Announcement

Collapse
No announcement yet.

What is the demarcation point in cM btw predicted 3rd cousin and predicted 2nd cousin

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

  • What is the demarcation point in cM btw predicted 3rd cousin and predicted 2nd cousin

    I know, "it's complicated..." but I'm hoping somebody's identified rough "cut points" (to use testing terminology) between, specifically, predicted 2nd and 3rd cousins - and, more generally, each level of cousinhood.

    My wife just got a new match (I'm so jealous) that is far closer than any either of us has had before on FF:

    "2nd Cousin - 4th Cousin 3rd Cousin 72.51 44.31"

    That is quite a bit larger in both measures (longest and total cM) than the previously-largest shared segment we had, which is:

    "2nd Cousin - 4th Cousin 3rd Cousin 48.32 26.94"

    Both are predicted 3rd cousins. So we're just curious what is the cut-off point for a predicted 2nd cousin? Anyone? Thanks.

  • #2
    Originally posted by dwight View Post
    I know, "it's complicated..." but I'm hoping somebody's identified rough "cut points" (to use testing terminology) between, specifically, predicted 2nd and 3rd cousins - and, more generally, each level of cousinhood.
    You can tell most 2nd cousins by the fact that there are multiple large segments shared vs. a 3rd with about one major segment shared.

    2nd cousins are determined by sum and not longest block. The lowest I have seen is 250cM. There is one I've seen in a rare category called 2nd cousin (2nd - 3rd range); not 1st to 3rd range but 2nd to 3rd. This one has a sum of 150.04cM and a longest block of 30.58cM. Most of the 2nd cousins have a longest block of 45 to 55cM approximately and approximately 250cM - 550cM in the sum. The sum is the primary way they predict 2nd cousins and closer.

    Your example has a large longest block but the fact the sum is low probably makes it 3rd and not 2nd. The best way to tell is to look at the shared segments in the chromosome browser. A 2nd has multiple larger segments, a 3rd has about one large segment (or two that were split in the test by a SNP poor area). I guess its hard to explain but easy to recognize so I'm enclosing examples of verified 2nd and 3rd maps. You will immediately see the difference. These are filtered at 5cM.



    MD
    Attached Files
    Last edited by mkdexter; 14th July 2011, 12:06 AM.

    Comment


    • #3
      Originally posted by mkdexter View Post
      You can tell most 2nd cousins by the fact that there are multiple large segments shared vs. a 3rd with about one major segment shared.

      2nd cousins are determined by sum and not longest block. The lowest I have seen is 250cM. There is one I've seen in a rare category called 2nd cousin (2nd - 3rd range); not 1st to 3rd range but 2nd to 3rd. This one has a sum of 150.04cM and a longest block of 30.58cM. Most of the 2nd cousins have a longest block of 45 to 55cM approximately and approximately 250cM - 550cM in the sum. The sum is the primary way they predict 2nd cousins and closer.

      Your example has a large longest block but the fact the sum is low probably makes it 3rd and not 2nd. The best way to tell is to look at the shared segments in the chromosome browser. A 2nd has multiple larger segments, a 3rd has about one large segment (or two that were split in the test by a SNP poor area). I guess its hard to explain but easy to recognize so I'm enclosing examples of verified 2nd and 3rd maps. You will immediately see the difference. These are filtered at 5cM.



      MD
      I have a known 1st cousin 2x removed and our shared/longest cM is 394.14cM/40.17cM. FF gives us a range of 1st-3rd and suggested 2nd.

      I also have some known 3rd and 4th cousins 2x removed that are coming up as suggested 3rd cousins (2nd-4th range) with shared cM in 60s and longest blocks in the 30s. And one 2nd cousin 2x removed with shared/longest cM of 82.58cM/36.12cM.

      I've attached a picture of my 5cM+ with my 1st cousin 2R (orange), 2nd cousin 2R (blue), and 3rd cousin 2R (green).

      I think what makes it a bit off in my case is that I'm 2 generations "below" the majority of my matches so in some cases I share less DNA and the suggested relationship is higher and in some cases large segments of DNA have been preserved over 2 more generations than expected and they are over-estimating the proximity of our relationship.

      But the picture is a good visual representation of 1st, 2nd, and 3rd cousins (1st and 3rd are both related from the same line).
      Attached Files

      Comment


      • #4
        Originally posted by mkdexter View Post
        You can tell most 2nd cousins by the fact that there are multiple large segments shared vs. a 3rd with about one major segment shared.
        MD
        I've been collecting some 3rd cousin data for my next 'Satiable Curiosity column for the Journal of Genetic Genealogy. Based on the rather small dataset so far, there seems to be a different distribution of segment sizes for known vs predicted 3rd cousins. The known cousins typically have more than one major segment (> 5 cM).

        I'll be collecting data for the next week or so (then again after the article is published) for 3rd cousin matches where there is also data available for the parent. If you'd like to have your data included, write to me off-list [email protected] and I'll send you instructions on how to submit it. Your data will be merged with others, and the identity of you and your matches will not be published.

        Comment


        • #5
          Ann I'll see what I can do when I get back. I left town for the weekend and this is the last of my Internet access till next week .

          Comment


          • #6
            Thanks to Matt and everyone for these responses. Obviously it is a huge jump to 2nd cousin - interesting how that works. My wife's match looks most like Matt's lower right image: Only 1 segment > 5cM and it takes up a huge chunk of chromosome 7. 14 segments altogether, but only 2 of the rest are even > 3cM!

            Originally posted by mkdexter View Post
            You can tell most 2nd cousins by the fact that there are multiple large segments shared vs. a 3rd with about one major segment shared.



            MD

            Comment

            Working...
            X