Announcement

Collapse
No announcement yet.

Which is more significant?

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

  • Which is more significant?

    I'm brand new to this, so please forgive me. I'm trying to decipher all my "matches." It looks like I have some "exact" matches (O steps) at 12 markers and 25 markers, but my closest match at 37 markers is "3 steps." Which is the stronger genetic link? (When I look at the percentages, it varies. For the same person -- let's call him Larry -- it shows a 33.57% chance of being related in the last 4 generations at 12 markers; but when I click over to 25 markers, the result goes up to 61.17% chance; and when I look at 37 markers, Larry and I only have a 10.16% chance of being related in the last 4 generations.) Can someone explain why the likelihood rises and then falls with more genetic info?

    Thanks!

  • #2
    More markers tested is the more accurate prediction.

    In a simple sense, think of the gender of people that you meet. You might meet a few women, and then a bunch of men, and so on. But if you just count the first 12 people you meet, it's unlikely that you met 6 men and 6 women. But if you count 37, you're more likely to get something closer to 50/50. And if you count all the people in the world (like looking at all of your DNA), you're going to get the exact answer.

    So, looking at markers is just an approximation, and the more markers you look at, the closer you are to the truth.

    Comment


    • #3
      Thanks, Taed. I understand what you're saying about more markers offering more precision. But with that in mind, why would the percentage of relationship likelihood go UP so dramatically with 25 markers compared to 12 markers? Shouldn't it go DOWN (as it does from 25 markers to 37)?

      Comment


      • #4
        The more markers you look at, the better and the clearer the picture gets. If you have a few "close" 37 marker matches, then take a look at 67 marker matches and see which of those 37 marker matches might disappear.

        But if you are new to this, then try going to a lower level of matches. For instance, on the FTDNA Markers Maps page try looking at 25 marker matches, which will show you your more distant genetic relatives, and that could give you a head start as to where you might refine your search.

        So yes, higher levels mean much, but lower levels can be useful, too.

        Comment


        • #5
          It also depends on what the markers matched are, which is where it gets messy.

          For example, let's say that we're only dealing with two STR markers, M-1 and M-2. The markers are found in the general population in this percentage:
          M-1: 13=90%, 14=9%, 15=1%
          M-2: 20=90%, 21=9%, 22=1%

          So, let's say that your markers are M-1=13 and M-2=20, and you have a match of both markers (2/2 or 0-step) with someone else. That's really not that interesting of a match because something like 90% of the population have those same two markers. So while you match, the likelihood of a close relationship is very low since you share those two markers with 6 billion other people.

          But now let's say that your markers are M-1=15 and M-2=22, both very rare marker values. If you had a 2/2 / 0-step match, that means that the person you're matched with is also in this very small population, less than 1%. (You cannot just multiply percentages since many markers are correlated with each other.) So, since maybe only 50 million other people in the world share those two markers with you, the chance of a close relationship is much higher.

          There's also another issue where some markers are more prone to changing than others. Some might change about about once in 1,000 years, but others might change about once in 100,000 years.

          So, they look at the number of markers tested, how many match, which markers match, how rare the value of that matching marker is, and probably correlations between your different marker values.

          And after all that, matching 12/12 with popular marker values is much more common than matching 33/37 with some rare matches in there.

          Comment


          • #6
            Originally posted by YoyoyB View Post
            I'm brand new to this, so please forgive me. I'm trying to decipher all my "matches." It looks like I have some "exact" matches (O steps) at 12 markers and 25 markers, but my closest match at 37 markers is "3 steps." Which is the stronger genetic link? (When I look at the percentages, it varies. For the same person -- let's call him Larry -- it shows a 33.57% chance of being related in the last 4 generations at 12 markers; but when I click over to 25 markers, the result goes up to 61.17% chance; and when I look at 37 markers, Larry and I only have a 10.16% chance of being related in the last 4 generations.) Can someone explain why the likelihood rises and then falls with more genetic info?

            Thanks!
            TiP calculations are based on what the average mutation rate is for each individual STR marker.
            Given your example, more then likely the marker differences at the 25 marker level have a higher rate of mutation which would allow the probability of closer relationship, but at the 37 marker level the differences happen on a STR marker(s) which have a lower mutation rate, which would place the probability of relationship further back.

            Comment


            • #7
              Originally posted by YoyoyB View Post
              I'm brand new to this, so please forgive me. I'm trying to decipher all my "matches." It looks like I have some "exact" matches (O steps) at 12 markers and 25 markers, but my closest match at 37 markers is "3 steps." Which is the stronger genetic link? (When I look at the percentages, it varies. For the same person -- let's call him Larry -- it shows a 33.57% chance of being related in the last 4 generations at 12 markers; but when I click over to 25 markers, the result goes up to 61.17% chance; and when I look at 37 markers, Larry and I only have a 10.16% chance of being related in the last 4 generations.) Can someone explain why the likelihood rises and then falls with more genetic info?

