Announcement

Collapse
No announcement yet.

How to interpret X results in Gedmatch?

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

  • How to interpret X results in Gedmatch?

    Please clarify what - if anything -- could be inferred by comparing A and B in the following cases:

    a) zero X match, A, B - males
    b) zero X match, A male, B- female
    c) zero X match, A,B - females
    d) significant (>? cM), A, B - males
    e) significant, A male, B female
    d) significant, A,B - females

    Thanks

  • #2
    Originally posted by goldschlager View Post
    Please clarify what - if anything -- could be inferred by comparing A and B in the following cases:

    a) zero X match, A, B - males
    b) zero X match, A male, B- female
    c) zero X match, A,B - females
    d) significant (>? cM), A, B - males
    e) significant, A male, B female
    d) significant, A,B - females

    Thanks
    For a),b),c)
    If two people don't have an X match, it doesn't mean there is no X pattern of inheritance, most of the time it just means that you can't tell for sure.

    Sisters or half-sisters who share the entire X usually share the same father because he only has one X to give each daughter. So you may be able to identify how half-siblings are related as long as the % shared is about right. A male who is a half-sibling won't share an X with a sister if they only share the father.

    For d),e),f)
    If two people have an X pattern of inheritance, you should follow the pedigree charts to see if you can find a matching ancestor in common along the X lines.

    Now why we like to know the sex of the individuals:
    See the following
    http://www.isogg.org/wiki/Family_Fin...elative_Finder . To be considered identical by descent (IBD) for the X chromosome at another company (23andMe) that uses the Illumina chip, the thresholds are/were (for version 2): to be considered a match there, at least one segment should equal or exceed 7 centimorgans (cMs). After that we have been given the following information regarding what you need for the half IBD segments:

    Autosomal: 700 SNPs, 5 cM
    X (male vs male): 200 SNPs, 1 cM
    X (male vs female): 600 SNPs, 6 cM
    X (female vs female): 1200 SNPs, 6 cM

    When you compare your X matches at GedMatch I noticed that you will see many matches when two females are compared but to get just as many males to match, you need to lower the thresholds considerably. That is because the males only have one X to compare. Males are already phased (sequence known) because the Y is not being compared and it is much harder to get a false-positive match. The females have two Xs. It makes a big difference because females need more SNPs and a longer segment to consider the segment a real match.

    I really don't think any large scale studies have been done using real X chromosome matches in distant cousins. I suspect that the 23andMe lab is either guessing or has proprietary algorithms for these thresholds that I can't explain.
    Kathy

    Comment


    • #3
      Originally posted by Kathy Johnston View Post
      For a),b),c)
      If two people don't have an X match, it doesn't mean there is no X pattern of inheritance, most of the time it just means that you can't tell for sure.
      Kathy
      I don't understand how this is entirely true? I agree with this when the situation is one where A is female and B is male and there is no match of X. But if both A and B are males, I think a non-match can be of value. Let's break it down further.

      We have two related males; possible relation is via:

      1 - mother of A and mother of B
      2 - mother of A and father of B
      3 - father of A and mother of B
      4 - father of A and father of B

      It seems that not matching on X rules out relationship 1. Is this not correct? This could be incredibly valuable information depending on what paper trail you have.

      For example, let's say we know B is related to A through A's mother; but we don't know if it's through B's mother or father. Then can't we use a non-match on the X to say that A is related to B through B's father (Relationship 1 is eliminated and 2 is confirmed)? This could be of tremendous value in narrowing down where to look.

      I'm learning so please elaborate if I am incorrect.

      Comment


      • #4
        Originally posted by Rena View Post
        I don't understand how this is entirely true? I agree with this when the situation is one where A is female and B is male and there is no match of X. But if both A and B are males, I think a non-match can be of value. Let's break it down further.

        We have two related males; possible relation is via:

        1 - mother of A and mother of B
        2 - mother of A and father of B
        3 - father of A and mother of B
        4 - father of A and father of B

        It seems that not matching on X rules out relationship 1. Is this not correct? This could be incredibly valuable information depending on what paper trail you have.

        I'm learning so please elaborate if I am incorrect.
        No that would not be correct. The fact someone does not match another relative simply means they did not inherit the same segment in the same place from an ancestor. They might be related on a different chromosome or related to a sibling or parent of the person being compared.

        Now of course males do not inherit X from their father but, they do from their mother and DNA recombination mother to son can easily make two brothers not match each other. It would be rare they don't match anywhere, but definitely possible.

        Now the opposite, a match, is important as it proves a link depending on the quantity, size, etc. of the segments that match (including the X).

        I'm not sure I need to elaborate further but I can if you need me to, or someone else can as well.

        Matt.

        Comment


        • #5
          Originally posted by Rena View Post
          I don't understand how this is entirely true? I agree with this when the situation is one where A is female and B is male and there is no match of X. But if both A and B are males, I think a non-match can be of value. Let's break it down further.

          We have two related males; possible relation is via:

          1 - mother of A and mother of B
          2 - mother of A and father of B
          3 - father of A and mother of B
          4 - father of A and father of B

          It seems that not matching on X rules out relationship 1. Is this not correct? This could be incredibly valuable information depending on what paper trail you have.

