haplogroup T

There are some haplogroups (like R1b) that garner lots of attention and public interest; then, there are others, like haplogroup T that most people don't even know about. This haplogroup's fame was largely due to the fact that Thomas Jefferson belonged to it, but, since at that time T was called K2, most people have probably forgotten about that too.

A new paper studies this mysterious haplogroup and uncovers some structure in it. Moreover, interestingly, it joins T with the somewhat less mysterious haplogroup L that is quite common from the Caucasus through Iran to South Asia. The internal structure of the Y-chromosome phylogeny is getting resolved.

The Lemba evidence is also interesting: this population was theorized to have a Jewish origin on the basis of its possessing the 6-marker Cohen Modal Haplotype within haplogroup J. When it was later recognized that this was insufficient as a Jewish marker because (a) it could be found on very different lineages (J1 and J2, even non-J), and (b) it could be found at how frequencies in non-Jewish Near Eastern populations, the case for the Jewish origin of the Lemba largely crumbled.

This is probably another blow to that theory, as T1b* does link the Lemba to the Near East but not particularly to Jews. Alternatively, modern Jews are a very restricted subset of ancient Jews, a possibility that probably needs to be entertained although it would be difficult to test.

Hum Biol. 2011 Feb;83(1):39-53.

Increased Resolution of Y Chromosome Haplogroup T Defines Relationships among Populations of the Near East, Europe, and Africa.

Mendez FL, Karafet TM, Krahn T, Ostrer H, Soodyall H, Hammer MF.

Abstract

Increasing phylogenetic resolution of the Y chromosome haplogroup tree has led to finer temporal and spatial resolution for studies of human migration. Haplogroup T, initially known as K2 and defined by mutation M70, is found at variable frequencies across West Asia, Africa, and Europe. While several SNPs were recently discovered that extended the length of the branch leading to haplogroup T, only two SNPs are known to mark internal branches of haplogroup T. This low level of phylogenetic resolution has hindered studies of the origin and dispersal of this interesting haplogroup, which is found in Near Eastern non-Jewish populations, Jewish populations from several communities, and in the patrilineage of President Thomas Jefferson. Here we map 10 new SNPs that, together with the previously known SNPs, mark 11 lineages and two large subclades (T1a and T1b) of haplogroup T. We also report a new SNP that links haplogroups T and L within the major framework of Y chromosome evolution. Estimates of the timing of the branching events within haplogroup T, along with a comprehensive geographic survey of the major T subclades, suggest that this haplogroup began to diversify in the Near East ∼25 kya. Our survey also points to a complex history of dispersal of this rare and informative haplogroup within the Near East and from the Near East to Europe and sub-Saharan Africa. The presence of T1a2 chromosomes in Near Eastern Jewish and non-Jewish populations may reflect early exiles between the ancient lands of Israel and Babylon. The presence of different subclades of T chromosomes in Europe may be explained by both the spread of Neolithic farmers and the later dispersal of Jews from the Near East. Finally, the moderately high frequency (∼18%) of T1b* chromosomes in the Lemba of southern Africa supports the hypothesis of a Near Eastern, but not necessarily a Jewish, origin for their paternal line.

Link

18 comments:

Andrew Oh-Willeke said...

Hurray! Y-DNA haplogroup T's distribution is the greatest mystery of any of the haplogroups yet observed (with Y-DNA haplogroup D a runner up mostly because it appears so old that there is more freedom to devise plausible stories to explain it) and is worthy of more data.

Eze said...

There could be a possible connection between the Lemba and the Somali. As Somalis also carry relatively high frequencies (10% to 21%, in various studies) of T, not sure if their main sub-clade is T1b though.

eurologist said...

Very interesting. Looks like much European T is indeed very old and likely neolithic.

However, as often, I find some of the derived time estimates a little too young. ~47-48 kya for K as whole makes no sense with its world-wide distribution and diversification. I think the absolute minimum should be 50-60 kya, based on the fact alone that each NO and P are probably at least 40 - 45 ky old given their wide distribution and locations, and also considering the presence of both C (often assumed to be at least 60,000 years old) and K in Australia.

~25-30kya for T seems likewise a bit young. Between ~15-25kya there would have been only a small population possible in the Near East - so it makes more sense that some of the diversification of T happened before then (e.g., 25-35kya for mutation PS21 rather than during the coldest and driest time of the ice age!). And then we can ask what would have made LT back-migrate to the west from South Asia? A better time window is 40-45kya when the climate improved and West Asia as a whole was much wetter for a while. Soon after that, the near East became very dry and cold, again - certainly neither an incentive to move there from South Asia, nor conditions for sufficiently large populations for diversification.

AdygheChabadi said...

Ahhhhhhhhhh! I love it!

Now, we need one of Y-Haplogroup L!

Andrew Oh-Willeke said...

Before they do Y-DNA hg L, they need to look at Y-DNA hg T in India. It seems unlikely to have originated there given its highly localized distribution around the proto-Dravidian language area, but where did it come from? The Horn of Africa? Egypt? The Levant? Europe? Details sublineage analysis should tell us provide a separate line of evidence of the age of hg T age in India.

FWIW, Eurologist, I think you have the direction and timing wrong. I think L which has a very distinct Indus River Valley distribution probably arrives ca. 7000 BCE with the Harappans (possibly as LT that becomes distinct there), and that T arrives in South Asia from somewhere ca. 3000 BCE-2500 BCE with African crops and expands with the proto-Dravidian langauge in South Asia. The European expansion looks like mostly 6000 BCE to 3000 BCE as part of the same wave that brings non-Iberian Y-DNA hg E to Europe, probably older rather than more recent.

African hg T looks like a 7000 BCE or more recent arrival (perhaps differentiated earlier), together with Afro-Asiatic languages and the Egyptian wing of the Neolithic.

