Announcement

Collapse
No announcement yet.

False positives?

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

  • False positives?

    I recently did a test for my mother-in-law on Family Tree DNA. She has about a dozen matches who are Finnish with a few more elsewhere in Scandinavia. The matches all share 20Cm and are suggested as fifth to distant cousins. She is Bulgarian and has no other matches outside the Balkans apart from people in the US and elsewhere obvious descended from emigrants. I am wondering if these are all false positives or if there is a real link?

  • #2
    Dr. Ann Turner posted here a few days ago indicating that anything between 15-20 cMs should be a valid match.

    Comment


    • #3
      That's good to know. I wish there was some book DNA for Dummies where all these basic questions are answered and methods to follow suggested. I have seen a few blogs online - one even written by someone who is a match - but they tend to be terribly involved. What is needed is the input not just of someone who knows DNA but also a good editor with lots of experience.

      I read "Seven Daughters of Eve" and it is a great read but is only the story of the discovery of mtDNA.

      Comment


      • #4
        Originally posted by mde View Post
        I recently did a test for my mother-in-law on Family Tree DNA. She has about a dozen matches who are Finnish with a few more elsewhere in Scandinavia. The matches all share 20Cm and are suggested as fifth to distant cousins. She is Bulgarian and has no other matches outside the Balkans apart from people in the US and elsewhere obvious descended from emigrants. I am wondering if these are all false positives or if there is a real link?
        Look at the size of the largest pieces, not just the total amount shared. It depends on the population, but in some instances, where there is a common but very ancient heritage, thousands of years in the past, you can share large amounts of similar DNA but the biggest pieces are relatively small.

        That means there has been lots of time for breaking apart and recombining, but there's not been much mixing with other, genetically less similar, groups. Though in your case you're European, so this seems unlikely.

        Comment


        • #5
          The best preparation for genetic genealogy is that you paid attention during the lesson on meiosis in your high school biology class, or even better, the same lesson at college level, or a beginning genetics course. But we seldom know what we are going to need later in life! Blogs are great, BUT as each blog develops, the blogger will tackle more and more difficult topics, so you have to go back to the early days of the blog to find the introductory remarks.

          If you have enough sense of logic and science to ask about false positives, you're on the right track. For autosomal DNA matching, segment sizes over about 10 cM are very unlikely to be false in any conventional sense, but it's still possible that the relationship is too far back to be proved. It may be useful to remember, also, that European history is full of migrations, as a result of wars, religion, famines, etc. At the same time, autosomal DNA samples are NOT distributed equally from across Europe, so the places where your matches come from may be misleading.

          First order of business in autosomal DNA tests, in my opinion, is to focus on any very close matches, on the order of 100 cM shared DNA or better. Those are the ones where the connection is likely within the past 3 generations, close enough that you and your match can probably figure it out, or at least get close.

          Comment


          • #6
            mde,
            If you're looking for a book, I've heard only good things about Blaine Bettinger's "The Family Tree Guide to DNA Testing and Genetic Genealogy." You can get an idea of the contents by the "Look Inside" feature at Amazon, and I've seen it at Barnes & Noble bookstores, too.

            It's also wise to keep up with the rapid developments by viewing webinars (Ancestry.com and Legacy Family Tree have various free webinars - look for those about DNA), and reading blogs, forums and other groups about genetic genealogy, as you are doing. ISOGG (the International Society of Genetic Genealogy) has a beginners guide, plus many other genetic genealogy topics in their Wiki that you can search to find information. The DNA Newbie Yahoo Group might be another good place for you to check out, too. FTDNA has a glossary of terms in their Learning Center, and ISOGG has their own Genetics Glossary.

            Comment


            • #7
              Originally posted by abuelita View Post
              Look at the size of the largest pieces, not just the total amount shared.
              Should I just ignore the generations back suggestion on Gedmatch? Whenever it finds any match that is not close it suggest give generations.

              Comment


              • #8
                Originally posted by John McCoy View Post
                First order of business in autosomal DNA tests, in my opinion, is to focus on any very close matches, on the order of 100 cM shared DNA or better. Those are the ones where the connection is likely within the past 3 generations, close enough that you and your match can probably figure it out, or at least get close.
                I only have one of those matches myself (I have Irish and British background) across all the companies I have uploaded to. In that case it was someone my mother knew anyway from my grandmother's home town.

                With my Bulgarian in laws it would be a similar situation. We know who people within that range of relation are. What I would ideally like to do is work out what went on further in the past. From a local history published a couple of generations back we have family trees that go well beyond the limited archives in this part of the world.

                I am very surprised how many 20Cm type links are coming up. The Finnish one is just the weirdest. There are quite a few people in America which is odd as there wasn't must emigration from Bulgaria to the US in the ninteenth century. I keep contacting people in the hope that they have worked back to a Bulgarian immigrant but they are always just surprised there is a connection. As with my own DNA the general feeling I get is that people haven't done a lot of research and just don't know those five generations back.

                Comment


                • #9
                  Originally posted by mde View Post

                  Should I just ignore the generations back suggestion on Gedmatch? Whenever it finds any match that is not close it suggest give generations.
                  The gedmatch figures are probably meaningful for most people who are European. But if one isn't...

                  Family member and a skeleton from some 12K or so years ago:
                  total 50.1cM largest piece 10.2 number of generations 4.08 actual number of generations? 500?

                  Same family member and a closer relative, someone with multiple, shared, 5x, 6x, great-grandparents, 18th cent. - triple 7th cousins. The surnames and geography match up. There might have been something more recent but undocumented.
                  total 148.0cM largest piece 26.9 number of generations 3.30

                  Family Finder, which seems to disregard the size of the segments, can make first-third cousins out of people who aren't remotely close. People from different continents, populations that have been isolated for hundreds/thousands of years.

                  This creates a majority of probable false matches, 195/202 or 97% for this particular person - to the extent that we had to turn off matching for all but immediate family because some people actually would get very angry at us when we tried to say "You really aren't a 2nd cousin."
                  How about this one: total 389cM largest piece 7cM
                  or this: 281cM largest piece 6cM
                  or a "2nd-3rd cousin" with total 187cM largest piece 5cM

                  On the other hand I think this is probably "real": total 191cM largest piece 11cM, because the surnames and geography make sense - though he's probably not a "2nd Cousin - 3rd Cousin" as the FF match list suggests, again likely shared ancestors 18th cent. or earlier. Ah - now this one, 199 and 41, certainly worth a letter.
                  Last edited by abuelita; 16th April 2019, 05:06 PM.

                  Comment

                  Working...
                  X