Announcement

Collapse
No announcement yet.

FF Matrix

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

  • FF Matrix

    I have been with FTDNA for a few years now and unfortunately am still unable to discern my mother's maternal matches from her paternal matches.

    There are a number of reasons for this.

    Firstly, I have no one on my maternal side still alive to test and compare.

    Secondly, my family tree on my maternal side is limited to 3 or 4 generations back in time. Most of my mother's matches seem to fall beyond that range. Her closest matches are 2nd to 4th, but she only has a few of those. Maybe a few more of 3rd to 5th... No match surnames seem to jump out as recognizable and those matches who have surname lists don't have surnames similar to any of mine. All their trees seem virtually unrecognizable to me with few correlations in either surname and even geography.

    Third, my mother's paternal side is unknown as she did not know who her father was.

    I am now in the process of using the FF Matrix to try to tease out relationships among and between each of her matches. I hope I am correct in assuming that the FF Matrix can tell me who is related to who in her match list? That is it's purpose, yes?

    Once I have found all the related "match clusters", at least from there I can assume they might naturally fall into two separate groups that will either be my mother's maternal cluster or paternal cluster.

    Is that a correct assumption to make?

    Are there any other tools I am unaware of that can help me tease out which matches are her paternal or maternal matches, in conjunction with the FF Matrix?

    (also, it does seem odd to me that the FF Matrix has a limit of 10 matches that it can handle at any given time. Because this means I have to spend many hours tediously going through each name and possible combination to see how each match is related to each other match - and when your comparing hundreds of other matches, it becomes a fairly daunting task. Is there an easier way?)

    Any help or advice is appreciated. Thanks.

  • #2
    I have an easy question, hopefully. Can anyone tell me what the function of the Family Finder Matrix tool is? What can it tell me and what are its limitations? I can't seem to find any information on it.

    Any help is appreciated, thanks.

    SMD

    Comment


    • #3
      I use the Matrix as a collating and inquiry tool. Here is an example, I matched Glen. I ran the list of matches in common with Glen. I put them into the matrix and a pattern emerged - there were folks who matched Glen, me and each other. I scooted them up to the top of the matrix, making a almost solid block. I dropped the outliers and found I had 8 in that solid block. I looked at that 8 in chromosome browser and found they all shared one long segment on the same chromosome, sort of a bar code.

      I focused on these folks, looking at their surnames and trees and emailing them. I found two were mother and daughter. So I relate through the mother. No more insight there to date. I got a surname list from Glen. I noticed my great great grandmother's surname. I tried emailing Glen and got nowhere (a year later he did contact me after he had his tree is great order) and went to the web looking for his line. I pieced it together and found a possible common ancestor - the immigrant first name fit a legend and a distant cousin's first name.

      Now that I knew the possible surname, I emailed all the matches in that matrix group and one, Chuck, said "Oh, I descend from that immigrant guy." Now we are working on the connecting pieces. Later a first cousin once removed tested and matched Glen and we share my great great grandmother as a common ancestor.

      Other times what I notice from the clumps in the matrix is common geography - Northern Ireland, Dutchess Co. NY, etc.

      Twice I have found out how two of my matrix matches match each other and put them together.

      Comment


      • #4
        Thanks for that JohnG.

        So I would be correct in assuming then, given my mother's circumstances as mentioned in my first post, that one way to distinguish paternal matches from maternal matches using the matrix would be to compare all matches via the matrix and see who matches each other?

        Theoretically, I should end up with two cluster groups of matches. These may not tell me which cluster is the paternal cluster and which the maternal cluster specifically - but it will show me two distinct clusters at least, yes? Cluster 'A' and 'B' will not share matches (hopefully).

        I mean, if Cluster 'A' turns out to be the maternal cluster, because of say a surname found in common, then it will be safe to assume Cluster 'B' is the paternal cluster (i.e. matches deriving from my mother's father's side).

        We know that my mother's father and mother were British. Unless there are matches that share both paternal and maternal sides, generally speaking two match clusters should emerge using the matrix on everyone in my mother's matches, am I right?

        Comment


        • #5
          I don't think so. I think you will get a lot of clusters on different chromosomes and you won't know which side any given cluster is on. I assume you can't test either parent?

          I would try this tool http://www.dnagedcom.com/ Some say the page is covered with ads, but with Firefox and adblock, I don't see a one. You will see clusters on each chromosome, but it can show if some of the people in a given cluster are matching one of your parents and some matching the other. It just can't tell you which parent.

          I think the matrix tool just shows you what matches are on each others list of matches. Nothing else as far as I know. I don't know where the experts who used to answer questions on this forum are these days. I don't really feel qualified to answer questions!

          Comment


          • #6
            Originally posted by SMD View Post
            Thanks for that JohnG.

