Announcement

Collapse
No announcement yet.

Segment matches

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

  • Segment matches

    Although I tested through FTDNA, my parents and sister tested through 23andMe and I've cousins that tested through other companies. We've all uploaded to gedmatch to compare.
    My mother, sister and I all match on the same segment with a distant German 5th cousin and her son.
    My sister, also matches at 8cm on two different segments with with the son, but not the mother.
    As his father has not been tested, I don't know if my sister matches him or not.
    The father of my sister and I have been tested.
    Both our mother and father show 3cm match with the son of our 5th cousin at or near this segment location.
    The father of the 5th cousins son is from eastern Germany, while the 5th cousin herself is from the lower Rhein near Duesseldorf.
    Our father does have known ancestry from eastern Germany.
    Would this be a combination of the dna of both our parents at this segment to show a stronger match of my sister and the son of our 5th cousin?

  • #2
    A 8 cM segment match (un phased) has a 50% chance of being Identical by Chance IBC or in other words jus a random match
    At 3 cM the odds are above 90% IBC.
    Since you there are parent child pairs at gedmatch.com you can use the Tier 1 phasing tool to phase the child's data into the 2 halves.

    This changes the odds of IBC by a factor or 2 so the 50% point for IBC is about 4 cM
    Using phased data on both sides does NOT give another factor of 2 it only helps a little bit.

    John W

    Comment


    • #3
      John,
      Thanks for the reply..., what does all that you stated mean? Does that mean it's most likely random? If so, is there a reason that caused it? If not random, is this a combination of parents DNA?

      Comment


      • #4
        MFromholt, perhaps you can answer a few questions, or at least ponder them.
        • How did you know the German 5th cousin is a fifth cousin, since GEDmatch does not give such estimates? Was this person a match at 23andMe to your sister, or to you at FTDNA?
        • You don't mention the size of the segment where you, your mother, and your sister all match the distant German 5th cousin on the same segment. Is that matching segment at least 7 cM or higher?
        • Does your sister share with the son on two separate 8 cM segments, or a total of 8 cM over two segments?
        • Does your father match the German 5th cousin, and if so, what are the segment sizes? The 3 cM segment that your parents share with the son may be Identical by Chance (IBC, also known as IBS, Identical by State*) if they do not also match the German 5th cousin, which would probably be with a larger shared segment (over 3 cM).
        • A fifth cousin, if the estimate is true, would have a 4th great-grandparent as a common ancestor to you and your sister. Can you and this match go back to 4th great-grandparents in your ancestry to find the connection? Of course the relationship may be otherwise, so you would need to know other relatives connected to your direct lines (such as siblings to your direct ancestors, and their children), particularly in the suspected German branches.
        The ISOGG page for Identical by Descent is worth reading, and shows a few tables for assessing matches by segment size.

        *Identical by state is defined in the FTDNA Glossary as
        IBS stands for identical by state, meaning the DNA matches by coincidence. When two individuals share numerous individual results without being related, those results are IBS.
        Last edited by KATM; 12 November 2020, 04:34 PM.

        Comment


        • #5
          Hello KATM,
          To answer the main question, the 5th cousin is known previous to DNA testing, with a common ancestor from the German state of Nordrhine-Westphalia. Both her and I as well as several others have been doing genealogical research for 40 years, so in that regard, we know what we're doing.
          My mother shares three segments with the 5th cousin, two on the 10th chromosome with lengths of 11.7cm(snp 1810), 11.0cm(snp 1154). and a 7.0cm(582 snp) on the 17th chromosome.
          My sister and I are pretty much identical on the 11.7cm segment with the 5th cousin while my mother's niece is identical with the 11.0cm segment. Also on 23andMe, a grandson of my mothers first cousin matches with my mother and sister on the 11.7cm segment.(I was tested on FTDNA, rest except for 5th cousin on 23andMe),
          My sister matches the son of the 5th cousin for two 8.0cm segments, one each on the 15th & 16th chromosome.(with SNP's of 331 & 488)
          Our father, using 3cm minimum, does not match at either segment to the 5th cousin, but does match the son of the 5th cousin on the same segment with my sister on the 15th Chr, a 5.1cm segment.
          His matching segment on the 16th chromosome with the son of the 5th cousin and my sister is a 6.4cm match
          My mother does not have a match above 3cm with the son of the 5th cousin, unless I lower the snp's to under 200(In this case, 100) on gedmatch,

          Comment


          • #6
            Ah, that clears things up. So you have determined the 5th cousin relationship using traditional genealogy, and DNA has proven a relationship with your mother, but you're not sure if there is a connection to your father.

