Announcement

Collapse
No announcement yet.

Sisters, same mother, X ?

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

  • Sisters, same mother, X ?

    I have a half sister. We share the same mother. Should we not both share on the X chromosome? We don't show any comparison on that chromosome.

  • #2
    To add to post above. I am female.

    Comment


    • #3
      Originally posted by Tenn4ever View Post
      I have a half sister. We share the same mother. Should we not both share on the X chromosome? We don't show any comparison on that chromosome.
      Occasionally there is no mixing of your mother's two chromosomes so you and your sibling get opposite X chromosomes, i.e. one just happens to get an X coming from the maternal grandfather and the other one gets an X from the maternal grandmother's side of the family.

      Comment


      • #4
        Thank you for that explanation. I have the charts for the X chromosome but it still drives me crazy.

        Comment


        • #5
          I have the same question. But my sibling is male and I am female. So what are the differences in this case?

          Comment


          • #6
            Originally posted by JamieEvan View Post
            I have the same question. But my sibling is male and I am female. So what are the differences in this case?
            The explanation that Kathy Johnston gave to the OP could be the same for you, even though the OP's situation was two sisters and your situation is a brother and a sister.

            Your brother received his only x from your mother. You received one of your x chromosomes from your mother. As Kathy explained, it's possible for a mother to pass an unrecombined x to her child.

            So, it could be that your brother received your maternal grandmother's x from your mother, without recombination, while you received your maternal grandfather's x from your mother, without recombination. Or vice versa. In that case, you and your brother wouldn't share any segments on the x, because they came entirely from different grandparents.

            It seems to be an unlikely possibility, but can happen. If one or both of your maternal grandparents is alive and will test, you can see if you or your brother share the entire x with one of the grandparents.

            Comment


            • #7
              Originally posted by Tenn4ever View Post
              I have a half sister. We share the same mother. Should we not both share on the X chromosome? We don't show any comparison on that chromosome.
              No one has really made a good case for what parameters to use for X-DNA matching. It does not make sense to use 700 SNP's and 7.0 cM like GEDmatch uses for regular autosomal. With regular autosomal the large segment size is used in attempt to get an estimate for generations to MRCA. X-DNA is not good for getting an estimate of generations to MRCA due to its quirky nature. All you are looking for is a connection. Even a small matching segment is proof of a connection. Use 150 SNP's and 1.0 cM minimum segment size and you will see a number of matching segments with your half-sister.

              I know I will hear an explosion over my use of short segments. With parameters of <150, 1.0> I can find pairs of people all day long that have no matching segments. How is that possible if my short segment parameters are said to spew out all kinds of false-positives? When it comes down to it, the chances of matching a string of only one hundred SNP's at random is almost nil. It's time for a major rethink of matching methodology.

              Jack Wyatt

              Comment


              • #8
                Originally posted by georgian1950 View Post
                No one has really made a good case for what parameters to use for X-DNA matching. It does not make sense to use 700 SNP's and 7.0 cM like GEDmatch uses for regular autosomal. With regular autosomal the large segment size is used in attempt to get an estimate for generations to MRCA. X-DNA is not good for getting an estimate of generations to MRCA due to its quirky nature. All you are looking for is a connection. Even a small matching segment is proof of a connection. Use 150 SNP's and 1.0 cM minimum segment size and you will see a number of matching segments with your half-sister.

                I know I will hear an explosion over my use of short segments. With parameters of <150, 1.0> I can find pairs of people all day long that have no matching segments. How is that possible if my short segment parameters are said to spew out all kinds of false-positives? When it comes down to it, the chances of matching a string of only one hundred SNP's at random is almost nil. It's time for a major rethink of matching methodology.

                Jack Wyatt
                Here are some observations to consider when a threshold of 1 cM for the X is used:

                If you are a male, your X is phased and unlikely to have many matches, even at 1 cM. A private study I did in 2014 on 110 women and 141 males in a large FTDNA project demonstrated that 6% of males had no X matches, 50% of males had only 1 to 5 X matches even at 1 cM. As you know, an autosomal match has to be present first before an X match is called.

                The male who had over 100 X matches was of Ashkenazi descent. He was not matching close relatives, but rather had an endogamous group in common. The tiny X matches appeared to be coming from a common genetic pool but these were not much use in identifying the common ancestor.

                I counted 110 women who had ordered Family Finder. Of those females who appeared to be “non‐Ashkenazi”, an average of 21% of their FF matches also had X matches. Of these X matches, the females in their match list outnumbered the males approximately 14 to 1. This demonstrates that having 2 X chromosomes each increases the false positive rate substantially. Phasing (in males) clearly made a huge difference.

                The average non‐Ashkenazi female had 105 X matches out of 492 total matches. I don't know how you can explain this result other than to say it suggests that there is a high false positive rate. The 1-5 cM X segments are unlikely to be coming from the same ancestors (MRCAs) as the autosomal segments. If you set the threshold higher than 5 cM, most of these matches disappeared. One Ashkenazi female had over 1000 X matches by the end of April, 2014 so that is why I had to separate out those with over 30% Jewish Diaspora. Too many people matched the same segments. I suspect even the grandparents matched the same segments but I was not looking at that issue here - I have seen that in other groups.

