02 January 2010

MacGregor DNA Project update January 2010

MacGregor DNA Project – update January 2010

Welcome to the annual update for the MacGregor DNA Project. One of the principal, but happy, difficulties in presenting a report on the MacGregor DNA group is in deciding how to break the results down into small enough groupings so that the diagrams do not become so dense they are unreadable. With 439 results posted as I write this, we remain one of the biggest of the Scottish clan projects. With so many results an attempt to put them all onto one diagram would create something that was not legible. This year I have therefore decided to concentrate on surnames and the surname groups which appear when viewing the Family Tree DNA MacGregor project website (www.familytreedna.com/public/macgregor). As a result I have not made charts which link some haplogroups together (for example there is no specific chart here with both R1b, the Atlantic Modal group – sometimes associated with the ‘celtic’ and the I group, generally applied to the Viking invaders). I am always prepared to generate specific charts for individuals if they would like to contact me.

Before presenting the charts just a few words about how they are generated. I take the raw data from the results page as published by Family Tree DNA and export it into an Excel file. For the purposes of identification I keep the kit number and surname. In many of the charts then each entry is identified by its kit number and part of the surname. For some of the charts however I have cut the surname label down to a letter key only which is given under the chart. With the Excel file complete I copy and paste the whole into the online Y-DNA Comparison Utility created by Dean McGee http://www.mymcgee.com/tools/yutility.html?mode=ftdna_mode using 50% as the Probability, 30 years per generation on average, and the McDonald Mutation Rate figure. I check the box which asks if I want ‘PHYLIP’ data. Once the program generates results I go to the PHYLIP box and copy the results – these I then paste into ‘Splitstree’ a downloadable program that creates the charts www.splitstree.org:

D. H. Huson and D. Bryant,

Application of Phylogenetic Networks in Evolutionary Studies, Mol. Biol. Evol., 23(2):254-267, 2006.

I use 50% most of the time because this seems to produce dates for comparing individuals which more or less conform to what the genealogies suggest concerning family splits. The programs use a combination of number of mutations related to mutation rate to produce likely time distances between each individual’s results compared with each other person’s results. This can sometimes produce less likely groupings for some more distant connections, separating those that probably ought to be together. However for more recent splits, as will be seen, the charts give some good indicators of family groupings.

CHART 1 - MacGregor 'bloodline' 67 markers

This first chart compares all those ‘bloodline’ MacGregors (spelling is irrelevant) who have undertaken the 67 marker test. It includes 3 others, Gelston, Reid and Peterson who have similar DNA results to the MacGregors – and the chart suggests that the connection of these three is somewhat earlier in time than for the others in the group. The main group contains a McNab (60111) and a MacFarland (120820) who seem to have at least a genetic connection, which suggests either a split just at the time of the adoption of surnames or two examples of the use of an alias (not unlikely as the McNabs were numerous in the Killin area, and the MacFarlanes in the Loch Lomond area). Any results which group together on a twig are likely to have come from the same root individual. A good example of this is the Stirlings who are believed to all come from a Robert MacGregor alias Stirling born in the 1660s. This group have a distinctive mutation in the third panel which no other MacGregors share. Therefore if kit numbers 3346, 44327 and 24029 are indeed related to this branch then the split off from the Stirling group is quite likely to have happened pre-1660. This chart was generated using 95% Probability – reducing to 50% does not change the groupings at all.

It’s worth commenting that the program does not always preserve the groupings in the same way – though not presented here I have another version of this chart whose result significantly alters the position of kit 1774 – to make it an isolated branch on its own. This is interesting because we believe 1774 to be of the Glengyle family which is the sub group of MacGregors to which Rob Roy belonged.

CHART 2 - all MacGregor 67 results other than 'bloodline' (those in Chart 1)

This chart groups all those in the project who are in the haplogroup R1b (with 67 marker results) but excludes the group of MacGregors who are in Chart 1 with the exception of kit 2124 MacGregor, used for reference purposes. It’s interesting that the program groups McKinlay (17621), a more genetically distant McGregor (28296) and a Skirling (13635) into a sub group, and then connects these with another group of more genetically diverse McGregors, a Greig (9690), a McWhannel (46397), and a Macadam (49834). On the other side of the chart the two Turks are clearly related, the two Starlings (41885 and 52255), the McGehees (74970, 51942, 121911) and Welcher 93788. Peterson 15223 and Reid 153633 seem also to be related more directly.

