No announcement yet.

Haplogroup T ("Tara")

  • Filter
  • Time
  • Show
Clear All
new posts

  • Haplogroup T ("Tara")

    Hi -first some info of hopefully general use, at least to other "T's". I extracted all the T's from the mitosearch database: there are 71 such people as of Dec 26, 2004. If the current estimate that "Taras" make up nearly 10% of native Europeans, then this suggests (assuming randomness and forgetting about small numbers) there are approximately 710 European samples on the database. Not very many, especially given the Y chromosome database nears 20,000 I am told, though perhaps not too bad given its fairly new. The oxfordancestry database seems to have even less Europeans on the mitochondrial database, though they claim that "soon" they will make availble 10,000 samples that have been collected from various research studies (though not all european).

    The breakdown of the number of people that have the different T subgroups on the mitosearch database are as follows:

    T 12
    T* 6
    T1 23
    T2 23
    T3 3
    T4 2
    T5 2

    According the data provided by Vincent McCauley, the criteria used in the Richards et al. 2000 study was that to be in Haplogroup T, one has to have mutations at 126 and 294. That is, all the different T sub groups must have those two mutations to be classified as T. (This info is availalbe on-line- go to the website linked in the 4 page pdf file that comes with your on-line results. The pdf file is under the link "understanding your results") (side issue: Does this imply that only T’s have both of these mutations? What about just one of the mutations? Is it in any way part of the criterion for any other Haplogroup?) In addition, the criterion for T5 subgroup is a mutation at 153.

    I mention T5 because my recent data (I was in batch 99- just got my results) was classified as Haplogroup T5, and I do have all three of those mutations (126, 153, 294).

    In addition, I have 4 more mutations in the HVRI region for a total of 7 mutations, and I'm not quite sure yet what to make of them. The additional 4 mutations are at 114, 192, 221, and 519.

    The only other person who is also a T5 on the mitosearch database also has a mutation at 519 (in addition to the 3 required mutations), but he has none of the other additional 3 mutations that I have and he has a mutation at location 292 which I don't have.
    I checked the rest of the McCauley database (the data that went into the 2000 study), and there is no exact match to my 7 mutations, or even to 6 of them.

    Best I can piece together, thinking out loud, is the following. I share in common with the other 70 individuals maternal clan mother Tara, who was believed to live approximately 17,000 years ago in Italy. She is the maternal ancestor of the 70 of us, more than 500 generations ago. Sometime between 17,000 and 10, 000 years ago (whenever T5 is beleived to occur) , our lines (mine and the other 69) go seperate ways, except for one of those individuals. At some point after T5 (let's say that was 13,500 years ago), probably rather quickly, my line goes a seperate way from that other (70th)T5 indiviudal's line. I say rather quickly because I still have 3 more mutations to accumulate, which doesn't leave much time!

    I guess the estimaste that there is 1 mutation every 10,000 years is obviously just an average. If there's a high variabilty to that estimate, I would think that makes it very hard to determine the timelines.

    One other detail is that i tried looking up my other mutations that aren't part of a typical tara pattern - so far I come up with those matching "helena"-but that's not suprising, since for that search i'm putting in one mutation at a time, and since helena is the cambridge reference standard, it's not suprising that only 1 mutation wojjld turn up helenas most often, especially since they make up 40% of eureopeans.

    Any comments on my pattern of 7 mutations most welcome (to summarize, my pattern for HVRI is 114T, 126C, 153A, 192T, 221T, 294T, 519C). Hard to know if this is an unusual pattern given so few people on the database. (If it is unusual, I guess this suggests than if I ever find an exact match, the odds are probably higher than the average that we share a reasonably near ancestor). Otherr comments on taras, or on the time line or on good articles to read also most welcome - if i seem to get anything wrong above please also let me know that too. (I also got the 2nd control region analyzed, but it deons't look like i need a high resolution analysis quite yet...though quick scan suggests i may have more matches to that region than to the main control region).

    guess that's it for now.

  • #2
    penguin, 114-126-153-192-294 found in Portugals


    • #3
      Originally posted by vraatyah
      penguin, 114-126-153-192-294 found in Portugals
      Do you mean Portugals as in people from Portugal? That's intersting since
      my maternal line is sephardic, which would mean spain or portugal- but I didn't think that many years ago. I beleive sephardics left for spain and portugal (iberian penusila) only 2500 years ago- could the mutations have occured after that?

      Do you have reference for the sequence you mention above in Portugals?

      Thanks for your help!


      • #4
        penguin, this is due to my bad English rather than multiple "Portugals" )

        There is 1 Portuguese sequence from Central Portugal in our database, I think it was published but the corresponding link is missing in the base.