                In summary:
                1. Males on average had far fewer X matches than women, 1.3% versus 21% of their total FF matches, unless they were from an endogamous Jewish population.

                2. Women found that most of their X matches were with other women. Only an average of 6.6% of their matches were with men. Ashkenazi women could see a much higher proportion of males in their match lists (compared to non-Ashkenazi women) because of endogamy. Those women with over 1000 X matches tended to have a high proportion of Jewish diaspora in their My Origins calculation.

                3. Women had many more X matches but these were disproportionately associated with Identical‐by‐State coincidental matches with other women. I think this is because women have two chromosomes that are not phased. Pseudo‐segment matches appeared to be the cause. Keep in mind, you are matching two pairs of SNPs at each position when you compare two females' Xs.

                Could this be a problem with the matching algorithms on the unphased autosomes too? Probably so.

                4. People from the Jewish Diaspora (and other isolated populations) tend to have significantly more matches in general as well as more X matches because of endogamy. So it is not just a matter of false positives causing the match; pile-up regions can contribute as well in some cases but that does not help you with close cousins or siblings. SNP density is another thing to consider especially in some populations where there is little variation.

                Kathy

                Comment


                • #9
                  georgian1950 and Kathy Johnston

                  Dealing with brother sister phased data on X(my Raw Data is separated into my Grandparents maternal and paternal X chromosome values), I found over 50% of these pseudo segments (under 5cM/500SNP) disappear. Algorithm is picking between females 2 values and matching it up to males single value (females 2 X chromosomes vs males single chromosome)

                  Matching algorithm also allows for a certain number of no calls and mismatching SNPs which results in a higher number of false segments on these smaller segments.

                  This also applies to the autosomal chromosomes 1 thru 22.

                  Yes some of the smaller segments can be valid, but majority in my opinion are population group related more so then common ancestor in the last 200 to 300 years.

                  Comment


                  • #10
                    Originally posted by prairielad View Post
                    georgian1950 and Kathy Johnston

                    Dealing with brother sister phased data on X(my Raw Data is separated into my Grandparents maternal and paternal X chromosome values), I found over 50% of these pseudo segments (under 5cM/500SNP) disappear. Algorithm is picking between females 2 values and matching it up to males single value (females 2 X chromosomes vs males single chromosome)

                    Matching algorithm also allows for a certain number of no calls and mismatching SNPs which results in a higher number of false segments on these smaller segments.

                    This also applies to the autosomal chromosomes 1 thru 22.

                    Yes some of the smaller segments can be valid, but majority in my opinion are population group related more so then common ancestor in the last 200 to 300 years.
                    georgian1950,

                    Your position seems to be that any shared segments of 150 SNPs and 1.0 cM are undoubtedly IBD. We've pointed out to you that none of the commercial testing companies go any lower than 5.0 cM and 700 SNPs in considering IBD segments. We've asked you to provide some paper trail verification that your contrary view is correct and that the testing companies are too conservative.

                    You've not offered any evidence, in the form of a scientific study or a proven paper trail confirmation of your tiny segments being undoubtedly IBD. So read prairielad's post above, especially those parts I bolded.

                    He's telling us that he's actually found that in the case of him and his sister less than half of the tiny segments you tell us are IBD turn out to be IBD, which becomes clear when the phasing of segments on a brother's and sister's x is taken into account. All it takes is one case to disprove your statement that all shared segments of 150 SNPs and 1.0 cM are undoubtedly IDB. prairielad has provided that one case.

                    So, you'll have to revise your statement and tell us what percentage of the tiny shared segments you claim are IBD are actually IBD.

                    Comment


                    • #11
                      One way to prove that a small segment is coincidental is to show that both grandparents (who are mates) carry the same sequence. For example, to map a sequence to a specific lineage, that X sequence cannot be in common with both maternal grandparents in the same place at the same time.
                      Attached Files

                      Comment


                      • #12
                        Originally posted by Kathy Johnston View Post
                        One way to prove that a small segment is coincidental is to show that both grandparents (who are mates) carry the same sequence. For example, to map a sequence to a specific lineage, that X sequence cannot be in common with both maternal grandparents in the same place at the same time.
                        How do you know that the set of grandparents don't have a common ancestor themselves?

                        Comment


                        • #13
                          Originally posted by georgian1950 View Post
                          How do you know that the set of grandparents don't have a common ancestor themselves?
                          They probably do a few thousand years ago. The X lineages do not converge within a genealogical time frame and there are DNA matches confirming the paper trails. The countries of origin are different. I have found this kind of ambiguity of small segments in most families where I have mapped the segments through the grandparents.

                          Comment


                          • #14
                            georgian1950

                            See my following post regarding pseudo/false segments

                            http://forums.familytreedna.com/show...168#post418168

                            Comment

                            Working...
                            X