CHART 3 - all MacGregors tested 37 markers (without 'bloodline', 'Irish' and 'Viking')

This chart includes all MacGregors who have tested at 37 markers but omitting 1) the main ‘bloodline group’, 2) those MacGregors believed to be specifically Irish in genetic origin (these are on a later chart) and 3) those MacGregors who are genetically Viking (these are also on a later chart). Here some particularly interesting groupings emerge. Kits numbered 99676, 164124, 155381 and 158870 are clearly all related and all the known genealogies point to a Rosshire origin (and they may have used Gregor as a surname at some point). Kits 129009, 94589 and 158917 are again clearly related – the only genealogy submitted there has its earliest generation in Clackmannanshire. Kits 84081, 29834, 137093 and 126138 seem to be related to the Rev William McGregor, purportedly from a place called Glen Ossian in Scotland, one of a band of brothers – Bartlett, John, Andrew and Alexander. Kits 2726, 138485 and 7183 seem to have a common origin in the vicinity (that is, within 15 miles) of Perth and may have used Gregor as a surname for a period. Kits 153532, 164088 and N62107 are certainly related but I do not have any genealogical data for these. Finally on the chart the 3 results on the far left are quite separate in origin and belong to different haplogroups from R1b to which the bulk belong. The next chart cuts out some of the other results to show three of these distinct MacGregor groups more clearly.

CHART 4 - as Chart 3 but with some detail removed

CHART 5 - Gregor, Grigor, Gregory, Greig, Gregg or Gragg.

This chart lists all participants (at 37 markers) called Gregor, Grigor, Gregory, Greig, Gregg or Gragg. The code used is as following: numbers only = Gregory; Gr = Gregor or Grigor; E = Gregg; EIG = Greig; A = Gragg. The letter N is a Family Tree DNA kit designation and has no significance. The most striking part of the chart is the collection of Gregories at the very bottom who are certainly from a common and quite recent ancestor. The program suggests that there may be some connection between Gregory 81125 and the Greig and Gregg results beneath – the time distance involved here could be as much as 500 years from the original split. Apart from these, other potentially related groups can be seen working clockwise round the chart. The two Greggs 130191 and 6979 are definitely related as are 110496 and 137236; the Graggs 81282 and 158127; the Greigs 64662 and 45360. There is a potential genetic relationship between Greig 9690 and Grigor 131056. Further information on many of these families can be found in the Gregg and Gregory DNA projects, and also some genealogies at:

CHART 6 - Irish DNA group (does not include 'Irish' McAdams)

This chart contains all the specifically Irish DNA (37 marker) results which are characterized by, in the first 12 markers, a 25 at marker 2, 11,13 at markers 4 and 5, and 14 at marker 11 though not all results will have all these mutations. The Irish results show a strong similarity to each other indicating a common ancestry in prehistory, but in more recent times the Greers (however spelt) seem to have common ancestor of that name, possibly linking with the Griersons 7874 and 94757. Again 2124 MacGregor is included for reference. White 120077, Skinner 57893 and McGregor 35727 are also more distantly related genetically. Compare this chart with the next which lists only Griers (however spelt) and Griersons – all results, and genetically Irish McGregors. The key for this is M = McGregor; S = Grierson; all others are Greer/Grier.

CHART 7 - 'Irish' results on 67 markers

This chart has the Irish based 67 marker results and shows two discrete groups all having common point of origin and indicating the close genetic ties of the Griers etc. Also on this chart are the Orrs – one in the Irish group and the rest a separate haplogroup R1a, who might also have their genetic origins in the Viking invasion of Ireland but who certainly have Irish connections at the present time.

CHART 8 - Grier(son) groups - different genetic origins

Again this is 37 marker results. The Grier/Grierson group on the right seem to share a common ancestor within the time of surnames; the Grier/Greer group in the middle has a separate origin and one might conjecture that these are the Griers of S.W. Scotland – connected to the Griersons of Lag. At the far left is a mostly very strongly related Viking group of Griers, and again these might be conjectured to have descended from a Viking settler in East Ireland.

CHART 9 - Magruder, McNie, McAdam and participants in the 'unattached' R1b group

This chart shows McAdam, Magruder, McNies and other participants whose results are listed at the FtDNA results page unattached to any specific group (37 markers). From this chart we can see that there are various groups of McAdams suggesting different origins: 83205 and 82874; 12683, possibly, with 5237, 76324, 144894, 16568, and 71022 (note this is a genetically Irish group); 54141 with 8857; 3714 with Thorn 48842. According to the groupings all other McAdams are singles or more genetically distant. The two McNies are related 66228 and 165873; the two Smiths 35819 and 32132; the Stalling(s) 21898 and 151153. All the Magruders are related 61472, 141702, 141700, 46179. MacGregor 2124 is again included for reference.

CHART 10 - MacGregor 'bloodline' on 37 markers

This chart contains all the MacGregor bloodline results with 37 markers. Some of the groupings are revealing as will be seen. The key for this is S = Stirling G = Gregory; Macp = Macpherson; C = Campbell; McNa = McNab; J = Jamieson; Ba = Bain; D = Drummond, McF = McFarland. All others are MacGregor however spelt. Kit 2124 is marked with an MM in front.

