Announcement

Collapse
No announcement yet.

67 Marker Results

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

  • 67 Marker Results

    I am wondering if anyone knows what the mutation rates of the top 30 markers are in the 67 marker Y-DNA test. I recently completed mine and am not sure if I understand the results. For those who matched relatively close at 25 and 37 markers, there were no mutations at all in the top 30 markers in the 37 marker test from those previous matches. As an example the number of mutations with one person were,

    1/25, 6/37, 6/67

    The above is typical of relatively close matches to me. Most of the mutations are in the 25-37 region and none in the 38-67 region. Are the mutation rates for the top 30 markers unusually low? If so, what good are they for genealogy?

  • #2
    The mutation rates for the upper 30 markers are indeed low. They are primarily useful for winnowing out 'coincidental convergence.'

    In other words, let's say that at 37 markers, you are a genetic distance of 6 from 3 different men. A GD of 6 is large enough so that some or all of these men may be very distant from you, and appear close only through 'coincidental convergence'. (An easy way to see this is to notice that members of unusual subclades, such as R1b1c6, may actually be only 6 steps away from R1b1c* men at 37 markers.)

    If you and the other 3 men upgrade to 67 markers, it should then become obvious whether you are genuinely related within the past 1000 years or not. Any 'coincidental convergence' at 37 markers will 'blow up' at 67 into a much larger genetic difference.

    In a similar vein, some of the upper 30 markers are useful for identifying particular subclades without having to purchase an additional SNP test from another vendor.

    Comment


    • #3
      Originally posted by lgmayka
      In a similar vein, some of the upper 30 markers are useful for identifying particular subclades without having to purchase an additional SNP test from another vendor.
      What lgmayka refers to here is the 492 marker for R1b's. So far, with a limited sample, it's been found that those who test S21+ (R1b1c9) with a SNP test performed by another company, Ethnoancestry, mostly have 492=13 (over 90%). The general R1b modal for 492 is 12 and 492 seems to be a slowly mutating marker. The rate of 492=13 among a general database of R1b's with over 600 results shows that only about 28% of R1b's have 492=13.

      While it seems to be a very strong indicator of S21+ status, it is not definite, since 10% of known S21+'s have 492=12. Only a SNP test will definitely give you that information. Although Ethnoancestry presently is the only company offering a SNP test for S21, most people expect FTDNA to start offering the test, perhaps later this year, since a research study recently released information about where to find the SNP on the y chromosome.

      Another interesting marker in the upgrade to 67 is 425. It's been found that there is a small percentage of people tested who have a "null 425." This means that some unknown change in the chemistry near that marker has occurred which makes it impossible for FTDNA's primers to detect a value for that marker. This has been found to occur in different haplogroups, although when it occurs in someone in the I1b subclade may indicate Scottish ancestry. I am also investigating a "null 425" cluster in R1b that may be a subclade of R1b1c9 (S21).

      So there are some results for a couple of the markers in the upgrade to 67 markers that have uses if you're in the I1b or R1b haplogroups and want to know about deep ancestry. Also, even if you're not interested in deep ancestry, a "null 425" result can be useful in sorting out things in a surname project. If one participant has a "null 425" and another doesn't, then that indicates some sort of mutation somewhere along the line, a very distinct mutation. Of course, individual markers can be ordered from the "Advanced Orders" section, so you could potentially order 492 and 425 without ordering the entire upgrade to 67 markers.

      Mike Maddi

      Comment


      • #4
        "They are primarily useful for winnowing out 'coincidental convergence.'"

        OK, but why do you need 30 more slow changing markers for this? Why not a mix of fast and slow? The problem is this. Introducing 30 slow changing markers will bias the FTDNATiP results. If, for instance, you are looking at someone who is 7 generations apart compared to 15, and those 30 markers don't mutate in that time frame, the effect will be to bias the calculation. Those people who are 15 generations back will appear to be closer than they really are. As a result, the 67 marker calculation might well be more innaccurate than the 37 maker calculation.

        "In a similar vein, some of the upper 30 markers are useful for identifying particular subclades without having to purchase an additional SNP test from another vendor."

        How do you do this?

        Comment


        • #5
          I checked into my idea of ordering either DYS425 or DYS492, which are useful markers in the upgrade to 67, as single markers. For some reason, neither of those markers can be ordered from FTDNA as a single marker.

          Mike

          Comment


          • #6
            Originally posted by JC399
            "In a similar vein, some of the upper 30 markers are useful for identifying particular subclades without having to purchase an additional SNP test from another vendor."

            How do you do this?
            The answer to your question is in my posting just before yours. However, I noted that relying on a single marker to predict a subclade will be just that, a prediction. And, of course, the prediction may be wrong. Only a SNP test gives you a definite answer as to subclade.

            Mike

            Comment


            • #7
              Originally posted by JC399
              Introducing 30 slow changing markers will bias the FTDNATiP results.
              FTDNATiP has been advertised as taking into account the specific mutation rate of each STR. I'm sure that when the 30 new markers were first released, FTDNATiP probably used some safe guess for them; but at this point, I have to assume that FTDNA has made a reasonable calculation of the new markers' mutation rates and factors those into the FTDNATiP calculation.