LT probably arises in the Fertile Crescent with L on the Eastern end and T on the Western end, differentiating in the Epipaleolithic.

I'd be curious to know what Dienekes has to say on the dating in this paper. Also, does anyone have more details on what the paper finds (paywalls still suck).

South Central Haplo said...

"We also report a new SNP that links haplogroups T and L within the major framework of Y chromosome evolution"

This is really surprising. It's a big statement. More studies need to be done.

Dienekes said...

I'd be curious to know what Dienekes has to say on the dating in this paper. Also, does anyone have more details on what the paper finds (paywalls still suck).

It uses the evolutionary rate, so divide by 3. That makes the age of T Neolithic.

"We also report a new SNP that links haplogroups T and L within the major framework of Y chromosome evolution"

This is really surprising. It's a big statement. More studies need to be done.


What new studies need to be done? It's a SNP that unites T and L, not much ambiguity there.

xx said...

So should we who are T's call ourselves LT's now? It says the two are now definitely linked by an SNP. I also notice similarities with T in some of the places L occasionally appears in Europe (though L is apparently at even lower frequencies in these places than T).

eurologist said...

"FWIW, Eurologist, I think you have the direction and timing wrong."

Andrew, I thought the highest diversity of L in Pakistan was well established - but I am certainly far from an expert on this. And with K so widespread and diverse in Southeast Asia and Oceania, it is hard to argue for any other origin for K than South or Southeast Asia. So, either LT or L and T must have back-migrated to West Asia at some point- no? To me, the most likely scenario is that LT spread in the entire area between the Levant and Pakistan, and subsequently the westernmost portion became T, the easternmost portion L. And during periods of severe droughts there is a convenient geographic border between them.

"It uses the evolutionary rate, so divide by 3. That makes the age of T Neolithic."

Dienekes, the thing is, though, the entire system must somehow work out as a package. Do you then also make K 16k years old, and C 20k years? Then you effectively postulate that a group of people entered Australia at the end of the last ice age and wiped out every male there of the people who had lived there for the prior 30,000 years. Is that really what you want to claim?

argiedude said...

I don't get it. In ISOGG all the SNPs tested here are already present. By the way, did the study include haplotypes? Though it doesn't matter too much, because most of these haplogroups are already well fleshed out in the y-dna T Project website.

PS: They should've tested the T cluster restricted to Ibiza (Baleares Islands) that makes up 10% of their y-dna.

AdygheChabadi said...

Andrew Oh-Willeke, if you want the paper I can send it to you.

Andrew Oh-Willeke said...

@ AdjgheChabadi

I would be indebted to you if you could. My personal e-mail is a hotmail.com domain and the first part is my last name, all lower case without punctuation.

"Andrew, I thought the highest diversity of L in Pakistan was well established - but I am certainly far from an expert on this. And with K so widespread and diverse in Southeast Asia and Oceania, it is hard to argue for any other origin for K than South or Southeast Asia. So, either LT or L and T must have back-migrated to West Asia at some point- no? To me, the most likely scenario is that LT spread in the entire area between the Levant and Pakistan, and subsequently the westernmost portion became T, the easternmost portion L. And during periods of severe droughts there is a convenient geographic border between them."

I would see some predecessor to both L and T and LT migrating from South Asia or points East to the Fertile Cresent, followed by LT or near total migration very basal L back migrating from the Fertile Cresent to the Indus River Valley/Pakistan where it subsequently experiences population expansion and becomes diverse - L is most diverse in Pakistan because it would have come into being there after LT arose in the Fertile Cresent from a predecessor to LT from I don't know where. LT wouldn't have backmigrated because it would have arisen in the Fertile Cresent from something else.

Then T would be the main successor to LT elsewhere and then back migrating from either SW Asia or East Africa to India quite a bit later.

AdygheChabadi said...

@ Andrew Oh-Willeke:
I hope I deciphered you clues correctly...hahahaha

I hope you received it...take care!

Let me know if you didn't...my e-mail is on my blogger page.

Romello said...

Why were Ethiopians labeled as "North Africans" in Table 1... lol? That really makes no sense at all.

terryt said...

"This is really surprising"

I thought that the connection between L and T has been accepted for some time, to the paper that separated KMNOPS from IJ and TL.

"And with K so widespread and diverse in Southeast Asia and Oceania, it is hard to argue for any other origin for K than South or Southeast Asia. So, either LT or L and T must have back-migrated to West Asia at some point- no?"

Not necessarily. As Y-hap F and a few descendant lines moved east, possibly from somewhere near Africa, some descendant lines may have simply dropped off along the way. G first of all, then IJ, then H, then TL. KMNOPS made it all the way to SE Asia where it broke up into K1, K2, K3, K4, M, S, NO and P. The last two moved on further: P back west and NO north.

Andrew Oh-Willeke said...

"G first of all, then IJ, then H, then TL."

If that is the scenario, TL has to split off before H which has great diversity and broad range in South Asia, while both T and L have very specific localizations in India and Pakistan respectively, and T has a broad range into Europe and Africa.

terryt said...

"TL has to split off before H which has great diversity and broad range in South Asia"

Not necessarily so. I doubt very much that any migrating population consisted of just one Y-hap. It could have already contained several derived clades. So if it contained Y-haps G, H, IJ, TL and KMNOPS any of them could have dropped off along the way, and become fixed in the remnant populations through the action of the old favourites: drift, bottlenecks and founder effects. We know that all the above haplogroups underwent later expansions of their own.

Eyal said...

Does anybody what company offers the SNP test for LT?

The Genetic Atlas website listed LT mutation as LT (Z1) for 2 years now, but no commercial company offers the SNP test.