          For example, let's say we know B is related to A through A's mother; but we don't know if it's through B's mother or father. Then can't we use a non-match on the X to say that A is related to B through B's father (Relationship 1 is eliminated and 2 is confirmed)? This could be of tremendous value in narrowing down where to look.

          I'm learning so please elaborate if I am incorrect.
          We have seen a few uncommon reports of siblings inheriting opposite X chromosomes from their mother with no recombination visible in either one. Therefore each got an intact X matching only one maternal grandparent. Usually there will be recombination, but not always. Some chromatids that become chromosomes remain "parental" types and not recombinant types. Now whether these happen to crossover only at the tips remains to be determined. The same thing can happen with the homologous autosomes but it is a rare event because usually one of the parents will show a crossover on that particular chromosome.
          Kathy

          Comment


          • #6
            Originally posted by Kathy Johnston View Post
            We have seen a few uncommon reports of siblings inheriting opposite X chromosomes from their mother with no recombination visible in either one. Therefore each got an intact X matching only one maternal grandparent. Usually there will be recombination, but not always. Some chromatids that become chromosomes remain "parental" types and not recombinant types. Now whether these happen to crossover only at the tips remains to be determined. The same thing can happen with the homologous autosomes but it is a rare event because usually one of the parents will show a crossover on that particular chromosome.
            Kathy
            I can somewhat vouch for this, not opposite x's but the same x uncombined. Two(sisters) out of six tested siblings (3 females, 3 males) received the same uncombined X from my mother (at least based on tested positions),which has to be my mothers maternal X since they show as an X match to my maternal uncle.

            David Pike Utilities shows the following when comparing them together
            Chr X has a Double Segment Match of length 17690 from position 1370495 to position 154551755 (153.18 Mb) (3 non-matching SNPs treated as matching)
            Total Mb: 153.18

            Chr X has a Single Segment Match of length 17690 from position 1370495 to position 154551755 (153.18 Mb) (1 non-matching SNPs treated as matching)
            Total Mb: 153.18

            As well when comparing each of my other siblings(Males) against them in FTDNA chromosome browser, the data shows they match each of us the same way on X.

            Sister A Sibling 1 X 17130740 40416467 37.19 3475
            Sister B Sibling 1 X 17130740 40416467 37.19 3475
            Sister B Sibling 1 X 92307251 140794953 58.31 5750
            Sister A Sibling 1 X 92307251 140794953 58.31 5750
            Sister B Sibling 2 X 1370495 124324753 142.26 13725
            Sister A Sibling 2 X 1370495 124324753 142.26 13725
            Sister B Sibling 3 X 1370495 69689573 95.31 8200
            Sister A Sibling 3 X 1370495 69689573 95.31 8200
            Sister B Sibling 3 X 143979644 154570039 22.89 1742
            Sister A Sibling 3 X 143979644 154570039 22.89 1742
            Sister A Maternal Uncle X 8880078 123309958 120.52 12675
            Sister A Maternal Uncle X 139796052 152335799 32.7 2200
            Sister B Maternal Uncle X 8880078 123309958 120.52 12675
            Sister B Maternal Uncle X 139796052 152335799 32.7 2200

            So I too will say can not assume anything when no match is shown as it all depends on how recombination happened or didn't happen on any given chromosome.

            Comment


            • #7
              Originally posted by Kathy Johnston View Post
              We have seen a few uncommon reports of siblings inheriting opposite X chromosomes from their mother with no recombination visible in either one. Therefore each got an intact X matching only one maternal grandparent. Usually there will be recombination, but not always. Some chromatids that become chromosomes remain "parental" types and not recombinant types. Now whether these happen to crossover only at the tips remains to be determined. The same thing can happen with the homologous autosomes but it is a rare event because usually one of the parents will show a crossover on that particular chromosome.
              Kathy
              We can assume that two sisters who match a father should show an entire X match. We can also assume that a mother should match the entire length of the chromosome with her son or daughter because that is how the chromosome browser works. It cannot distinguish which marker goes on each chromosome of the mother without further information. But I have seen errors, or less likely mutations that can even cause small holes in the parent-child relationship matching. I know a mother who does not match her son on the entire length, but it is almost a full match. So I hope people don't freak out if the parent has a tiny segment that doesn't match. Two errors could occur in nearby SNPs that can create a false negative result. The current chips that all the companies use are not perfect.

              Comment


              • #8
                I have learned the last few weeks that whilst an x match can be useful, it is not always guaranteed even when expected.

                I am a male and have no x match whatsoever with my mother's brother even though we share my maternal grandmother (his mother).

                I do however have a large x match with both my mother's maternal (female) and paternal (male) cousins, both are about 61cMs.

                My maternal uncle doesn't match any of us on x even though he matches us all on most of the other chromosomes. After receiving advice on another thread in this forum i now accept that this can happen sometimes so beware of making assumptions if you don't match on x! I think it is best used positively rather than to disprove something.

                Comment

                Working...
                X