No announcement yet.

New Buryat G2a subgroup?

  • Filter
  • Time
  • Show
Clear All
new posts

  • New Buryat G2a subgroup?

    Long ago posted my wife's mtDNA and Vraatyah had predicted G2. (She is a Western Buryat). We now have an *almost* complete sequence that seems to indicate G2a but does not perfectly match anything I can find. Vraatyah?

    73G-151T-152C-237G-263G-315iC-489C-709A-750G-1189C-1438G-2706G-4769G-4833G- 5601T-(6800-7400)-7600A-8701G-8860G-9377G-9540C-9575A-9932A-10398G-10400T- 10804G-10873C-11719A-12372A-12705T-13563G-14569A-14766T-14783C-15043A-15301A- 15326G-16129A-16223T-16274A-16278T-16362C

    The 6800-7400 region is not done just yet but will likely have it in the future.

    Note that this sequence was done privately and independently outside of FTDNA, and that it CONFIRMS the 237G found by FTDNA, which is apparently unusual.

    Last edited by dentate; 14 October 2006, 09:38 AM.

  • #2
    Hello Jeff, it seems I overlooked your topic. Just checked my collection of the complete sequences and found no matches. Unique variant, indeed.



    • #3
      Thank you Valery! Would you agree that her best match is to G2a, with back-mutations at 14200 and 5108? If she has those back mutations then she would branch off right after the five mutations that Tanaka says define G2a. Despite our discussion a few months back about whether G2 was Central Asian or Siberian in origin, if this is a variant close to the root it suggests that there may be more diversity and more time depth in the Baikal region.

      Interesting that G did not make it to the New World along with all the other groups in that West Baikal/Tuvinian region. I presume it arose later and arrived in the area after that migration was over.

      Very interesting about your personal T1, too. I am in a very different situation. I have well over 100 exact matches on FTDNA, with my N1b. Very little doubt about my maternal ancestry! Must be a reflection of societal restrictions on the degree to which mates could be randomly selected.

      We appreciate your time and your help.


      • #4
        Jeff, here is a mj-network for Gs (image in g.jpg and descriptions in Another variant you'll find in Tanaka's paper, I know you have read it. The second must-read paper in Kong 2003 with a revision published this year. Both papers include Bandelt's mj-networks with indepth phylogenies for G. I emphasize: as usual, the authors published processed diagrams where from many competing possibilities they chose the most-likely one and cleaned the remainder.

        I left the "raw" network unchanged, you can see a rectangle formed by 2 sites 16227 and 14200. Your wife's lineage (marked ****) represents a "wide" state where both sites have the same value as CRS, it's the root; in Tanaka's paper you'll find sequences with 16227 without 14200; Kong published a couple of sequences with reversed composition of these sites. Finally, both papers have lots of 16227-14200 variants that form well-known G2a1 haplogroup. So, all 4 variants do exist and constitute a classical reticulation in the network which is forced to be a general graph rather than a tree. A priori you may choose every refining of that figure, and each decision can fit the parsimony formalism after an appropriate reweighting of both sites (I left their weights equal deliberately, to show this figure). From the parsimony view, all 4 variants are equally possible until we choose weights which break one link. However, there is one unrealistic variant, it's the case when the left upper 227 link is broken, it means that 14200 emerged independently in 2 branches; that's too unrealistic; intuitively, its back-development is more possible. The algorithm we employ does not know where the root is and so not aware of what "back" is, it's a kind of indirect-graph algorithms. We do know that and are already limited in our choice of pathways, only 3 candidate link are left. Tanaka solved this problem by breaking the right bottom link which is the same as to say "14200 mutated back". I don't know what is correct. Maybe the Buryat lineage forms a cluster common in Siberia, and we would name it "pre-G2a"?

        So, my point hasn't changed: somewhere near the G2a root.

        Attached Files
        Last edited by vraatyah; 24 October 2006, 01:10 PM.


        • #5
          Valery, as usual I am grateful for your thoroughness, for the time you devote to this, and by your unbiased logic. Not to mention, very impressed by your flawless English. Thank you for reproducing the diagrams.

          Yes, I have seen those papers and I agree with your conclusions.

          The Buryats are clearly a polyphyletic collection of peoples. I suspect that surveys of the Buryat gene pool conducted in Ulan Ude are going to miss some interesting material. Logistically it is much easier to work there, but there may be some interesting heterogeneity in the outlying areas such as Ust Orda and Aga Buryat.

          Thank you again!