Announcement

Collapse
No announcement yet.

Family Finder matching question

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

  • Family Finder matching question

    Hello

    Does anyone know how do the results behave if, let's say, someone is both my father and my mother's sibling?

    Would it state both relationships under the match, or only one, or would that break the testing process and make matches unreliable?

  • #2
    if someone is both my father and my mother's sibling
    This is rather vague, but assuming that you mean your results compared with this someone-sibling:
    there is not a reliability issue, the cM in atDNA shares show a variety of possibilities at various levels below the fairly unmistakeable full parent, but endogamy ("in-breeding') and twins might disallow an accurate TiP or Genetic Distance. The relationships on tables are interpretations of results excluding the situation you describe.

    Since you do not identify the sex of the someone-sibling, I cannot comment on Y or mito for that person with the info provided, and probably would not in any case.

    A most common aberration is "double cousins," so abnormally high shares have to be resolved with knowledge outside of DNA information. The range of estimations of relationships cannot always be accurate. For most people, they are, but, for instance, an aunt and a grandparent share to any one person cannot be resolved without genealogical research.
    Last edited by clintonslayton76; 16 February 2018, 02:29 PM.

    Comment


    • #3
      No, I mean, for example, when my father had kids with his sibling.

      Comment


      • #4
        Originally posted by jötunn View Post
        Hello

        Does anyone know how do the results behave if, let's say, someone is both my father and my mother's sibling?

        Would it state both relationships under the match, or only one, or would that break the testing process and make matches unreliable?
        If someone is sibling to both your mother and father, will your mother and father not be sibling as well then?

        Comment


        • #5
          Originally posted by jötunn View Post
          No, I mean, for example, when my father had kids with his sibling.
          Well that is well within the definition of endogamy. The language estimation cannot go above "parent" for any abnormally sized share, but there are problems below "parent" level for abnormally large shares, look up homozygosity for the key indicator for closely-related parents.

          The father's Y would be passed unchanged from grandmales, the mother's mito would be passed from grandfemales, in any case. The relationship of parents to one another if tested would be a normal sibling relationship in atDNA.

          But the atDNA relationship of sibling-parents to children might be normal (50%) or might show a higher percentage, in this particular case. The language designation above 50% share cannot be anything but parent. Therefore, the relationship of child to grandparent can also go above "normal" share size, but presumably would not indicate as near as parent. But all of this has to be tested in individual cases, because of recombination. Plus, you will not see the issue of more than 50% unless either of the parents have atDNA posted. I cannot answer if sibling shares are abnormally high to one another, but if that is the case, you would have to wade into the literature on that, and I frankly have not found a lot on that subject outside of twins. Most of what I find is not scientific but sensational in nature.

          Testing to determine if parents are related is also possible at GEDMatch, but all of this presupposes that you have access to all atDNA for everyone concerned.

          Phasing would, I assume, be "broken" by these types of results, since it tries to filter out each parent's shares, and this would be problematic, due to homozygosity. If parents are first cousins, children's SNPS should show about 72% homozygous: for sibling parents, there would be so much homozygosity that the offspring is prone to any otherwise recessive gene trait shared by sibling parents that could cause significant medical problems. Another DNA service specialized in marketing themselves as interpreters of medical conditions based on DNA, but I think the FDA put the stoppers on a lot of that.

          I cannot think of any other result that would be "broken," but others might have other theories.
          Last edited by clintonslayton76; 16 February 2018, 04:42 PM. Reason: additional details

          Comment


          • #6
            In regards to endogamy, I have a few examples in my autosomal match list, one of them being to someone who is related to me through 6 different routes in my maternal family tree. Through paper trail we would be fairly distant cousins however the shared segments places us somewhere around 5th to 6th cousins, it seems various intertwined families in a small area in the Netherlands maintained a level of "keeping it in the family" for generations. The algorithm of the system won't be able to exactly determine what the relationship may be when considering endogamous relationships.

            In regards to an incestuous relationship and understanding the genetic relationship, this article seems fairly decent.

            Comment


            • #7
              If referring to a child whos parents were siblings, both would show as a Parent/child relationship to said child.
              They would both show around 3500cM shared DNA with child

              3500cM is half identical amount. We receive roughly 3500cm from mother and the same from father. Amounts are always given in half identical amounts, due to fact it is usually only siblings who would be matching on both their maternal and paternal chromosomes.

              Gedmatch has a tool which one can run DNA through to see if parents are related. You run child kit through and it looks for Runs of Homozygosity (areas that match on ones maternal and paternal chromosomes)

              in case where parents where siblings, I believe it should show Runs of Homozygosity of 1700cM or more...

              Comment


              • #8
                Originally posted by jötunn View Post
                No, I mean, for example, when my father had kids with his sibling.
                The children would be matched the same as the parents do to each other they would all be matched as siblings

                Comment


                • #9
                  More or less what I said in the first place: looking at percentages and language describing the relationships is misleading.

                  Normally parents share 50% (~3500 cM) with each child, but if a child is born from incest between siblings, that child might share more than 50% of their atDNA with each parent, whereas in parent-child incest, the child would share more than 50% cM of their atDNA with one parent/grandparent, and the 25% expected for a grandparent might be skewed to higher cM than usual.

                  But siblings other than identical twins are predictably in their ~2250 cM. As stated in my 2nd response, if your see higher-than-normal share with a non-identical sibling, it might be a milestone for further research.

