Loading ...
Sorry, an error occurred while loading the content.

179A study that links pedigrees & the phylogenetic tree of Hamilton Groups A and B

Expand Messages
  • Bill
    Jan 27, 2012
    • 0 Attachment
      This note is written for readers who are interested in the details of how pedigrees and Y-DNA results can be made to work together to yield vastly more complete information than either approach can give alone. We now have an approach that can compare pedigrees with clusters on the tree. Moreover, both pedigrees and trees now have compatible time scales.

      As part of my study of haplotypes and how they can be used to form dated phylogenetic trees, I recently concentrated on two clusters of Hamiltons, one called Hamilton A and the other called Hamilton B. My Maymore Hamilton line is among those in Hamilton B. The Hamilton B lines connect through both pedigrees and Y-DNA to Scottish royalty around the time of James I of Scotland. You can read about this fascinating pedigree line, complete with its non-paternal event at: http://dl.dropbox.com/u/59120192/HamiltonBHistory.pdf
      and the note (just below) contains the URL that ties the pedigrees to the tree.

      I belong to a Hamilton chat room and I recently sent to their members the following notice about the phylogenetic tree containing Hamilton As and Bs. Here is the note I sent:
      There has been considerable discussion recently about the genetic relationship between the Hamilton A and B groups. Here is a brief discussion of what I have found. First, open the phylogenetic tree located at:
      The tree contains Kit Numbers vertically on the right, along the time line RCC=0 which is about 1945 CE. The RCC time scale is at the bottom of the chart and time increases to the left so that 10 RCC is about 433 years. My application program uses 37 marker haplotypes to place the testee results on the tree, and it optimizes the positions of the testees on the tree. Just to the right of the Kit Number is the Hamilton Group designation. A and B are Hamilton A and B. You will note that the program has separated the A group into subclusters A1 and A2. These subclusters, as well as A and B, are well-separated from each other on the tree. The designation AR shows where the haplotypes of the Group A Robertsons fit on the tree. They are quite mixed among the Hamilton A; some are listed among the A1 sub-clustering and some are among the A2 sub-clustering. The A1 subcluster tends to group near the bottom of the tree, although some appear mixed among the A2 group. While the AR group is scattered, you can see a grouping of 6 or 7 of them among the A1 subcluster at the bottom of the tree.
      Now, what about the time relationships? We now look at the junction points on the tree. Those junction points show where mutations have caused evolutionary branching from the progenitors down to today's testees.
The B group shows little, if any, sub-clustering. But there appear to be groups of B that are probably related by pedigree because they are within an RCC of 5-10. There are about 4-5 of these groups in B that probably tie together with pedigrees. The most recent common ancestor (MRCA) of most of the Bs is at RCC ~ 8, but then other MRCAs show up at RCC ~ 12 (the lower half of the B are tied with the middle B at that point). The MRCA of all the B group is at RCC ~ 22 (an estimated 950 years ago). Then the MRCA of the B group AND the A group is estimated to be at RCC ~ 66 (2850 years ago, or about 2850-1945=900 BC).
Similarly, the MRCA of the A1 and A2 groups is located at about RCC ~ 10 (430 years ago) with the CA of the A2 group at RCC ~ 18-23 and the CA of the A1 group at RCC ~ 10.
In the papers I have written in which pedigrees are compared with the clusters on the phylogenetic tree, there is virtually a one-to-one correspondence between each cluster and the pedigrees of their cluster members. Note, however, that mutations still cause uncertainties in the placements on the tree, particularly when single pairs are compared. But the overall form of the tree should match reality pretty well. Individual uncertainties amount to a standard deviation of RCC=3 or less. Statistically, out of 100 placements one might expect a couple of testee pairings to be off by an RCC of about 9.
      I invite the testees and Gordon Hamilton to comment on these clusters and pedigree groupings. 

      - Bye from Bill Howard (Hamilton B line, with Tom Hamilton, Kit No. 111729, my 2d cousin, once removed, as my tie to this group.
      Gordon Hamilton is the Hamilton surname administrator. You can view his Hamilton website at the following URL:
      This is one of the very best web sites I have seen that deals with this topic. It investigates various Hamilton pedigree lines and haplogroups analyzing the differences in marker values in detail. My investigation of the Hamilton A and B lines treats only those lines. This web site extends the analysis to additional lines in different haplogroups. I commend this site to you. It's an excellent tutorial. Gordon Hamilton responded to my note in the Hamilton Clan chat room. Here is his commentary of my analysis and the Hamilton A and B phylogenetic tree.

      Your analysis of results for Groups A and B certainly indicates that they separated some time back as expected from the differences in the STR values. Your time to a most recent common ancestor for the two groups of about 3000 years is similar to what others have estimated as well.
      Turning specifically to the Group A results first, the two outliers at the beginning (kits 46116 and 9248) are almost certainly due to RecLOH** events in each of their lines. Participant H-123 (46116) is known to be closely related to H-046 (15116), and participant H-014 (9248) is known to be closely related to H-054 (19325). You will note that 15116 and 19325 are far removed from their close relatives in your phylogenetic tree as one would expect if a RecLOH event occurred in the two lines. What this is really telling us is that the results for the two outliers (46116 an 9248) should be ignored. This then gives a tree for Group A which looks reasonable with the most recent common ancestor for this group living about 17 RCC or about 730 years ago. Since we believe that those in Group A are derived from Walter Fitzgilbert who was born about 1270, this number is almost too good to be true. However, it is certainly consistent with our earlier conclusions.
      As indicated on our web site, the separation of Group A into two tables, A1 and A2, is fairly arbitrary; attempts were made to cluster those with similar STR results close to one another. Your finding that the A1 and A2 results separate to some extent is therefore consistent with that.
      Turning to the Group B results, the first three results (14095, 21774, and 18060) are clearly outliers but for these I do not have a good explanation for why this should be the case. I note that in two of these cases (14095 and 21774) the value for one of the STR sites (CDYb) is off by two (35) from the modal (37). Possibly your program does not allow for the possibility that this difference may be due to only one mutation event. The other outlier (18060) is known to be closely related to H-053) which has two less mutations so maybe something unusual happened in the case of 18060. In any event I would tend to ignore the three outliers, in which case the rest of the tree looks reasonable with a most recent common ancestor of about 12 RCC or about 520 years. Again this is not too far off from what is expected if all of Group B, as we suspect, is derived from one man (James1) who lived about 1400.
      As you imply, one should not put too much emphasis on the exact position of each person in your tree but overall the trees look reasonable with the foregoing caveats.
      -- Gordon (H-003, Group B)
      ** A good description of what is meant by the term RecLOH can be found at:
      I would appreciate any comments that our chat room may have on this work and these exchanges. – Bye from Bill Howard
    • Show all 2 messages in this topic