        • #5
          OK, I recalled. The paper is

          Diversity of mtDNA lineages in Portugal:
          not a genetic edge of European variation
          L. PEREIRA et al.

          Ann. Hum. Genet. (2000), 64, 491-506

          hvs1 114-126-153-192-294
          hvs2 73-150-263-303.1-311.1


          • #6
            >I beleive sephardics left for spain and portugal (iberian penusila) only 2500 years ago- could the mutations have occured after that?

            to calculate the coalescence correctly we should collect all T 114 153 192 sequences together. You see we have only 2 samples so that is out of the question. Since direct calculation is not available we have to restrict ourself to different ways - such as analogy, mutations in other clades etc. - which are all inexact. I think 2 transitions (153 is quite a common) found in more than 1 population might originate in more ancient times than recorded history.
            Last edited by vraatyah; 28 December 2004, 05:32 AM.


            • #7
              ps. your additional transition at 221 may indicate recent event since its virtually absent in T background and quite a rare elsewhere.


              • #8
                Hi- thanks for the ref of the patterns in portugal. I read the article and checked out the appendix at the end. Saw the individual you were referring to. My HVR2 mutations are also 73, 150, 263, but with 2 C insertions at different regions than his/hers at 309 and 315, but i guess some of all of these are hyper hyper variable and unstable from what i've seen.

                Looks like in that study for HVR1 they only went up to around 16385, so
                I wouldn't be able to match the 519 anyway (which is quite common anyway i understand), so i may only be differnt trom that individual in hvr1 at that 221 position.

                interesting in the article he wasn't classified as T5, but instead the unidifferentiated T*. I know some people don't agree with the classification scheme after T1, in part because of instability at regions 292, 296, 294 (?), althouth these authors claim they use the classificaiton of McCauley richards etc., but then don't really do so.

                One question i do have is can you tell me the size of the data base you checked? If you checked 200 and I had one near match I'll think about that differently than if you checked 8,000 and I had one near match.

                thanks again for your help.


                • #9
                  >If you checked 200 and I had one near match I'll think about that differently than if you checked 8,000 and I had one near match.

                  we've collected 35K(!!!) sequences from all eurasia in our database, about 1/6 unpublished.

                  >but with 2 C insertions at different regions than his/hers at 309 and 315

                  your HVS2s are exactly the same: the paper authors use 3'-5' alignment instead of common 5'-3' which FT and most labs use. you can check it trying to collate both sequences between 303 and 315:

                  mtDNA light strand, 5'-3' 301-320 bp

                  CRS (1)
                  or (2)
                  gaps where insertions may occur are introduced in another manner

                  your mt:

                  from the paper:

                  Last edited by vraatyah; 29 December 2004, 11:23 AM.


                  • #10
                    Interesting - thanks. I didn't look closely enough at the methods. And thanks for info on number of samples in your database.

                    I'm wondering if my mutation at 221 is an artifact, perhaps a sequencing artifact as discussed in articlerender.fcgi?artid=385090 (article on phantom mutations). It was one of the artifactual mutations in the '96 italian paper on ladin speaking populations.
                    If 221 is an artifact, and since the portugal study didn't look for mutations above approx. 16385, and even since the 2nd control region of that individual matches mine (though i don't know how common that is), that person looks like it could be an exact match. I guess that's one match in 35,000 samples of largely europeans - seems like that's not many matches! all very intersting.


                    • #11
                      >I'm wondering if my mutation at 221 is an artifact, perhaps a sequencing artifact as discussed in articlerender.fcgi?artid=385090 (article on phantom mutations). It was one of the artifactual mutations in the '96 italian paper on ladin speaking populations.

                      Your question concerns a problem which is too deep for this forum. As you know Bandelt's study was based on 2 methods. The first, his own RM algorithm, seems to be quite a sensitive to find most recombinations in particular dataset (becose they usually result in redundant homoplasy). The second is plain (but not a stupid) "frequency" check which filters rare mutations out, as well as transit/transvers ratio violations. Both methods we can apply simultaneously - you see, all rare sites are at least "extremal" in Bandelt's terms, your 221 is a "peripheral" and doesn't cause homoplasy. Your sequence will be a final "leaf" in the T5 haplogroup tree. So this sophisticated RM method has nothing to do with your 221 site.

                      Another approach, Bandelt's MJ method not mentioned in "Phantom" paper, orders sites ("links") according to their weights and builds the median network which includes some possible trees.

                      Actually we have only 3 sequences: ancestral T3, Portuguese and yours, and 2 sites: combined 114-192 and 221.

