May 29, 2010

Comparison between morphological and genetic data for Egyin Gol Mongolians

I had first blogged about Egyin Gol in 2003, that paper is freely available here. From the current paper:
The Egyin Gol necropolis is located in the Egyin Gol Valley (northern Mongolia), near the Egyin Gol River, close to its confluence with the Selenge, a main tributary of Lake Baikal (see Fig. 1). This site has been the subject of a French-Mongolian interdisciplinary research project from 1997 to 1999, which allowed the excavation of 84 graves containing skeletal remains of 99 individuals buried from the third century B.C. to the second century A.D. The graves were organized in three main sectors (A, B, and C) that, based on AMS carbon-14 dating of human bones, progressively expanded from south to north (i.e., Sector A is the oldest followed by Sector B and Sector C). The development of Sector C corresponds to the end of the necropolis and may reflect a Turkish influence on the Xiongnu tribe (Keyser-Tracqui et al., 2003).
The results showed, however, that individuals buried in sector C represent a specific kin group clearly differentiated from the rest of the necropolis based on nonmetric
data (Table 4), and confirmed by the genetic data. This might be explained on the basis that these individuals are suggested to be of Turkish origin, based on their shared single paternal lineage, unique in the necropolis and affiliated with Turkish populations (Keyser-Tracqui et al., 2003). However, the sector C individuals share the same maternal lineages with individuals buried in sectors A and B, which could explain the global homogeneity of the population as a whole. The particular characteristics of the sample from sector C suggests a possible shift in the population demographics, caused by the emergence of a Turkish component in the Xiongnu population at the end of the necropolis use and at the end of the first steppe empire led by the Xiongnu. The fact that this particular subgroup of the population buried in sector C was detected by nonmetric traits analysis demonstrated that nonmetric traits are an efficient tool when analyzing population microevolution.
The Y chromosome results are found in Table 2 of the original paper.

I ran the Y-STR profile of the shared patrilineage over the haplogroup predictor, but I don't get a clear estimate for the Y-STR profile (grave 46 in Table 2).

A YHRD search gave the following results:
The high frequency in Kazakhs and Yakuts, with a little spillover in both China and eastern Europe is certainly consistent with a Turkic origin of this haplotype.

Finally, I gave it a try at ysearch, getting a match with a Pole and a 1-step match with a Czech, both of which are listed as C3-tested.

So, there you have it, archaeology, non-metric data, Y-chromosomes, and a little use of online tools gives us a first glimpse on what may have been a group of ancient Turkic individuals. Of course here are theories-a-plenty about what language the Xiongnu originally spoke, so it would be premature to arrive at any firm conclusions.

Interestingly, C3 is also present in a different 2,000-year old Xiongnu individual from NE Mongolia from the Duurlig Nars site, but an earlier group of Xiongnu-related individuals from China (Pengyang) belonged to haplogroup Q.

American Journal of Physical Anthropology doi:10.1002/ajpa.21322

Comparison between morphological and genetic data to estimate biological relationship: The case of the Egyin Gol necropolis (Mongolia)

François-X. Ricaut et al.

Osseous and dental nonmetric (discrete) traits have long been used to assess population variability and affinity in anthropological and archaeological contexts. However, the full extent to which nonmetric traits can reliably be used as a proxy for genetic data when assessing close or familial relationships is currently poorly understood. This study represents the unique opportunity to directly compare genetic and nonmetric data for the same individuals excavated from the Egyin Gol necropolis, Mongolia. These data were analyzed to consider the general efficacy of nonmetric traits for detecting familial groupings in the absence of available genetic data. The results showed that the Egyin Gol population is quite homogenous both metrically and genetically confirming a previous suggestion that the same people occupied the necropolis throughout the five centuries of its existence. Kinship analysis detected the presence of potential family burials in the necropolis. Moreover, individuals buried in one sector of the necropolis were differentiated from other sectors on the basis of nonmetric data. This separation is likely due to an outside Turkish influence in the paternal line, as indicated by the results of Y-chromosome analysis. Affinity matrices based on nonmetric and genetic data were correlated demonstrating the potential of nonmetric traits for detecting relationships in the absence of genetic data. However, the strengths of the correlations were relatively low, cautioning against the use of nonmetric traits when the resolution of the familial relationships is low. Am J Phys Anthropol 2010. © 2010 Wiley-Liss, Inc.



