Announcement

Collapse
No announcement yet.

Gedmatch Cousin Prediction? Large Seg Match

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

  • Gedmatch Cousin Prediction? Large Seg Match

    Hi All,

    I uploaded my mothers data to Gedmatch and she is matching with someone with 104cm - The largest matching segment is 91.9cm on Chr 4. SNP 3184. 12cm on Chr 2.

    Do you have any ideas on what Predicted relationship my mother has with this person?

    Does this mean they share a close (ish) common ancestor?

    Any advice would be greatly appreciated

  • #2
    That is a very odd DNA match. With largest matching segment at 91.9 cM you would expect the total to be way higher than 104 cM.

    ISOGG says "Second cousins once removed, half second cousins, first cousin three times removed, half first cousin twice removed "

    Who does your mother have in common with this match?

    Comment


    • #3
      Hi jimbrik,

      Thanks for your reply. Hmm I know I thought this too...

      Personally I think the reason for this may be my mother tested with LivingDNA, because of this I could only upload her raw data to Gedmatch Genesis and not the main match database on Gedmatch.

      The only reason I noticed the match is because my brother who tested with 23andme was matching this person by 85.7cm over 3 segments, those segments being 12cm on chr 2, 44.4cm on chr 4 and another 29.3cm on chr 4 also.
      It shows a break within the segments on chr 4. When comparing them on my mothers LivingDNA kit number on the one to one compare it showed no breaks within the segments in chr 4 making it 91.9cm.

      What was interesting was my brother showed the 12cm match on chr 2 but when comparing my mothers to this person it did not show the chr 2 matching segment but I know it’s there because my brother is… What I have noticed is that the when comparing Genesis kit numbers to the normal Gedmatch kit numbers sometimes it underestimates the match cm amount.

      My mother and brother are also matching someone else at 79cm and this person also matches this new person I have just mentioned on the same chr 4, but not as much.

      I have emailed them but no reply just yet.

      So do you think there must be more segments matches we are not seeing just yet?

      Thanks for the advice

      Comment


      • #4
        Not to get too far off the topic here, but I think this is an increasingly important issue when comparing results from different testing companies and chips:

        To begin, when did your brother test at 23andMe? According to the ISOGG page on 23andMe, these are the dates of the various chip introductions:
        • v1: November 2007
        • v2: September 2008, ~555K SNPs
        • v3: November 2010, >900K SNPs
        • v4: November 2013, ~570K SNPs

        Most recently (August 2017 or so), their chip v5 was introduced. It has about 640K SNPs, according to a post by Debbie Kennett on her blog. From what I've read, both 23andMe v.5, and Ancestry.com v.2 both have customized chips containing more medically-relevant SNPs - at the loss of genealogically-relevant SNPs.

        I ask with which version your brother tested, because this may make some difference. Living DNA says they test 650K SNPs, but we are now getting into the main question: of any of the companies, and their various versions of chips, what is the overlap in common SNPs? In other words, we are now getting into the territory of comparing apples to oranges (or at least, different varieties of apple). This is illustrated quite clearly by the chart on ISOGG's Autosomal SNP comparison chart page.

        Living DNA uses a type of chip, "Global Screening Array" or "GSA," which is what the chip maker, Illumina, is now offering to DNA testing companies. They are no longer offering the previous, Omniexpress chip, which FTDNA is still using (as well as what was used by 23andMe and Ancestry.com before their most recent chip versions). You can read the post at Debbie Kennett's blog, above for details, but here is the pertinent content:
        The Illumina Omniexpress chip, which was previously used by all the genetic genealogy companies, is being phased out, so we are likely to see other companies moving to the GSA in due course. The OmniExpress worked well for European populations but was not so good for other populations. The GSA provides much better global coverage and should improve the results for people with non-European ancestry.

        The GSA is designed for imputation. Imputation is the process of inferring the missing markers in a DNA sequence. This can be done by statistical methods because DNA is passed on in chunks, which means that markers travel together. Researchers can use imputation to reconstruct an entire genome sequence, and it provides a much more cost-efficient way of doing large-scale studies. However, I know of no scientific papers which have looked at the efficacy of imputation for cousin matching. The companies will have a significant challenge ahead in the months to come as they adapt to the new chip and test out their imputation pipelines.

        Because there are so few overlapping markers between the GSA and the OmniExpress this change will also present problems for companies and third-party websites that accept autosomal DNA transfers. A choice will need to be made as to whether to do comparisons using only the overlapping markers or whether to experiment with imputation. GEDmatch is already experimenting with its new Genesis database, which can now accept GSA transfers. It will be interesting to see what solutions are found by other companies.
        See Roberta Estes' post on "Concepts - Imputation," to get an idea of how imputation will affect the consumer DNA genealogy testing industry.

        The matches at GEDmatch (regular database, not Genesis) have tested with the older, Illumina chip. Depending upon which version of chip from 23andMe your brother was tested, he may be more "compatible" with other matches (showing the optimal amount of matching segments). I think 23andMe's v3 chip was the most compatible with FTDNA matches, in that the overlap of SNPs in common was the largest (see the ISOGG chart linked above).

