No announcement yet.

MRCA for 5 individuals

  • Filter
  • Time
  • Show
Clear All
new posts

  • MRCA for 5 individuals

    My Y-DNA group project is for the Weisel surname.

    What I'd like to know, is it possible to determine the range of generations to the MRCA for the 5 test results shown below. Their haplogroup is R-M269. I didn't find the TiP report useful when comparing two individuals with a known common ancestor.

    I have 3, Y-DNA37 results for descendants of Weisel-0 (~1690-1770), likely from the Rhineland-Palatinate area of Germany. The 3 Weisels are about 9 generations of descent from Weisel-0. The 3 who tested, two descend from the oldest son of Weisel-0, and one (Weisel2) from the youngest son of Weisel-0. The origins of Weisel-0 are unknown even though a Weisel descendant in the 1930's made 2 trips to Germany trying to trace his ancestry. The test results confirm the 3 Weisels are related.

    In addition, the 3 Weisels are also a match with a Myers and a Moyer (US spellings).

    What I'd like to report on, is an estimate of a MRCA for the 5 individuals. I'm assuming the surname for the MRCA could have been either Weisel or a Meyer.

    Test results: (Genetic distance - GD)

    Weisel1: Weisel2 - GD 2; Weisel3 - GD 3.; Myers - GD 3; Moyer - GD 3

    Weisel2: Weisel3 - GD 1; Myers - GD 1; Moyer - GD 1; Weisel 1 - GD 2

    Weisel3: Weisel2 - GD 1 ; Myers - GD 2; Moyer - GD 2; Weisel3 - GD 3

    Thanks for your comments and opinions, BD Weisel

  • #2

    Not from the information that you posted.

    However, it is possible to calculate time to the most recent common ancestor using STR and SNP data.

    So..... I looked and found that there is a public Weisel project on FTDNA.
    It's very small and I'm thinking it is the project that you are referring to.

    Now, while Y37 STR results can be used to "match" people, they do not have enough information to perform a reliable calculation of the time to most recent common ancestor.
    Generally, I've found that 111 STRs are needed to get an age + or - 150 years.
    Of course, it depends on just how old the common ancestor is.

    Anyhow, on the project I saw 3 kits that you might be referring to:
    2 of the kits have only Y37 results.
    So, I ran a quick calc, just to see and found a common ancestor living 550BC (800BC to 250BC).
    Not very satisfying I know

    BTW, I have a Third Great Grandfather whose surname was Weis.
    He married a Hudlemeier, which could have been shorted to Myers.
    So, wondering if he might be a match.


    • #3
      Hello Andrew,
      Thanks for taking the time to look at my query.

      Actually the 3 kits I was referring to are: 558148 (1)-Michael; 398768 (2)-Karl; 844324 (3)-Mark

      All three are well documented and share a common ancestor to immigrant M. Weisel (1690-1770). This is about 9 generations. Calculating from 1700, their MRCA is 320 years.

      Because the two Meyer matches are as close as the 3 Weisels, I would think I could estimate the MRCA for all 5 as between 350-500 years.

      This isn't using a formula, but does it sound reasonable?

      BTW - Unlikely to be a match. Meyer and all it's variants is a VERY common surname. And Weis was most likely Weiss, another common surname. Likely assumed by someone with white hair.

      thanks, Bonnie


      • #4
        I tend to use Excel's geometric mean function for this type of analysis. It's typically used for heterogeneous populations, and I figure that's what we're looking at: different # of generations, generation spans, individual actual (vs. population theoretical) mutation rates.

        So what I'd do is perform a haplotype comparison in McGee's Y Utility, using 32 year average generation span and FTDNA mutation rates. The rates from the Macdonald calculator are probably more conservative in theory, but usually the results are off only a generation or two vs. FTDNA.

        Then I'd take the column of figures for TMRCA from the mode, and multiply them by two, because these figures represent only half the span of time since the MRCA in McGee.

        Then, separately, I'd construct a distribution curve in Excel in a new table using the geometric mean from the previous (doubled) TMRCA figures as the average, and the sample standard deviation of the individual (doubled) TMRCA figures from the previous table as inputs for the NORMDIST function. I would recommend calculating the sample standard deviation manually, in a column off to the side of the previous table, using the geometric rather than mathematical mean, just to be theoretically consistent, although my guess is that the difference between this result and the mathematical mean employed by Excel's canned function will probably be trivial.

        Just to have a nice round figure, I extend this distribution to 100 rows, calculating the 'number' input for the NORMDIST function as the geometric mean from the first table divided by 50 and multiplied by the current iteration number (i.e., number of rows since the head of the table).

        You can even include another column in this distribution table for an estimated year of birth for the MRCA, and use it as an axis label for any charts you might want to construct. This I calculate as an assumed birth year for a typical donor--just use your own if it's too big a deal to work out, the difference between alternative measures is unlikely to be large--minus the geometric mean from the first table, divided by 50 and multiplied by the current iteration number.

        Use the non-cumulative NORMDIST.

        The result will be a table estimating the probability that the MRCA is removed, on average, from all the donors, by that number of generations. It won't add up to 100% because you're using the non-cumulative distribution, and the average TMRCA in generations is unlikely to be a whole integer. But it's easy to add a column to calculate the cumulative form of the distribution if you really want to. You can create a line chart from these figures that will resemble the standard bell curve.

        If you're worried that the results could be skewed by the fact that some of the donors may be more closely related to one another than to other members of this analysis, you might want to consider aggregating by mathematical or (if there are groups of more than two) the geometric mean values. This would be important if you're talking about sub-groups whose MRCA is significantly more recent than the overall MRCA, but my feeling is that this doesn't apply here, where it sounds as if the difference between MRCA of sub-groups and the overall MRCA is only one or two generations.