  1. There aren't any ethnic Poles with C3 in Ysearch, just people with ancestors born in the former eastern Polish provinces, like in some town called Ruskati. God knows where that is now.

  2. The above comment is factually incorrect on two different counts. One Polish C3 in Ysearch is from Minsk Mazowiecki, which is within modern Polish borders. The other Polish C3 in Ysearch is recorded as from Ruskati. I have not identified the precise location of that town either, but the individual is ethnically Polish regardless of recent border changes.

  3. "The high frequency in Kazakhs and Yakuts, with a little spillover in both China and eastern Europe is certainly consistent with a Turkic origin of this haplotype".

    Without really questioning this conclusion, I must mention that you are in fact getting almost double apportion of matches in Moscow (1/81) than in Yakutia (1/183) at the YHRD search. However the Kazakh connection seems solid.

    Notable for me is that 11% of mtDNA lineages are of South/West Eurasian origin (misleadingly reported as "European" in the paper): U2, U5a1a and J1. It is somewhat notable because the location of the site is right south of Lake Baikal, suggesting that these lineages could have arrived either in Neolithic or even Paleolithic flows (there is a mild penetration of "Western" techno-cultures from Altai into Mongolia c. 30 Ka ago).

  4. Y-STR matching of the samples:

  5. Sorry, the link I provided seems dead, here's the new link, Y-STR matching of the samples from 3 sectors:

  6. If the samples of grave 69 and 76 are really R1b(...), anyway their YCAII=18-21 and 23-23 demonstrate, I think, that they belong to the most ancient R1b1* of Asian descent (probably generated from the Western one with 18-23 or 18-22 by a RecLOH) and have nothing to do with the ancestor of all Western European subclades. See what I have written in another thread speaking with Argiedude.
    If Maju is right, probably they came from Europe at least before the Younger Dryas.

  7. I may be wrong but a quick comparison with the haplotypes of Alonso'05 (using only DYS 19-390-391-392-393), suggests that, if individual #65 was R1b1b2 (as Natsuya suggests with 60% of chance), then he should belong to a very frequent haplotype of, probably, R1b1b2a1(xR1b1b2a1a) that is common in Central Europe, specially in Belgium and Austria.

    But they would be as logical, at least, to be O2b. Sincerely, it is most difficult to discuss only on Y-DNA STR haplotypes, we really need SNP-based classifications to make any sense in most cases.

  8. Yes, I think the samples are more likely to be O2b.
    If it's really R1b, R1b-P25(xM269) or R1b-M73 are also possible, because these haplogroups are more common in Central Asia. What's your opinion? I remember it was said that SNP testing will be done on these ancient samples.

  9. "If it's really R1b, R1b-P25(xM269) or R1b-M73 are also possible, because these haplogroups are more common in Central Asia. What's your opinion?"

    In truth I don't really know. STR haplotypes give me headaches, really. I have looked only to some extent those belonging to R1b1b2 but they are still quite messy and can only be addressed in large populations. I would suggest to SNP-test the bones... but guess it won't be. :(

  10. Grave 69 is likely M269, and grave 81 is a possibility of some R1b variety if other markers hadn't failed. AFAIK R1b1* has not been found in Central Asia with the exception of a single Mongolian on ysearch. All the R1b has been R1b1b1 or R1b1b2 to date.

  11. To Aaron:

    Wasn't grave 81 confirmed by SNP testing to be Q-M242?

  12. Aside from C3(xC3c), N1c-Tat and Q-M242, those R1a1a are not so surprising.
    We all know the rise of Xiongnu empire has something to do with IE-speaking Scythians.
    The presence of H, J, R1b1b2, O2b and O3 says Xiongnu is united clans of different ethnic origins.


Stay on topic. Be polite. Use facts and arguments. Be Brief. Do not post back to back comments in the same thread, unless you absolutely have to. Don't quote excessively. Google before you ask.