Announcement

Collapse
No announcement yet.

Other Spouses

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

  • #16
    RE "I am related to the first three, probably within the last 350-500 years, .."

    If we assume the "best case" and the added 13 markers match 13 for 13, then the 90% interval would be from 9 geneations
    back to 70 generations. At 25 years per genearaion the 90%
    interval would start about 250 years ago. So the statistics would
    support your hypothisis of 350-500 years.

    But until you have taken the test you do not know what the results are. Statisticly if you are 3 off at 12 you will "most likely" be 6 off at 25. The only way to know if it is 3 or 6 or something
    else is to have the test done.
    Ask your self these questions -- "What would I do if I found
    a 22 for 25 march? And what would I do if I found a 19 for 25 match?"
    If you would have a different course of action the test is worth something - if not then why do it?

    RE "My problem is I match about 10 per cent of all the European people's DNA, a 25 marker test would not prove any closer relation to these three, but would disprove some of the other surnames I am matching correct?"

    That is correct. Even if you upgraded to 35 markers and matched
    on 32 of 35 you would only pull the edge of the 90% interval back
    to 7 generations from 9 - not worth mutch in my book.

    Comment


    • #17
      Thanks for the line by line response, John. Our project has two identified lines, and several other probable (too different to be closely related). One of the participants conjectured that he was in the "John" line, and is off by one mutation on 391 (not a fast mutating allele). His paper records are not good enough to place him clearly in the line, but it is certainly possible. Do you think that being off by one in 12 makes it likely? Same surname and generally similar geography in his past. He would like to know, and I am not knowledgeable enough about this to answer him with any degree of confidence.

      Comment


      • #18
        RE "One of the participants conjectured that he was in the "John" line, and is off by one mutation on 391 (not a fast mutating allele). His paper records are not good enough to place him clearly in the line, but it is certainly possible. Do you think that being off by one in 12 makes it likely? "

        With a case like this it is all a matter of statistics and the individual
        definition of "likely"
        Some people want more proof than others.
        To me 11 of 12 and a surname match is not enough.
        Look at it this way from father to son it is UNLIKELY there will be
        a mutation. But over 100 generations it is UNLIKELY we have NO
        mutations. That is what makes this technology hard to apply.

        I can run the calculations and give you the statistical percentages.
        The part we have no way of knowing is what percentage should
        we give to a match in surname and a match in location.
        You wrote "generally similar geography in his past."
        This is also subject to subjective evaluation based on how
        general is "general"
        For example from 1750 to 1890 one "general" path is
        VA to KY to IL. That to me would not be helpful even it
        all the families had it.
        On the other hand if it is the came county in each state
        that make it higher probability.

        One of the items you did not mention is the values of the other
        markers. If they are all the modal for the haplogroup and that
        is R1b we have a much higher chance of an accidental match.
        If there are one or more that are in the range of 1% we have
        increased the odds of a family match.

        I do not like to "confirm" relatedness based on even 12/12 markers
        because I know there is a 20% chance an upgrade to 25 markers
        will show they are not related - this is even with a last name match.

        If the families cannot find a paper link given the 11 of 12 match
        I would recommend upgrading to 25 for two of he men.
        Then look at the results and see what we have.
        After the upgrade to 25 markers we will have these possibilities
        24 of 25 Highly probability of being related
        22/23 of 25 probably related given the last name match
        Less than 22 of 25 - probably not related.

        That is a lot of words and I am not even sure I have answered your question.

        Comment


        • #19
          JOhn. Thank you again. The individual with the single mutation is R1b both for an exact (non-project member) match and one-step. The line he is only one marker off, is also R1b, but only at one step matches (they have no exacts except, presumably, each other). Another line, that doesn't even come close, is also R1b and has an exact (non-project member) match. I have asked for more participants, and one has joined. With a few more, I will feel more comfortable asking for people to upgrade. I agree that we need more precision. The best part of this surname project is the vast quantity of written genealogical record that exists, back to around the 1580s. I could give you all the numbers for each marker, but I think it is a good idea to hold off a bit until I have more information. If I submit information to youfor analysis, I would prefer to do this privately, and not on the Forum.

          Comment

          Working...
          X