CHART 11 - Sept or like names grouping (37 markers)

This chart has all those who are listed in the group ‘Sept or like names’ with 37 markers. Kit 2124 is included for reference and interestingly there is a group of Stirlings with similar DNA to 2124 – but this group of Stirling are quite distant from the earlier group of Stirlings whose ancestor is Robert MacGregor alias Stirling. The Turks are all closely related as are all those at the bottom of the chart whose name is McGhie (however spelt) (and Welcher). Although there are Magees with the typically Irish signature, this group do not have that DNA signature and may be Scottish in origin.

Chart 12a - 'Viking' groups on 37 markers
Chart 12b - 'Viking' groups on 67 markers

These two charts show the Viking-related genetic results from the project, the first chart with 37 markers and the second, those of the group who have increased to 67 markers. There are two separate genetic groups: those on the left are I2 and those on the right I1. The testing company Ethnoancestry (www.ethnoancestry.com) says the following about each group:

‘Haplogroup I1(*) is an indigenous European group and is hypothesised to have spread out from the Iberian refugium after the last glacial maximum 18,000 years ago. I1 reaches its highest frequency in Scandinavia and has a decreasing gradient in frequency to the East and to the South and West. I1 is relatively common in the British Isles, having been taken there by Norse and Danish Vikings as well as Anglo-Saxons’.

Haplogroup I2a(*) appears to have originated in the Balkans, perhaps from a glacial refugium there; I2a is very common in Croatia and Bosnia today and decreases in frequency across Eastern Europe. A rare offshoot branch of I2a is also found further West, including in the British Isles’.

‘Haplogroup I2b(*) appears to have originated near modern day Germany, where it reaches it peak frequency. I2b is found spread across a broad area of NW Europe including the British Isles, where it has been brought by numerous historical migrations’.

For note – apart from the above results the following participants have different predicted haplogroups: Kammer R1a1; Seuch R1a; Westran R1a1; Bennett N6998 J2; Stivadoros G

If there are any comments or questions about this update please address them to me: richardmcgregor1ATyahoo.co.uk


  1. Comments by David Grierson in Melbourne
    Part 1

    Richard, following our private discussion I'm posting some comments relating in particular to Charts 6, 7, and 8 in your study. I know that you were illustrating particular points when you designed the charts, but there are some aspects which merit further discussion. I have three points that I would like to bring to notice, although I appreciate that we (the Grier(son)/Greers - Grier for short) are to some extent fringe dwellers in the study. I will first reiterate my position on the Grier/MacGregor relationship - there is no evidence whatever in any DNA test so far published that any of the Griers descend from mainline MacGregors, and the (so-called) Gracie Tree - the only known source of the supposed relationship according to Robert Grierson - is discredited in this area. The recent on-line publication of a small book written ( and published in the early 20th Century) by Robert Grierson (a Glasgow solicitor although born in Ireland) states quite clearly that there is no reason to believe that particular story. He also says that his cousin, Sir Phillip Hamilton-Grierson, the person most familiar with the Lag Charters, thought that particular story to be false. I think I have identified the Charter (as have many others) that was misrepresented by Gracie in building his tree. Further, there are no YDNA results of tests (that I have seen) that clearly relate MacGregor and Grier at some point since the first extant records, say 600 years. That is not to say, of course, that either name was not used as an alias at various times. What it does suggest is that people who used aliases ultimately reverted to their birth name - but of course, the sample is small.

    To my points, which concern: the persistent use of the "Irish Related" statement; the phylogenetic tree in Chart 7; and the suggestion that (in the discussion relating to Chart 8) L21 Greers might be connected to the Griersons of Lag.

    Irish Related?
    It seems clear to me that the MacGregor founder was L21 in SNP terms (otherwise called R1b1b2a1b5 by FTDNA), based on the few SNP tests completed within your study and also a close comparison with the L21 samples within the WTY project. L21 is, of course, ancestral to M222, the Clade central to my study. A question that follows from this proposition is where and when did the M222 mutation occur? Dr Anatole Kyosov thinks that L21 occurred up to 5000 years ago in France, and the "group think" for M222 seems to be that the mutation occurred up to 2200 years ago. This is several hundred years before M222 is thought to have arrived in Ireland. France seems to be a logical place for it have occurred also. There is no doubt that L21 went to Ireland early in it's life, as it is the predominant R1b Haplogroup there.