            This is a situation for which images of the segment matches might be helpful (with blurred names), but I'll try to keep things straight. Let's see if I understand it so far (it's a bit tough to format these tables, so that's why they look as they do):

            For your family members' segment matches to the 5th cousin, in cM and with (SNP amount):
            Chromosome # Your mother Your father You Your sister Your mother's niece Grandson of your mother's
            1st cousin (at 23andMe)
            10 11.7 cM (1810)
            11.0 cM (1154)
            no match ~11.7 cM seg. ~11.7 cM seg. 11.0 cM seg. 11.7 cM seg. to your
            mother and sister
            17 7.0 cM (582) no match






            For your family members' segment matches to the SON of the 5th cousin:
            Chromosome # Your mother Your father You Your sister Your mother's niece Grandson of your mother's
            1st cousin (at 23andMe)
            15 no match (3 cM
            only w/SNP
            < 200)
            5.1 cM, same seg.
            as your sister
            none 8.0 (331)
            16 no match (3 cM
            only w/SNP
            < 200)
            6.4 cM, same seg.
            as your sister
            none 8.0 (331)







            It's not surprising that bigger matching segments for your mother and the 5th cousin don't survive to the son on chromosomes 10 and 17, considering the distant relationship. And it's interesting that your sister matches the son on two different 8 cM segments on other chromosomes, where your mother does not.

            It's also of note that your sister actually shares more than your father does on those two chromosomes, with the son of the 5th cousin. This is generally a red flag, since children inherit segments passed down from a parent, so the parent should have at least the same sized segment, usually larger.

            At GEDmatch, when you use this son of the 5th cousin and compare your father and sister (and any other relative) to him, do you see any other smaller broken segments for your father, near the start or end points for the 5.1 and 6.4 cM segments? I think that the match with the son, for both your father and sister, are likely NOT evidence enough to say that your father is related. I assume that you compared your father to all chromosomes for the 5th cousin, and he did not match any. So these matching segments on chromosomes 15 and 16 with the son are likely to be false, perhaps due to what is known as "pile-up regions." It's possible that these segments which your father and sister share with the son may be due to the segments being common in the ancestral population, and thus not really proving a relationship to the specific family.

            Do you have any other evidence that your parents may have a common ancestor in Germany back in those generations? I know you wrote that your father has eastern German ancestry, but does your mother?

            Getting back to the previously mentioned ISOGG page for Identical by descent: scroll down to the "Excess IBD sharing" section (pile-up regions) to look at the table showing start and stop positions for the excess IBD regions. It includes one segment for Chromosome 10, and two segments each for Chromosomes 15, 16, and 17 which are known to have such excess sharing; the start and stop points are given, so you could see if the segments for your matches fit within them.

            Roberta Estes discusses pile-up regions in her blog post "Concepts – Identical by…Descent, State, Population and Chance," in the Identical by Population section; see #5 in her list of possible match outcomes, where it says:
            Some DNA testing company runs academic or population based phasing software against your DNA and removes that segment entirely because they’ve decided that particular segment in your results is “too matchy” so it must therefore be “invalid” and population based. This is often referred to as a “pile-up” and means that you have proportionally more matches on that segment than you do on other segments. If your “pile-up” segments are removed in this case, again, you won’t match at all. This is exactly what happened to my Acadian matches when Ancestry implemented their Timber phasing software, which removes pile-ups.
            Last edited by KATM; 13 November 2020, 12:36 PM.

            Comment


            • #7
              First, thanks for the response, I, as well as other family members, greatly appreciate it.
              As far as I know, the most eastern of my mothers line would be central Bavaria(which has not been proven) and the Nordrhine-Westphalia along it's eastern border with Lower Saxony, still quite a ways from where my paternal line came from, so my mother has no known recent ancestry from Eastern Germany.
              My mother does have small segments with low snp's near both segments that my sister matches with my 5th cousins son. But, as you pointed out, and after reading more, it's most likely a "pile-up" match. In both cases, my sister and the son of the 5th cousin, the matches overlap what is mentioned on the link you sent me.

              Comment


              • #8
                Originally posted by MFromholt View Post
                Although I tested through FTDNA, my parents and sister tested through 23andMe and I've cousins that tested through other companies. We've all uploaded to gedmatch to compare.
                Results from almost all companies can be uploaded to FTDNA. The only exception I know of is MyHeritage if it is more than six months since the results came in, Strongly suggest having those who did not test here now upload their results to FTDNA.

                Comment

                Working...
                X