Sunday, 19 June 2016

The Origins of R1b-GF2 (Part 3) - our nearest neighbours

(Please note: for added security, project members are described by their initials and the last 4 digits of their kit numbers)

We have previously examined the DNA profiles of several new members of group R1b-GF2 and seen how this has resulted in a reassessment of those in the Ungrouped section. This re-evaluation resulted in 15 potential R1b-GF2 candidates being identified among Ungrouped members, 4 of whom succeeded in "passing the test" and are now nestled comfortably in the warm embrace of R1b-GF2. In this article, we look beyond R1b-GF2 to see who are our nearest neighbours and does this tell us anything about the origins of our group.

Our (current) position on the Human Evolutionary Tree

We saw in Part 1 of of this series of articles, that the likely terminal SNP* for all members of R1b-GF2 is FGC20561 - a SNP downstream of Z253, one of the major R1b sub-clades. The SNP progression for R1b-GF2 would be as follows:
  • R- ... M269 ... > L21 > DF13 > ZZ10 > Z253 > S847 > S844 > S856 > S845 > S846 > Z17685 > FGC20561
Where this places us on Alex Williamson's Big Tree, the YFULL version of the haplotree, the FTDNA & ISOGG versions of the haplotree is illustrated in the diagrams below. As we are at the cutting edge of science when it comes to using SNP data to explore our roots, FGC20561 is so new that it does not even appear on Alex's tree (yet). The more upstream SNPs do (i.e. Z17685 and above). Similarly, the YFULL haplotree only includes S845 but nothing further downstream. Both FTDNA & ISOGG versions do include FGC20561, but none of the trees include any SNPs further downstream of FGC20561. In other words, for now it is the terminal SNP of the group ... until further SNP testing (such as the Big Y) reveals a new terminal SNP.

There are some minor differences between the trees because we are still at the experimental stage of putting together the Tree of Mankind - for example, YFULL has the SNPs S844 and S856 as being equivalent SNPs (i.e. more or less at the same position on the haplotree) whereas FTDNA and ISOGG have S856 as a SNP further downstream of S844.

But bit by bit, we will continue to move further downstream along the Human Evolutionary Tree until we identify the SNPs specific to the Farrell's of R1b-GF2, and maybe even SNPs specific to sub-branches within this particular group, or specific to individual R1b-GF2 Farrell's.

R1b-GF2 on Alex Williamson's Big Tree

The YFULL haplotree stops at S845 (formed 1900 years ago)

FTDNA's haplotree - some minor differences in branching structure

FGC20561 and adjacent SNPs on ISOGG's haplotree

Insights from a Haplogroup Project Administrator 

Ray Murta is one of the Admins of the Z253 project. Ray kindly contacted me earlier in the year to share some of his project's findings in relation to the Farrell surname. Ray noticed that there was a close genetic connection between the Farrell's of R1b-GF2 and people with the surnames Kelley and Harrell. 

In brief, Ray put together a family tree for a subgroup of members of his Z253 project based on the SNP results of their recent Z253 SNP Pack tests.  This subgroup is known as the Z253-1716-11 sub-group and consists of over 300 people who have contributed their SNP & STR results* to help the further analysis of the branching structure in the Z253 subclade. Within this Z253-1716-11 sub-group is a further sub-group of about 50 people, which Ray refers to as the Farrell-Harrell-Kelley Cluster.  So we have the Z253 sub-clade (583 members), and below this the Z253-1716-11 subgroup (300 members), and below this the Farrell-Harrell-Kelley Cluster (50 members), and below this our own Farrell group R1b-GF2 (20 members).

Two people within this cluster have taken SNP tests and are positive for FGC20561. One of these people is our very own project member JMR-8902 who is the subject of an earlier blog post describing how Harrell DNA became associated with the Ramey surname.

Below is Ray's version of our portion of the haplotree, showing the line of descent from S856 to FGC20561 (in green) and it's adjacent branch FGC20562. The Big Y testers are in yellow (these are the same people as in Alex Williamson's tree above, namely Yorke & Roderick) but there are additional people included who have not tested on the Big Y - two other Yorke's, Cain, Ramey, Finch, Farrell, Harrell, & Kelly. Because several different surnames share the SNP FGC20561, this suggests that it is not specific to the Farrell surname of R1b-GF2. Currently, the three members of R1b-GF2 who have undertaken SNP testing have done either a single SNP test (1 person) or the Z253 SNP Pack (2 people). We would need two or more people to undertake the Big Y test in order to determine downstream SNPs that may be specific for the Farrell surname of R1b-GF2.

The surnames associated with the adjacent FGC20562 branch are Christie, McKenna, & Hazleton (Big Y testers) as well as McReynolds, Reynolds, Bankstown, Robbins & Gilchrist. There is a third branch, currently defined by FGC20560 (aka Z17685) with surnames Finlay & Robbins.

Interestingly, Ray dates FGC20561 & FGC20562 to about 1100 AD, and the slightly upstream SNP FGC20560 to 1000 AD.

Ray Murta's version of the haplotree showing the FGC20561 branch (in green)
(click to enlarge)

Ray’s analysis has linked several surname projects together, including ours. The other Surname Projects include the Harrell Project and the Kelley DNA Project. The particular Harrell group is Harrell Line 2 with roots in Augusta county, Virginia, US (Roberta Estes, renowned blogger, is one of the Admins and the project results can be viewed here). And the particular Kelley group is their group 19A (results here), again with roots in the US.

Farrell & Harrell

You will immediately be struck by the similarity of the surnames Farrell and Harrell. And this similarity begs the question: is there a possible connection between these two surnames? And the answer is a resounding yes.