              Comment


              • #8
                Originally posted by lgmayka
                FTDNATiP has been advertised as taking into account the specific mutation rate of each STR.
                True enough, but regardless of whatever mutation rate(s) they take into account, if you add 30 markers that don't change in the time frame you are looking at, it's going to add a bias to the result of any time predicting calculation. The most useful data is from those markers that actually change in that time frame.

                Comment


                • #9
                  Originally posted by MMaddi
                  The answer to your question is in my posting just before yours. However, I noted that relying on a single marker to predict a subclade will be just that, a prediction. And, of course, the prediction may be wrong. Only a SNP test gives you a definite answer as to subclade.

                  Mike
                  I was making my post at the same time you were so I didn't have a chance to read what you wrote. I was predicted to be r1b1. Here are my top 30 markers. Can you tell what subclade I am?
                  Attached Files

                  Comment


                  • #10
                    Originally posted by JC399
                    I was making my post at the same time you were so I didn't have a chance to read what you wrote. I was predicted to be r1b1. Here are my top 30 markers. Can you tell what subclade I am?
                    As you yourself pointed out, there's not much mutation or many interesting markers in those 30. I did point out 492 and 425. You have 12 on 492, which means you're probably not S21+. And you don't have a null 425, so there's not much that can be said about that.

                    It's hard to say what subclade, if any, you're in. Most European descent R1b's turn out to be R1b1c, so you're probably that. But then the estimate is that if all R1b1c subclade SNPs, including S21 and S28, are tested, probably 50% of R1b1c's will be negative on all of them. (Either they actually have no SNPs beyond M269, which defines R1b1c, or they have a SNP that hasn't been discovered yet.)

                    Your first 37 markers may tell something. Have you uploaded your results to ysearch.org? It's a lot easier taking a look at a haplotype there, plus you can compare it to other haplotypes easily.

                    Mike

                    Comment


                    • #11
                      Originally posted by MMaddi

                      Your first 37 markers may tell something. Have you uploaded your results to ysearch.org? It's a lot easier taking a look at a haplotype there, plus you can compare it to other haplotypes easily.

                      Mike
                      Yes I have uploaded those to ysearch. Those matches say r1b, r1b1, r1b1c, or unknown.

                      I guess I was trying to figure out what value getting the extra 30 markers was. For me it doesn't seem to have given me anything more I didn't know before.

                      Comment


                      • #12
                        Originally posted by JC399
                        True enough, but regardless of whatever mutation rate(s) they take into account, if you add 30 markers that don't change in the time frame you are looking at, it's going to add a bias to the result of any time predicting calculation. The most useful data is from those markers that actually change in that time frame.
                        Just the opposite. The markers that didn't change are just as important as the markers that did. If two men are, say, 5 steps apart at 37 markers, but then at 67 markers they are still only 5 steps apart, then they are obviously genetically closer than one would have thought merely from the 37 markers.

                        You seem to be assuming that the upper 30 markers are absolutely constant, and hence useless. That is most certainly false. The upper markers do change, albeit at a slower mutation rate; and as long as that slower mutation rate is taken into account, they are quite useful in calculating the Time to Most Recent Common Ancestor (TMRCA).

                        Your reaction, frankly, is very similar to what mine was when the new markers were first released, and I and another I1b1-Dinar got our results. I was horrified that there were hardly any differences between us in the upper markers (perhaps 2). Over time, as I saw that other people were indeed seeing plenty of differences in the upper markers, I realized that I1b1-Dinar (at least north of the Balkans) is younger than I thought, we are closer together than I thought, and our common ancestor occurred later than I thought. This was difficult for me to swallow, especially because it flew in the face of 'conventional wisdom' (which was made up several years ago, on the basis of SNPs alone). But I now realize that the markers are correct, and the 'conventional wisdom' is both obsolete and just plain wrong.
                        Last edited by lgmayka; 8 January 2007, 08:51 PM.

                        Comment


                        • #13
                          JC, compare your ysearch ID to mine, DZ7XA. My 446=16 is very rare, and rare 406s1=9. My 492=12 seems to indicate I'm not S21+ (though I am testing for it).

                          There is some info in the upper 30. We match on quite a few of them though, but not all, which I think is common for people who are R1b1c.

                          Comment


                          • #14
                            Originally posted by Paulie
                            JC, compare your ysearch ID to mine, DZ7XA. My 446=16 is very rare, and rare 406s1=9. My 492=12 seems to indicate I'm not S21+ (though I am testing for it).

                            There is some info in the upper 30. We match on quite a few of them though, but not all, which I think is common for people who are R1b1c.
                            We are pretty distant with 30 mutations in 67 markers.

                            Comment


                            • #15
                              Originally posted by lgmayka
                              Just the opposite. The markers that didn't change are just as important as the markers that did.
                              I disagree. The extreme example of that is where you only compare the last 30 in someone you are related to and they all match even though you may be 18 generations apart. In that case your most probable common ancestor is 0 generations with a wide distribution curve. Not much useful information there.

                              Comment

                              Working...
                              X