February 15, 2011

Y chromosomes of Libyan Tuareg

The gene pool of these samples seems quite simple composed entirely of E1b1a8-U175 (sub-Saharan), E1b1b1b-M81 (Northwest African), and R-V88. I find the absence of other clades of Middle Eastern or European origin as extremely interesting, pointing to the extreme value of this population as a key to North African prehistory.

Haplogroup U175 is most frequent in Sub-Saharan Africa and it belong to clade E-M2 which trace their descent to east Africa but became more frequent in Sub-Saharan Africa.

In the next Dodecad Project update, I will probably have a North African sample of more varied composition than the HGDP Mozabites and Behar et al. (2010) Moroccans. Note that North Africans are part of the current call for submissions in the project, so I would love to have more samples from that region of the world.

Am J Phys Anthropol DOI: 10.1002/ajpa.21473

Deep into the roots of the Libyan Tuareg: A genetic survey of their paternal heritage

Claudio Ottoni et al.

Recent genetic studies of the Tuareg have begun to uncover the origin of this semi-nomadic northwest African people and their relationship with African populations. For centuries they were caravan traders plying the trade routes between the Mediterranean coast and south-Saharan Africa. Their origin most likely coincides with the fall of the Garamantes who inhabited the Fezzan (Libya) between the 1st millennium BC and the 5th century AD. In this study we report novel data on the Y-chromosome variation in the Libyan Tuareg from Al Awaynat and Tahala, two villages in Fezzan, whose maternal genetic pool was previously characterized. High-resolution investigation of 37 Y-chromosome STR loci and analysis of 35 bi-allelic markers in 47 individuals revealed a predominant northwest African component (E-M81, haplogroup E1b1b1b) which likely originated in the second half of the Holocene in the same ancestral population that contributed to the maternal pool of the Libyan Tuareg. A significant paternal contribution from south-Saharan Africa (E-U175, haplogroup E1b1a8) was also detected, which may likely be due to recent secondary introduction, possibly through slavery practices or fusion between different tribal groups. The difference in haplogroup composition between the villages of Al Awaynat and Tahala suggests that founder effects and drift played a significant role in shaping the genetic pool of the Libyan Tuareg.

Link

12 comments:

astenb said...

Interesting, 1 R1* sample as well.

Andrew Oh-Willeke said...

"I find the absence of other clades of Middle Eastern or European origin as extremely interesting, pointing to the extreme value of this population as a key to North African prehistory."

Wouldn't one have expected something very similar in a sample this size from this area if asked to guess the results in advance? What is it that makes this "extremely interesting" and gives it "extreme value" as opposed to being just one more confirming piece of data?

Gioiello said...

@ Astemb “Interesting, 1 R1* sample as well”.
Have you got the paper? Can you send it to me, at least the Supplements?
Anyway R1* is due to the most ancient dispersal, probably present all over the world. We have some samples among the !kung and others. To understand the origin of an haplogroup it is more important the pathway of it, and North and Central Africa has above all R1b1a (V88+).

gioiellotgnn06@gmail.com

astenb said...

They supporting files are in the link below:
http://onlinelibrary.wiley.com/doi/10.1002/ajpa.21473/suppinfo

Are the !King samples really R1* though? I am weary of any R1* samples produced prior to the identification of R1b1a or (V-88). Sure it could be valid but the sampling of R1b1a ALONG SIDE R1* shows this to be pretty legit.

Gioiello said...

Many thanks, Astemb. I found the !kung R1* in a paper of Rozen I spoke about many times on Worldfamilies and perhaps also here. Unfortunately I use remember the papers read and I haven't it on my PC, but I have the paper printed somewhere. I remember that that !kung was found in an American hospital (he had many diseases) and was SNP tested, but perhaps he was a R1b, anyway one of the most ancient in our haplogroup (at least mine).

Gioiello said...

The paper is: “Remarkably little variation in proteins encoded by the Y chromosome’s single-copy genes, implying effective purifying selection” (AJHG, 2009, 85) by Rozen et al.

Gioiello said...

The !Kung is R*:
M207/UTY+325+18
R M207/UTY2
but I don't know if he has other SNP of some subclade.

Umi said...

Did they test for both R1a and R1b? I'm not sure if this R1* is legit.

pconroy said...

I see that sample "AW60", reported as Haplogroup=Q - based on Whit Athey's Haplogroup Predictor - and SNP=R1* - based on bi-allelic marker arrays.

I just ran the STR values for sample AW60 through Whit Athey's Haplogroup predictor myself, with the setting:
Area Selection=Equal Priors

And found that some markers could not be entered, as follows:
1. had to set 607= >18, as no 19
2. 724 not included in analysis
3. GATA H4.1 not included in analysis

Prediction:
R1a=90.9%
N=8.9%
Q=0.1%

So not R1* afterall, but R1a

Gioiello said...

Yes, the Athey’s predictor gives R1a. Anyway he wrote to me he was working to a new predictor, but I don’t know if this is the old or the new, but he said that the new would have had more subclades, then this is the old, I think. The old was good, but lacked where it was more necessary, i.e. on the borderline haplogroups.
DYS724 is CDYa,b, GATAH4.1 =22 corresponds to 12 in FTDNA values.
If the authors have SNP tested AW60, it is R1*.

pconroy said...

Gioiello,

But my point is that the paper says that using Whit Athey's Haplogroup Predictor, sample "AW60" = Q

But that is patently NOT the case. So either they made an error, or I'm using a later - read more accurate - version of the predictor.

Now since we know R1a is attested in Egypt and Sudan, this would not be earth shattering news

Styan said...

Styan said...
In the last several years I have spent a lot of time studying the STR data in Ysearch, especially in the 67 marker haplotypes from which I usually study 36 of the slower or more stable STRs. The aim is to find out what groups exist and where they are found. I was pleased to find that the supplementary material for this paper includes the first 16 of the STRs I usually study, namely: DYS 393, 19, 385a, 426, 388, 392, 459a-b, 455, 454, 437, 448, 460, YCA IIa + IIb, DYS 438, considerably more than in many comparable papers. The comments below apply only to consideration of these 16 STRs.
E1b1a8-U175: The 12 examples include 8 with the same figures for all 16 STRs and 3 more with the same deviation at one point (DYS 19 = 15 instead of 16). However, TAH4 has different figures for 7 of the 16 STRs. He must belong to a different branch of E1b1a8, only distantly related to that represented by the others. Unfortunately I have given little attention to Sub-Saharan haplogroups so I cannot comment further on this group.
E1b1b1b-M81: The 19 examples include 15 sets of 16 identical figures and 4 with 1-2 deviations, different in each case. The 15 identical sets of figures are also identical to the following haplotypes found in Ysearch: 8mjah and C59BN from Algeria, QGVMV from Mali, P6C9U from Morocco, RY9WZ from Spain, 4YMQ6 from Portugal and TEP27 from England. One of the clearest distinguishing features of this group is DYS 454 = 12 (instead of 11). The new information confirms what I had already concluded from Ysearch: that E1b1b1b-M81 has unusually little variation and so it must be one of the youngest haplogroups.
R1b1a-V88: There are three examples, each a little different from the other two. These examples combined with those from Ysearch confirm that this group usually has DYS 393 = 13, 454 = 12 and YCA IIa = 21. AW32 is very close to Ysearch CEFQX from the USA. They differ only at DYS 460, the least stable of the markers considered here. R1b1a-V88 appears to be a rather varied group. It would be good to have considerably more extended haplotypes from African members of it.