        You may want to have your mother do a Family Finder test at FTDNA, and upload that to GEDmatch, to see if it reveals any more matching segments.

        Just out of curiosity, have you transferred your brother's 23andMe file to FTDNA?

        Comment


        • #5
          Hi KATM,

          Thank you for your post, very interesting.

          My brother tested with 23andme a few months ago, the V4 chip.
          Thank you for the insight regarding compatibility between different testing companies, makes perfect sense and an eyeopener. I am fairly new to this so please bear with me

          Yes I have transferred both my brothers and I raw data from 23andme over to FTDNA. Unfortunately the person from GEDmatch I think was tested with AncestryDNA.

          When I get more funds I think I will get my mother tested with FTDNA or Ancestry as you have suggested.

          What is frustrating is that I believe this match is from an illegitimate line I have been trying to figure out to no avail.

          My great grandmother who was born in 1913 had no father listed on her birth certificate and was illegitimate.
          The other match on GEDmatch I mentioned in my previous post also links via DNA to this new match and my mother. So there must be a common ancestor linking us. This previous match gave me access to his tree on Ancestry a few weeks ago but I couldn’t see a common ancestor although his tree was comprehensive with most lines dating back to the early 1700’s. The only thing of interest by viewing his tree was a surname that keeps on popping up with other matches - The surname Morgan.
          I have contacted the new match but so far have heard nothing back.

          Do you think this new match of my mothers could date from 1913? Or would the DNA shared be a lot higher if a common ancestor was shared from this date?

          Comment


          • #6
            Originally posted by galdo View Post
            Hi All,

            I uploaded my mothers data to Gedmatch and she is matching with someone with 104cm - The largest matching segment is 91.9cm on Chr 4. SNP 3184. 12cm on Chr 2.

            Do you have any ideas on what Predicted relationship my mother has with this person?

            Does this mean they share a close (ish) common ancestor?

            Any advice would be greatly appreciated
            This is something that I have seen a lot of with my Appalachian ancestors. If there is a match that has a ver large segment and few other segments the match tends to be much farther apart than the total shared cm predicts. I would expect this to be at closest 2nd cousin 2x removed, 3rd cousin 1x or 4th cousin. Those with ancestors from a small population pool will find these types of matches more frequently than those with ancestors from large population centers. For example I have a match that has 3 shared segments total of 117cm that largest segment 92cm and the common ancestor puts her at 4th cousin 1x removed. She matches my father at 103cm with 3 segments the largest segment is identical to mine. Do to random recombination of dna I actually have more shared dna in common with her than my father does.

            Comment


            • #7
              Originally posted by rwaldrop View Post
              Those with ancestors from a small population pool will find these types of matches more frequently than those with ancestors from large population centers.
              Hi rwaldrop,

              Makes sense to me regarding small population clusters but I can't imagine that would be the case for my mothers side. Her ancestry is fairly mixed British esp her fathers side being close to large cities.
              But I guess anything is possible

              Comment


              • #8
                Originally posted by galdo View Post
                Do you think this new match of my mothers could date from 1913? Or would the DNA shared be a lot higher if a common ancestor was shared from this date?
                It's hard to say, because DNA matches are so variable. Pinning it down to a specific year is difficult, due to length of generations in some families being longer.

                If we go by a couple of references:
                then the 91.9 cM largest shared segment for your mother and her match seems to fall within possible relationships of second cousins once removed, half second cousins, first cousin three times removed, half first cousin twice removed, to third cousins, or second cousins twice removed. It would be good to use a relationship chart (such as this one by Alice Ramsey) to figure all this out.

                But, there are other relationships which could include 91.9cM within their reported ranges, including a 5th cousin, half GG Aunt or Uncle, and probably more. If your great-grandmother's mother and the unknown father (your 2nd g-grandparents) were both part of an endogamous population, or related otherwise, all these relationship estimates could be misleadingly close.

                Things like where your great-grandmother was born, and who lived nearby (possibly with the surname Morgan?) would help to solve the puzzle. Have you checked the 1910 census for your maternal great-grandmother's mother (your 2nd g-grandmother) and/or her family, to see any possible fathers who may have lived with them, or nearby? Was your maternal 2nd g-grandmother working for anyone at the time of your grandmother's conception, possibly as a servant? You could check that family's census record for clues, too.

                Are there any ethnic groups in your matches other than what you know of your mother's father's side, or of her other known ancestors, which might be from the unknown 2nd g-grandfather?

                Are you sure that "Morgan" does not relate to anyone on your mother's father's side, your maternal grandmother's father's side, or anyone in your 2nd g-grandmother's family? Apparently "Morgan" doesn't appear in your research of your paternal side.
                Last edited by KATM; 29 September 2017, 02:16 PM.

                Comment

                Working...
                X