Announcement

Collapse
No announcement yet.

How Far Back in History? Accuracy?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • How Far Back in History? Accuracy?

    Hi,
    The lowest % in my results is 3%. I always assumed that this meant one of my great-great-great grandparents was of that origin. I.e. 5 generations back there would have been 32 ancestors who had contributed to my DNA, 3 x 32 equalling 96%.
    If I guesstimated that each ancestor had, on average, been 30 years of age when I was born that would also mean that ancestor would have been likely to have lived around about 150 years before my birth.
    I also have one ethnic origin of 4%. I assumed that I had just inherited more of that relative's genes.
    Have I understand correctly? I can't find any explanations on this site.
    I am currently awaiting the results of a DNA test with 23andme in order to compares results but I know each company uses different methods. According to their website the ethnic origins will be apportioned on the basis of where ancestors would have lived 500 years ago before common sea and air travel.
    So is that the way Family DNA works?
    All contributions welcome.
    Thanks,
    Sally

  • #2
    The way I understand it, Family Tree's results refer to populations that lived 1,000 to 10,000 years ago. That makes it just about useless for us genealogists, but interesting to those who study population movements, I suppose.

    I don't think the ethnicity parts of these DNA tests are very accurate anyway. I heard one speaker, I forget who, refer to them as "parlor games".

    Comment


    • #3
      Admixture tools may claim to measure ethnicity in some distant past, but this can't yet be based on actual DNA from people who lived in those places at those times. Consequently, I'm not convinced that we can take admixture results at face value. I'm not even sure what we would mean if we said that someone's ancestors came from France 10,000 years ago, because I think it is very likely that most people who live in France today have at least some ancestors who migrated from somewhere else over the course of 10 millennia.

      Perhaps more important for the case where admixture tools report a small percentage, we have to be aware of a complication, frequently discussed here. Recombination at each generation is fairly limited, maybe a few dozen recombination events per generation. That means DNA is transmitted in fairly large chunks (connected to make chromosomes, of course). A big chunk from one ancestor on a chromosome means you didn't get a big chunk from some other ancestor at that same location. The statistical models suggest that the odds of getting any DNA at all from a particular ancestor, say 6 generations back, begin to fall off significantly. You can't be sure that you actually got any DNA from a particular ancestor that far back. Conversely, the same statistical models suggest that you may end up with a lot more than the computed percent from other ancestors equally far back. The expectation of a 3% contribution 5 generations back is a statement about the expected average, not about what we will see in any individual case.

      Comment


      • #4
        10,000 years ago is more into our past than most of us can imagine.

        6,000 years ago Sahara was green ( cf. http://en.wikipedia.org/wiki/Neolithic_Subpluvial ).

        9,000 years ago there was a shortcut for hikers from Denmark to England over a dry land (cf. http://en.wikipedia.org/wiki/Doggerland ).

        Broadly speaking 3,000 years ago, Germanic tribes were only in the northern part of Germany (so there were no Franks in France yet!), Celtic peoples were still before their expansion, there were no Slavic tribes in Central Europe, etc. etc. etc.

        Also, even when talking about recent 2,000 years, one should give a thought to population genetics equivalent of Theseus' paradox.

        Mr W

        Comment


        • #5
          Mr. W,

          It is only helpful then at higher percentages or is that questionable as well? Is the first most important thing to lay the foundation of your research and if you find a link a long the way does that explain to explain it or does it really not matter? Is it really just a way for FTDNA to make money?

          Comment


          • #6
            Originally posted by TXbama View Post
            Mr. W,

            It is only helpful then at higher percentages or is that questionable as well? Is the first most important thing to lay the foundation of your research and if you find a link a long the way does that explain to explain it or does it really not matter? Is it really just a way for FTDNA to make money?
            Please read the original description of clusters in myOrigins available through the Family Tree DNA Learning Center http://www.ftdna.com/learn/ftdna/myo...tion-clusters/. There are no dates given.

            The dates were an answer to a question during 2014 FTDNA Project Administrator conference. No details were given.

            I would assume that all the European clusters are the youngest ones. Possibly the Native American cluster is the oldest one.

            Yes, of course high percentages are more meaningful

            And yes, it is a good service (no advertising meant here), when you consider the target audience: Americans. If one knows nothing about own heritage, they may really appreciate knowing that their heritage is, for example, 50% British Isles and 50% Eastern Europe, as one may not be able to figure it out by any other means.

            Mr W

            P.S.
            In my opinion, some cluster names are grossly incorrect or misleading. Only Finland is located where the Finland and Northern Siberia heat map is - Northern Siberia is far away from there. Eastern Europe is right in the Central Europe.
            Last edited by dna; 6 February 2016, 01:51 AM.

            Comment

            Working...
            X