                      --------153---------/T5/------114------/x/---192---/Portuguese one/

                      where can we place yours? The link corresponding to 221 transition has length at least the same as 153th. Without weight "relaxation" we can obtain only one tree with final 221 link:


                      --/T5/------114------/x/---192---/Port/-------221----/your mt/

                      we can add 1 median type 153-221, set "threshold" > 0 and then "downweight" both 114 and 192 sites that makes this network very doubtful:


                      /Port/--192--/x/--114--/t5/---221---/x/--114--/x/--192--/your mt/

                      (in the most extreme case we have a rectangle with 2 possible evolution pathways for your type)

                      I don't think we should ignore the weight of 114 and 192 since both transitions are uncommon in T background. After all, any "weighted" transition such as 221 may occur even as final link in the tree, so I prefer the scheme (1).

                      Frankly, we cannot absolutely exclude artificial origin of 221, but I've seen at least 15 groups, each of 1-4 identical sequences with transition at 221, and about 90 unique sequence types with 221.

                      Last edited by vraatyah; 17 January 2005, 02:31 PM.


                      • #12
                        New T5

                        I just found out I'm also T5.

                        HVR1 - 126, 153, 183, 189, 294, and 519

                        HVR2 - 73, 150, 247, 263, 309.1, and 315.1

                        I have no exact matches to anyone or any country at the present time, which is disappointing. Hopefully, as more people submit results that will change?!

                        My family is originally from The Netherlands.


                        • #13
                          T5s seafaring people?

                          Dear fellow T5s:

                          After googling T5, I've found a real lack of data on this haplogroup. The data that is available is somewhat conflicting. In Gonzalez et al. "Mitochondrial DNA Affinities at the Atlantic Fringe of Europe" ( the conclusion is that T4 shows greater frequencies in northern europe while T5 shows greater frequencies in southern europe and presumably north africa. However, the greatest single frequency shown for T5 is in Norway at 1.2%, followed by North Portugal at 1.1% and 0.9% in France. The other studies I have seen have even less data on T5 than Gonzalez et al. Could this be accounted for by Viking settlement on the Atlantic fringe?

                          My maternal ancestors are from Wieringen in the Province of North Holland in the Netherlands. Wieringen until recently was an island populated by west-frisians but was also subject to both Viking raiding and likely settlement (Viking silver hoards have been found locally).

                          The local population in Wieringen is west-frisian. The Frisians have a long history as a seafaring people. The Frisians were early empire-builders among the scattered tribes of the Dark Ages and by the sixth century were the most coherent and prosperous tribe among the Germanic people, controlling an area stretching from northern Jutland in Denmark to Flanders in Belgium. Most of inland Europe was in those days impassable due to poor roads and lawlessness, and the Frisians dominated sea-going trade, ranging as far a field as the Baltics, Russia, Scandinavia, England and southern Europe. Indeed, the North Sea was then known as Mare Frisicum.

                          Wieringen, being an island, was somewhat isolated for centuries. There is no evidence of any immigration from southern europe or north africa in the region for the past 2 millenia of which I am aware.

                          Hopefully at some point an effort will be made to identify a distinctly west-frisian or even a frisian haplogroup distribution. My guess is that it will show a higher prevalence of T5 similar to that in Norway.

                          My Y results indicate I am in the I haplogroup (Athey suggests I1a but I will need SNP to confirm) which is consistent with a Frisian or Scandinavian heritage.

                          As my most of my ancestors can be definitively tied to the northern part of The Netherlands back to at least the early 17th century and there is no indication of anything other than Frisian origin, it would seem rather anomalous for my T5 haplogroup to originate in southern europe or North Africa in the relatively recent past.

                          Is anyone aware of any studies with a larger sampling of T5s than Gonzalez?



                          • #14

                            Penguin, I noticed that you mentioned that your mother's line was Sephardic. My husband's mother's line is Sephardic/Romaniote from Rhodes, Greece. And his mtDNA has been lumped into T, like yours (but T1, though)...*and* he has the unusual 16519C, like you, which we can't find any other matches for within his haplotype on He's...

                            HVR1 Mutations

                            HVR2 Mutations

                            I wonder if there's a connection?

                            - Brooke Schreier Ganz
                            Los Angeles, California


                            • #15
                              Hi Brooke-

                              The 519 mutation is quite common (its a couple of the others that are unusual).

                              I would really love to see breakdown of sephardics from Greece
                              (ours is from Volo greece, which could be romaniote also-not clear where they were before volo) on mitochondrial DNA. Have you come across anything relevant? Theres plenty on y chromosome, and some stuff on MT dna of Moroccons.