              Thanks!
              All of these matches are ancient.A 34/37 match could end up at 50/67 if you were both tested at 67 marker level.

              Comment


              • #8
                Originally posted by 1798 View Post
                All of these matches are ancient.A 34/37 match could end up at 50/67 if you were both tested at 67 marker level.
                Or a 34/37 match could end up 64/67. Regardless of which is more likely either scenario is possible.

                Comment


                • #9
                  I am Ancient - NOT!

                  Originally posted by 1798 View Post
                  All of these matches are ancient.A 34/37 match could end up at 50/67 if you were both tested at 67 marker level.
                  WOW!!! My fourth cousin twice removed and I are ANCIENT. And I just thought I was older than dirt.

                  Or put another way, I don't believe all 34/37 matches are ancient.

                  Comment


                  • #10
                    Originally posted by 1798 View Post
                    All of these matches are ancient.A 34/37 match could end up at 50/67 if you were both tested at 67 marker level.
                    The 38-67 panels of markers are well-known to have very slow mutation rates when compared to the 26-37 marker panel.

                    Can you give us the kit numbers for two men who have a 34/37 match, but whose match turns out to be 50/67? That would mean they differ on 16 of the 30 markers in the 38-67 panel.

                    It sounds like you just pulled this idea out of a hat, without any known case to justify it.

                    Comment


                    • #11
                      Originally posted by MMaddi View Post
                      The 38-67 panels of markers are well-known to have very slow mutation rates when compared to the 26-37 marker panel.

                      Can you give us the kit numbers for two men who have a 34/37 match, but whose match turns out to be 50/67? That would mean they differ on 16 of the 30 markers in the 38-67 panel.

                      It sounds like you just pulled this idea out of a hat, without any known case to justify it.
                      Is your first name Thomas?!!!
                      They are on my home page, two at 33/37 and one at 34/37 and I lose them at 67.One of them doesnt belong in the same subgroup,he is P312.

                      Comment


                      • #12
                        Originally posted by 1798 View Post
                        Is your first name Thomas?!!!
                        They are on my home page, two at 33/37 and one at 34/37 and I lose them at 67.One of them doesnt belong in the same subgroup,he is P312.
                        Actually, my middle name is Thomas.

                        And you know for a fact that their distance from you is 50/67 markers? Do you realize that for someone to show up as a match at 67 markers, the distance must be 60/67 or better?

                        There's a big difference between 50/67, as you stated to be possible (even though you match 34/37), and 59/67 which both of these cases might be. So, please tell us specifically what their distance is to you at 67 markers. And provide your and their kit numbers, because I'd like to verify myself that they're 50/67 to you.

                        Comment


                        • #13
                          Originally posted by MMaddi View Post
                          Actually, my middle name is Thomas.

                          And you know for a fact that their distance from you is 50/67 markers? Do you realize that for someone to show up as a match at 67 markers, the distance must be 60/67 or better?

                          There's a big difference between 50/67, as you stated to be possible (even though you match 34/37), and 59/67 which both of these cases might be. So, please tell us specifically what their distance is to you at 67 markers. And provide your and their kit numbers, because I'd like to verify myself that they're 50/67 to you.
                          I had to do a lot of searching but I have found one of the men who has a GD of 12 between the 37 to 67 from me. He is L21 and he is 50/57 and his relative is on my home page and he is the 33/37 match.He used to be at ysearch but he must have deleted his profile.

                          Comment


                          • #14
                            Originally posted by 1798 View Post
                            I had to do a lot of searching but I have found one of the men who has a GD of 12 between the 37 to 67 from me. He is L21 and he is 50/57 and his relative is on my home page and he is the 33/37 match.He used to be at ysearch but he must have deleted his profile.
                            The fact that you had to do a lot of searching to find an example to back up your claim shows why I and one or two others objected to your claim. It's very uncommon, in other words rare, for a match to go from 34/37 to 50/67, as you insisted.
                            Last edited by MMaddi; 3 October 2013, 01:44 PM.

                            Comment


                            • #15
                              Originally posted by 1798 View Post
                              I had to do a lot of searching but I have found one of the men who has a GD of 12 between the 37 to 67 from me. He is L21 and he is 50/57 and his relative is on my home page and he is the 33/37 match.He used to be at ysearch but he must have deleted his profile.
                              The editing time on my previous post ran out, but I wanted to add this. Now you're talking about 50/57, which I've bolded above. So, maybe you were getting confused between 50/57 and 50/67 - big difference.

                              Comment

                              Working...
                              X