In a previous blog post, I described the different variants of the Farrell surname and one of these is Harrell. You can see how this evolved by looking at the Gaelic version of the surnames: Farrell was derived from Ó FEARGHAIL which - owing to the aspiration of the initial f - mutated into Ó hEARGHAIL from which was derived Harrell. This reasonably explains how the similarly sounding surnames were derived from the same root.

STR data in the Farrell-Harrell-Kelley group

Ray has entered the 50 odd members of the Farrell-Harrell-Kelley Cluster into a spreadsheet and analysed their STR data. This exercise produced some interesting results:
  • The cluster includes 4 members of the Kelley Surname Project, one of whom tested negative for FGC20561. This suggests that some Kelley's branched away from the rest of the FGC20561 group around about 900 AD.
  • Excluding the 4 Kelley members, the common ancestor for everyone else in the group is estimated to be about 600 years ago (i.e. 1400 AD).
  • There are smaller clusters within the group with STR profiles suggesting a much more recent common ancestor. For example, the Harrell's are estimated to have a common ancestor who lived within the last several hundred years. 
  • The DNA evidence suggests that the Harrell group branched off from the Farrell group. Based on STR data, this branching roughly occurred about 9 generations back, or about 1680 AD (90% range 1380-1860). A more accurate assessment of the TMRCA will hopefully be available in the not too distant future. The one Lakey member is a suspected NPE who probably descends from a Harrell male.
  • Certain STR profiles appear to be solidly associated with the group. Here is what Ray writes in a recent email:
One aspect that I'm pleased with is that the analysis on FGC20561 and it's brother FGC20562 has highlighted a STR profile that appears to uniquely identify prediction of belonging to one or other of the Haplogroups. It involves just two Markers, namely DYS464, and DYS422. I had never previously considered DYS464 to be a reliable indicator, but in this case it appears to have been pretty stable for the past 1000 years, and DYS442 is a bit of a flip-flop marker in the Z253-1716-11 Cluster.
The unique marker values first appeared in FGC20560/Z17685 which lies immediately above FGC20561 and FGC20562 with a mutation in DYS464 from (14-15-16-17) to (14-14-16-17) combined with a value for DYS442=13. There was a subsequent mutation that occurred under FGC20561 of DYS442=13>14. I have used Mike Walsh's R-L21 Haplotype spreadsheet to check the prominence of these values, and every single one out of more than 14000 results occurs in a member of the Z253-1716-11 Cluster, with an overall STR profile that matches FGC20560, or FGC20561 or FGC20562.
Based on this information, I have located some additional FGC20561 suspects who also match the Farrell/Harrell/Kelley profile and have added them to the Spreadsheet which is attached. they include a Fagan group, a Pickett, May, Rock, Grace and Devine. Incidentally, Farrell Kit # xx9133, currently in the ungrouped section of the Farrell Project, belongs in the GF2 group.
In the previous post, member xx9133 was identified as an additional member from the Ungrouped section who belonged in R1b-GF2. The fact that he matches the "unique STR profile" (which is akin to rare marker values) provides additional internal consistency and lends further support to his inclusion in R1b-GF2.

More individuals will become apparent over time, and each addition to the group brings us a step closer to figuring out where they came from.

Links to Ancient Genealogies

Ray points out that the Lordships (territories) of the O'Farrell and O'Kelly clans were side by side, with the O'Farrell's in Longford and the O'Kelly's in neighbouring Roscommon and Galway. This could very well point to Central Ireland as the ancestral origins for the members of the Farrell-Kelley cluster.

But there are other clues that might help us pinpoint the ancestral origins of R1b-GF2. If we take a closer look at the Farrell-Kelley Cluster and the surnames associated with it, we may get a better idea of the likely ancestral origins for this group. This Surname Distribution Exercise is a big undertaking and will be described in a separate article.

Lessons Learned

There are several important lessons to learn from this new information:
  • Haplogroup Project Administrators have a much broader view of DNA results than Surname Project Administrators and have valuable contributions to make to surname projects by linking them with their closest neighbours.
  • This underscores the importance of all project members joining their relevant Haplogroup Project, which for R1b-GF2 include the following major projects (there may be others as well):
  • The current terminal SNP for R1b-GF2 is FGC20561 and this is dated to about 1100 AD.
  • The R1b-GF2 Farrell group appears to be at least 600 years old (1400 AD).
  • The Farrell - Harrell branching point is about 1680 AD.
  • FGC20561 is not shared uniquely by the Farrells of R1b-GF2 - other surnames sharing this SNP include Yorke, Cain, Ramey, Finch, & Kelly ... and all appear to be Irish surnames.
Next Steps
  • Several members in R1b-GF2 should undertake the single SNP test for FGC20561 (or preferably the Z253 SNP pack) in order to confirm that they are positive for this SNP. This applies in particular to the 4 members who were recently moved from the Ungrouped section (i.e. kit numbers ending 7344, 9133, 8478, & 0126).
  • Several members should also consider doing the Big Y test in order to discover additional SNPs downstream of FGC20561, one (or several) of which will eventually be definitive for the Farrell's of R1b-GF2.  Two members would be optimal in the first instance, as far apart genetically as possible (i.e. a Genetic Distance of 10/37).
  • Upgrading to 67 or 111 markers will help facilitate the construction of a Mutation History Tree and define the branching structure within R1b-GF2, and indeed the whole Farrell-Harrell-Kelley Cluster. This in turn will help date the various branching points in the tree and identify which R1b-GF2 Farrell's are most closely related to each other.

* Remember: there are two types of DNA marker (SNP markers and STR markers) and therefore two kinds of DNA data (SNP data and STR data)
Maurice Gleeson
June 2016

No comments:

Post a Comment