                  Your question does not establish how many have been tested and if there are results that suggest incest; if this is an ancecdotal relationship being tested with DNA; or if this is a purely theoretical question. The bell curve of these cMs is just that, and there is some speculation (Bettinger) that female atDNA contributions might be more prone to recombination affecting the cM shares. I mention this because I failed to note that the Y and mito from each sibling-parent is also from the same grandparents.

                  In other words, using a sibling relationship ("horizontal") to posit evidence of incest is severely crippled without a test from at least one parent or "vertical" ancestor.

                  My source is the same as spruitean:
                  http://genetics.thetech.org/ask-a-ge...ing-incest-dna
                  Last edited by clintonslayton76; 17 February 2018, 12:03 PM.

                  Comment


                  • #10
                    Originally posted by msc_44 View Post
                    The children would be matched the same as the parents do to each other they would all be matched as siblings
                    So my grandparents would be shown as parents then, but... 2 fathers and 2 mothers?

                    Generally I'm asking about how would the FTDNA system describe the match. Would it just say "parents" or something like "father and uncle"?

                    Comment


                    • #11
                      FTDNA cannot describe an anomalous relationship, period. I do not understand why you are asking a speculative question about results that we cannot see. Because I do not have access to algorithms used by FTDNA to determine if a 1900 CM share is biased toward 1) Full Siblings or 2) Grandparent/grandchild, aunt-or-uncle/niece-or-nephew, half-siblings, wondering about it seems pointless unless such shares have been seen on FTDNA.

                      The important value is the cM share, not any WORDs attempting to categorize a result. There is nothing authoritative about DNA estimations of either origins or relationships, but interpretations of results so far have resulted in various (conflicting) origin theories and tables (ISOGG as an example) where you will see the normal possibilities. A population where endogamy is the cultural norm would no doubt show more vague estimations if it develops its own tables. And IV fertilization will continue to challenge "norms."

                      The highest share can only be for a parent or twin with no accounting for suspected incest. Much of the world's couples are closely related, but not that closely related. If someone is sharing enough with a grandfather to make the gf look like a parent, he will show as a parent even if another male person shows as a parent. The share estimations are not concerned with mimicing a family tree, they are simply translated into WORDS from algorithms.
                      Last edited by clintonslayton76; 17 February 2018, 02:30 PM.

                      Comment


                      • #12
                        Misstatement

                        Because I do not have access to algorithms used by FTDNA to determine if a 1900 CM share is biased toward 1) Full Siblings or 2) Grandparent/grandchild, aunt-or-uncle/niece-or-nephew, half-siblings,
                        This was poorly expressed: 1975 cM will show Full Siblings, Half Siblings, Grandparent/ Grandchild, Aunt/ Niece, but cannot determine which it is. Whereas, 1877 cM will show Half Siblings, Grandparent/ Grandchild, Uncle/ Nephew or Niece. Note that only 100cM loss drops the Full Sibling possiblity from the second match.

                        What I meant to express was: I do not know the "breakpoint" that FTDNA uses to determine one from the other, but there are overlaps, acc to genetic literature, where a clear-cut determination is probablistic but not proveable for one or the other. So throwing in the idea of incest cannot vastly alter the cM share between the siblings who produced children or the children-siblings between themselves, but between parents and children, refer to the links at my previous posting or spruitheans (identical in location but not in form).

                        Run of Homozygosity (ROH) is the issue jotunn needs to look at.

                        Comment


                        • #13
                          So if I understand properly, father´s and mother´s common siblings may be shown as parents too, right?

                          And, for example if my father had an identical twin, it will show two fathers for me too, right?

                          Comment


                          • #14
                            Originally posted by jötunn View Post
                            So if I understand properly, father´s and mother´s common siblings may be shown as parents too, right?
                            Not necessarily, as siblings do not inherit equal amounts of DNA from their parents.

                            Comment


                            • #15
                              Originally posted by jötunn View Post
                              So if I understand properly, father´s and mother´s common siblings may be shown as parents too, right?
                              Example.
                              Code:
                              +--------+      +--------+
                              | Grand  |      | Grand  |
                              | Parents|      | Parents|
                              +---+----+      +----+---+
                                  |                |
                              +---+---+       +----+---+
                              | Father|       | Mother |
                              +---+---+       +----+---+
                                  |                |
                                  +-------+--------+
                                          |
                              +-------+   |   +---------+
                              |  [COLOR="Red"]You[/COLOR]  | --+-- | Sibling |
                              +-------+       +---------+
                              Let say all of you have tested.

                              Now when you login to your account it will tell you the following.

                              GrandParents = Half Siblings, Grandparent/ Grandchild, Uncle/ Nephew
                              Father & Mother = Parent/Child
                              Sibling = Full Siblings

                              Observe there is no gender information.

                              And, for example if my father had an identical twin, it will show two fathers for me too, right?
                              You will not see two father you will have 2*Parent/Child

                              Something for you to understand, FTDNA dosn't make a familytree for you. It will only suggest relationship range. Then you have to make "linked relationship" to distinguish father/mother/child/siblings etc.

                              After you make "linked relationship" FTDNA will attempt to sort out your matches to father/mother branches nothing more (you still have to make your "linked relations").

                              Comment

                              Working...
                              X