    You have correctly identified the three major groupings of Griers, and for this discussion I am ignoring the Vikings. I have generated modals for the L21 Griers and for the M222 Griers. A generic Scots modal also exists. The L21 Grier modal and the Scots modal have a GD of four, which could indicate a split as long as 2000 years ago as I understand it. So the L21 Griers are only distantly related to the Scots. Kevin Campbell has suggested that 389ii at 29 is not Scots, and neither is it necessarily Irish. Of more interest, in 25 markers there is only one difference between them and the so-called "French Cousin" from WTY. Samples of other L21 modals show little correlation. This is interesting given the argument below. How and when did the L21 Griers get to Scotland? Did they come direct from France or via Ireland like the Dalriadic Scots?. When one compares the L21 Grier modal with the M222 Grier modal, it is clear that these two lines went their separate ways well before the M222 mutation occurred. But that doesn't preclude the notion that they remained in the same area and continued a relationship during the thousands of years implied by the genetic differences.

  2. David Grierson comments Part 2
    (NB the blog comments can only extend to 4096 characters)
    Now the M222 Griers have 8/67 differences between them and the YSearch M222 modal, and these differences are remarkably consistent. All of the 19 tested M222 Griers bar one are 18 at STR458, a remarkable result in a "fast-moving" marker. The majority are 14 at 389-1. All bar three are 17 at 364c. All bar two are 18 or higher at 570. Only two are 38 at cdya, the rest being 37 or 36. All 11 tested at 444 are 13, a very rare mutation in M222. What can we derive from this? Firstly, given the preponderance of Niall derivative surnames tested, I think it is fair to say that the YSearch M222 modal is very close to representing the Niall family modal. So comparing our modal to the YSearch (Niall) modal can perhaps give us something. You will be aware of the theory of off-modal comparison as being the best guide to MRCA, especially when the family group represented is consistent in its results. On this basis, it is my view that our modal describes a separation from the line that went to Niall back near the founder of M222. In other words, I think the evidence says that our line is not a part of the "Irish" descent, and may be directly French (ie not via Ireland). To take this argument to the logical conclusion, it seems to me that the MacGregor mainline is more "Irish Related" than we are! However, I would ask you to consider my arguments about Chart 7.

    Chart 7:
    To me, Chart 7 exhibits two useful aspects of the name relationships - but you haven't mentioned either of these aspects. In any case, I wonder what you were trying to illustrate in mixing four discrete haplogroups (or clades thereof) on one chart. From my perspective, your chart tells me that in M222 you have identified three very distinct groups. There are the clearly related Grier group on the right, the group above them that are separated by, what, 150 generations? And the odd man out, McGrew. Leaving him aside, because I can't explain him (other than to say that he represents as early a split as we do), 150 generations, up and down, is about 2250 years. That fits remarkably well with previous comments. More importantly, as I understand your own accepted "Irish MacGregor" modal, it is virtually identical to the "Niall" modal. So, in the case of the M222 clades, perhaps you can correctly identify the top-right group as "Irish Related". I don't think you can say the same about the Griers. Now, centre bottom, you identify two U152+ members of R1b - R1b1b2a1b4 in the FTDNA nomenclature - nothing to do with M222. Different names, different clades, and a, what, 5000 year split? Why are they there without comment (other than Irish related)? You have the R1a1s out to the left, who knows when they split with the R1bs, but almost certainly greater than 10,000 years ago. No comment? And, strangely, you have the Q1a3 on a shorter lead than the R1a1s. Now, given that by definition Q and R split before R1a and R1b split, what this shows is the danger due to convergence of using haplotype alone (without reference to haplogroup) to identify relationships. That is, M222 cannot possibly be more closely related to Q than to R1a. I think these particular aspects should be placed before your readers.

  3. David Grierson comments Part 3

    I don't know how you arrive at the suggestion that the Lag relationship might be found in the L21 group. As I read your notes, there is one family in L21 who seems to have made this claim. But, there are several in M222 who have that family tradition. Further, many of the Griersons (all of whom are M222+) have long Scots traditions, albeit the paper trail is poor beyond about 300 years (and there are no northern Irish Griersons, I believe all Irish families of the name went to Ireland from Scotland relatively recently (+/- 400 years), see Robert Grierson etc). Why would you look in the L21 "Greers" when there is a simpler explanation? I think that most of the tested Griersons, together with some of the M222+ Greers, are relatively close to Lag (say 450 years). Some of them are perhaps an older connection, say 600 years. We have just received the test results for a Grierson who may have the best claim to a Lag connection so far. He is connected on paper to the migrant Argentinian family of Griersons. A member of that family - the famous Dr Cecilia Grierson - years ago claimed the Lag descent (possibly with documentation, although who knows where that is now?). I am working to understand the use of the McGee TMRCA utility, and will publish my results with respect to the M222+ Grier relationships when complete.

    David Grierson in Melbourne