            So I would be correct in assuming then, ....
            Unfortunately, no. It is more complicated and quirky because each birth is a roll of the dice, tossing genes around. It would work like that, more or less, if all the matches were 1st or 2nd cousins, maybe 3rd too. But even then there are quirks. My father's one surviving 1st cousin matches everyone they should on my side of the family, and that makes a strong case that matches in common with her are through my maternal grandmother's parents. But then my wife matches her too! They have a common ancestor back in New England in the 1700s. So our son matches Dad's cousin two ways. This does yield a matrix a little like what you suggest in that the matrix based on my son and Dad's cousin's matches in common has one solid block that is me and my kin and a small outlier block that is my wife and her first cousin. That would suggest to my son that he relates to that person in two ways. When he found that my wife and her cousin share a set of grandparents, he knows part of the story of how he relates. I have reason to suspect that I also relate to Dad's cousin in two ways, via Dad and via Mom.

            If you start with two siblings, they have about 50% of the same DNA from their two parents. Their children, first cousins, have about a 4th of that 50% in common, or 12.5 %. All this is average, might be more, might be less. Depends on the dice rolls.

            Anyway, as you go down the generations the odds that two cousins of yours relate to you with the same long DNA segments goes down. There is no one 'bar code' for all descendants of a couple, some have one set of segments, some have another. I may match my 5th cousin with one 10 cM segment and my brother matches the same 5th cousin by another 11 cM segment. So the matrix and long shared segments orient you and give you clues, but not often absolute answers. It is both fun and frustrating!

            Comment


            • #7
              Originally posted by MoberlyDrake View Post
              I don't think so. I think you will get a lot of clusters on different chromosomes and you won't know which side any given cluster is on. I assume you can't test either parent?
              Okay, but when you enter two matches into the matrix, it either shows a check mark, or it does not. That suggests to me that either they are related or they are not.

              If a 'cluster' of 7 matches all seem to have check marks indicating they are related somehow to each other, that would tell me that those 7 matches belong either to the paternal OR the maternal side of the person whose matches they belong.

              I'll try to illustrate further...

              My mother has matches A, B, C, D, E, F, G, H, I, J.

              After using the matrix she finds that A, B, C, D, and J are related to each other and her.

              She also finds that E, F, G, H, and I are related to each other and her.

              However, both groups are NOT related to each other. That suggests that one must belong to her maternal matches and the other must belong to her paternal matches.

              It does not tell her which one is which.

              For instance, if a group of 7 matches all have check marks for each other, it could NOT be the case that 3 are maternal and 4 are paternal. They would all either be maternal OR paternal matches, no?

              UNLESS... 1 or more of the group of 7 are related to BOTH sides, maternal and paternal, of the person whom they are matches of.
              Last edited by SMD; 9 April 2014, 06:27 PM. Reason: correction

              Comment


              • #8
                Originally posted by JohnG View Post
                Unfortunately, no. It is more complicated and quirky because each birth is a roll of the dice, tossing genes around. It would work like that, more or less, if all the matches were 1st or 2nd cousins, maybe 3rd too. But even then there are quirks. My father's one surviving 1st cousin matches everyone they should on my side of the family, and that makes a strong case that matches in common with her are through my maternal grandmother's parents. But then my wife matches her too! They have a common ancestor back in New England in the 1700s. So our son matches Dad's cousin two ways. This does yield a matrix a little like what you suggest in that the matrix based on my son and Dad's cousin's matches in common has one solid block that is me and my kin and a small outlier block that is my wife and her first cousin. That would suggest to my son that he relates to that person in two ways. When he found that my wife and her cousin share a set of grandparents, he knows part of the story of how he relates. I have reason to suspect that I also relate to Dad's cousin in two ways, via Dad and via Mom.

                If you start with two siblings, they have about 50% of the same DNA from their two parents. Their children, first cousins, have about a 4th of that 50% in common, or 12.5 %. All this is average, might be more, might be less. Depends on the dice rolls.

                Anyway, as you go down the generations the odds that two cousins of yours relate to you with the same long DNA segments goes down. There is no one 'bar code' for all descendants of a couple, some have one set of segments, some have another. I may match my 5th cousin with one 10 cM segment and my brother matches the same 5th cousin by another 11 cM segment. So the matrix and long shared segments orient you and give you clues, but not often absolute answers. It is both fun and frustrating!
                Hi JohnG,

                It sounds as if your family tree has some recent ancestors that overlap between the paternal and maternal side. I understand that is a possibility in some families. If 4th and 5th cousins from either side marry or what have you.

                I understand that the autosomal test though only goes as far back as maybe 5 or 6 generations at the most. If anyone in those generations don't intermarry and have children, I don't see why a Swedish 3rd cousin of mine from my father's side should have any connection whatsoever with a British 4th cousin on my mother's side. They would not show up as related in the matrix. At least, I can't fathom how they could since they were born in separate countries with completely different ancestors.

                Comment


                • #9
                  Actualy Family Finder autosomal test can go 20 generations back but Family Finder program will only give relationships suggestions to the 5th Cousin range any farther then that you have to guess the range.

                  Originally posted by SMD View Post
                  Hi JohnG,

                  It sounds as if your family tree has some recent ancestors that overlap between the paternal and maternal side. I understand that is a possibility in some families. If 4th and 5th cousins from either side marry or what have you.

