No announcement yet.

Comments please, on "half" and "true" siblings

  • Filter
  • Time
  • Show
Clear All
new posts

  • Comments please, on "half" and "true" siblings

    Determining true siblings from half-siblings by share seems a reliable test, in that two generations back from the tester is a fairly predictable "quartering" of same-grandparents and same-parents shares. Half-siblings do not normally share grandparents. But when are half-siblings truly "half"?
    What if the "male in the middle" contributes to a son by wife 1, then produces a son by the sister of that wife? This scenario is quite common when a man is widowed with children, and the sisters could also possible be twins (sorry to go there). Would that not give the appearance of true siblings if those two were tested for atDNA? Wouldn't the descendants also show closer shares than descendants of different-grandparent shares?
    I use sons only because these two brothers would share the Y from the father. I do not see how atDNA would be able to confirm or deny such an occurrence. I am also not clear if mtDNA is of any value here, since it seems very non-specific.
    I realize that these shares become more and more randomly distributed as they are diluted down the gens, but restricting this to 3 gens, is this a reasonable stumbling-block to knowing whether these are from the same mother or from sibling mothers?
    If this has been broached here before, I apologize, I find SEARCHING on these forums to seldom yield a history of such questions.
    Last edited by clintonslayton76; 16th October 2016, 11:06 AM. Reason: correction

  • #2
    Because the (full?) sisters who are hypothetically the mothers of two different children by the same man will end up with different parts of their parents' DNA, the total of "full matching segments" (use the One-to-one comparison tool with chromosome painting, on GEDmatch, to see this) for the sons of these two sisters will be (significantly?) less than expected for "full siblings". I haven't worked out the numbers, but I'm sure someone will be willing to chime in here.


    • #3
      Thank you

      This answer makes sense in light of the randomly distributed shares from each mother, but I am not sure that it would be detectably less than with non-sibling mothers, on a 1-to-1 for cM shares, esp if the mothers were twins.
      I know that in my own family we have has serial sibling marriages, but no (known) test subjects alive to test the theory.
      Thanks for responding.