                  I understand that the autosomal test though only goes as far back as maybe 5 or 6 generations at the most. If anyone in those generations don't intermarry and have children, I don't see why a Swedish 3rd cousin of mine from my father's side should have any connection whatsoever with a British 4th cousin on my mother's side. They would not show up as related in the matrix. At least, I can't fathom how they could since they were born in separate countries with completely different ancestors.

                  Comment


                  • #10
                    Originally posted by SMD View Post
                    Okay, but when you enter two matches into the matrix, it either shows a check mark, or it does not. That suggests to me that either they are related or they are not.

                    If a 'cluster' of 7 matches all seem to have check marks indicating they are related somehow to each other, that would tell me that those 7 matches belong either to the paternal OR the maternal side of the person whose matches they belong.

                    I'll try to illustrate further...

                    My mother has matches A, B, C, D, E, F, G, H, I, J.

                    After using the matrix she finds that A, B, C, D, and J are related to each other and her.

                    She also finds that E, F, G, H, and I are related to each other and her.

                    However, both groups are NOT related to each other. That suggests that one must belong to her maternal matches and the other must belong to her paternal matches.

                    It does not tell her which one is which.

                    For instance, if a group of 7 matches all have check marks for each other, it could NOT be the case that 3 are maternal and 4 are paternal. They would all either be maternal OR paternal matches, no?

                    UNLESS... 1 or more of the group of 7 are related to BOTH sides, maternal and paternal, of the person whom they are matches of.
                    I think everyone has to work with the match data and learn what it says to them. The check mark means that person matches both the person whose kit is being looked at and the other person in the matrix. They could relate the same way or in different ways.

                    If you have those two clusters, they are certainly worth looking at. They could be maternal and paternal, or maternal father and maternal mother, etc. But I think you pick a theory and then work with it and be alert for new information.

                    And lots depends on who tested. I have some clusters that seem to be due to the large number of immigrants from a given place or at a given time - Northern Ireland in the period after the American Revolution, the Great Migration to New England in the 1620s and 1630s, ...

                    Most of the new information I have gained is from matches further out than 4th cousin, some are 8th or further. Part of that is I have pretty good trees out to great great grandparents or great great great grandparents, or further. So further out is where I am fishing.

                    Comment


                    • #11
                      I guess I'm operating under the assumption that ones father's ancestors and ones mother's ancestors - generally, for a majority families - will not have overlap; that is, an ancestor who can be traced back to being both a father's ancestor and a mother's ancestor. Perhaps I am underestimating the number of families with that kind of overlap; i.e. distant cousin marriages.

                      My only goal right now is to distinguish between my mother's father's ancestors and my mother's mother's ancestors. Even if I don't know which group is which, I will at least have two fairly large and distinct groups/clusters of matches related to each other, yet not to the other group... that I can safely say belong either to my mother's ancestral lines OR my father's ancestral lines. Then once I get a surname clue that matches an individual in one group or the other, I will then see which group is which.

                      I appreciate all your responses to help clarify. Thanks.

                      Comment


                      • #12
                        My parents were 3rd cousins. I am lucky to have a 2 cousin match on my mothers side back o the shared ancestry. I have found matches coming down my fathers side that are "not in common" with the 2nd cousin on my mothers side. So I ended up getting different dna segments through each parent back to their shared ancestry. I also have lots of matches on my fathers side that are "in common" to my 2nd cousin on mothers side due to intermarriage of various families.

                        A lot depends on geographic location and how long families stayed in the same area or migrated together . More generations in the same location the more likely a couple are related.

                        As I have quite a few matches I have my match chromosome data in a spreadsheet. I have a color code for different lines of my family tree and I use a couple of columns for notes. In my case matches being "in common" is most significant if they share the same dna segment.

                        Comment


                        • #13
                          Originally posted by a shepherd View Post
                          My parents were 3rd cousins. I am lucky to have a 2 cousin match on my mothers side back o the shared ancestry. I have found matches coming down my fathers side that are "not in common" with the 2nd cousin on my mothers side. So I ended up getting different dna segments through each parent back to their shared ancestry. I also have lots of matches on my fathers side that are "in common" to my 2nd cousin on mothers side due to intermarriage of various families.

                          A lot depends on geographic location and how long families stayed in the same area or migrated together . More generations in the same location the more likely a couple are related.

                          As I have quite a few matches I have my match chromosome data in a spreadsheet. I have a color code for different lines of my family tree and I use a couple of columns for notes. In my case matches being "in common" is most significant if they share the same dna segment.
                          Hi a shepherd,

                          Yes, I can see where my theory falls short with a family like yours - where your parents are 3rd cousins. I wonder how common such occurances are.

                          I should probably expect the likelihood of a close cousin marriage in my mother's tree also, since it is presumed that her father and mother were both British. Being an island nation with a relatively large population, it stands to reason that there would be more inbreeding over a shorter span of generations. I suspect my mother's father's ancestry though came from a distinctly different part of Britain than my mother's mother's ancestors. Nonetheless, there may still be 'crossover' the further back in time I explore.

                          My main concern though is with the most recent 5 or 6 generations. Approximately 1st to 4th cousin range. Within that scope there should be very little likelihood of 'crossover' between her father's and mother's side.

                          Comment

                          Working...
                          X