The MacGregor DNA Project

MacGregor DNA Project  blog update 2018

This year’s blog will be slightly different from those done previously in that I want to spend a portion of it explaining how genetic genealogy helps us understand intra-clan relationships but in order to continue I need to say exactly what a Scottish clan is, as there are some particular misconceptions concerning the exact nature of clan in the 21^st century.  At the end, I go on to discuss how SNPs are helping to build up a better picture of family groupings within the main line MacGregor family tree after the mid 14^th century.

What is a clan in Scottish usage?

Six hundred years ago this question was quite simple to answer. You were associated with a clan if you had been born with the name – in the MacGregors’ case that might be expressed as Gregor, Grigor, MacGregor, McGrigor, McGregor and a whole range of alternative spellings such as, for example, McGreagor (an attempt to render the Gaelic phonetically into English?). At that time too spelling had not been standardised - so one might find Mckgregor, M’gregor and so on. You were also a member of the clan if your name was an accepted variant, such as Grierson, or Grier, Greig/Grieg/Grig etc. These were considered to be shortened or anglised versions of the main clan name. So, Grier-son equals Gregor-son and Grier is the same name without the ‘son’ on the end.

Whether or not these accepted names were genetically related to the main line was not the point, as a clan was a collection of related surnames. Members of the clan recognised as Chief the head of the main line (the Chief of the MacGregors for example), and often, especially in the early days relied on him for protection, or rather, on his ability to pull a ‘federation’ of individuals together to ensure, usually armed, protection, or, as a means of seeking retribution on another group for some offence.

There were others associated with individual clans. People who belonged to individual septs. Sometimes the same name would appear in two or more lists of accepted septs of different clans – such is the case with surname King, among others. Again, some descriptive words used as surnames were understood to have been borne by people associated with the clan, and these surnames are found in many clan lists: Bain (or Ban) or its anglicised equivalent White; Roy meaning Red; Dhu or Dow meaning dark or black, are some examples.

Finally, there were people who answered none of these ‘qualification’ but who lived on the land which was under the Chief’s influence as ‘part-takers’. Grant and Menzies rental documents of the 18^th century reveal instances where individuals adopted the name of the local chief where formerly they were called only by their patronymics. A patronymic shows the genealogy of an individual back two generations – so my patronymic would be Richard McEwan VicPeter: there McEwan is not a surname but shows that my father was Ewan (not THAT Ewan). On a rental document you might find John McGregor VicPatrick which means that John’s father was Gregor and his grandfather Patrick: if John were a very poor inhabitant – a cottar – it might be that he had lost the knowledge that he was a genetic MacGregor and so ended up taking the surname Grant. It’s unusual to find this situation among MacGregors because of their turbulent history but it happens in other clans. Paradoxically because the MacGregor name was proscribed [forbidden] for so long [1603-60 and 1693 to 1774] it seems many families held on to the knowledge that they were MacGregors despite having been forced to take other surnames. Some families never changed back to MacGregor when it was finally possible to do so – which is why in the DNA project we see individuals called Drummond, Stirling, Campbell etc who are genetically MacGregors - their ancestors never readopted the name when it was safe to do so.

What is clan in the 21^st century?

To some the concept of clan in the present century is an irrelevance. We no longer need the Chief to protect us and we no longer live in defined communities of inter-related individuals offering support to each other in the same way. In the 21^st century a clan is made up of people who value the bonds of kinship often promoted by Clan Societies, and they value the, sometimes surprising, connections that clan association brings. At its gathering in 2014 the Clan Gregor Society of Scotland had 11 nationalities represented, some of whom like the Philippine and South African contingents actually shared a recent common ancestor. Members of a clan still recognise the Chief as head of the clan, but, as I have explained, it is likely that relatively few share a common ancestor with him (or her). Most Clan Societies recognise that finding paper evidence of relationships becomes increasingly difficult before 1750 and therefore accept both female line connection (e.g. ‘great grandmother was a MacGregor’, or a King, or a Bain etc), or, that there is a tradition within a family of MacGregor clan connection.

What does genetic genealogy tell us?

1)    That people called by a certain surname are not necessarily descended from a common ancestor, although as we have found in the DNA project approximately 50% within a surname subgroup do share a common ancestor. (I will explain my subgrouping in a moment)

2)    That there are different groups of individuals called, say MacGregor, who share a common ancestor in the fairly recent past but the connection of one group of these MacGregors to each other is likely to lie thousands of years in the past and pre-surnames

3)    That there are multiple origins for surname groups

4)    That clan sept names have varied genetic origins and not just one single origin within the period of the existence of surnames (surnames become more common after the mid 14^th century, particularly in England following the Black Death as there was more movement of labour to replace those lost to the plague)

The MacGregor DNA project at this date has 1427 members. In order to make the results easier to navigate I have divided them up into subgroups. The advantage to this is that within a subgroup it is easier to see individuals who are potentially related, in the more recent past, to each other, because they appear in the surname grid near to each other. It also shows up where there are multiple origins for surnames, and is particularly useful where some members of the subgroup are able to indicate an earliest known ancestor. In a good number of cases that ancestor will be shared with other members of the subgroup and some of the group may not have the genealogical information that others have and so find their link to the past. It can also suggest geographical location for the ancestor. The disadvantage is that members of different subgroups in the results grid cannot quickly compare their results with members of other subgroups, but in a Y chromosome study that is less of a problem. In any case everyone has access to a ‘Matches’ tab on their DNA results page and also everyone can use Ysearch and Mitosearch.org.

To further illustrate what can be done I will demonstrate with one subgroup. Regular readers of this blog know that I have offered to make any comparisons between individuals that are wanted. It would be helpful if this could normally be limited to 10-12 results for comparison. Also, please remember that I have to be able to compare like for like – so I cannot compare a 37 result with a 67 – I can only use the first 37 of the latter for the comparison.

Example 1: Greig/Grieg/Gregg etc surname 37 marker result grid

The first example shows all the 37 marker results for this subgroup. No colours are present in this grid to show mutations because we are not trying to compare individuals to one ‘master’ result or even to an ‘average’ or modal result. It can be hard without colours to see how one individual is related to another but in general if results are closer together on the grid they would tend to be more closely related genetically (but just being close on the grid does not necessarily pick up results that are more related than others). That is why we use a graphics program to make comparing results more visual in what is known as a phylogenetic tree. Also the program used picks up similarities across results that might not be immediately obvious to the eye.

Example 2: Greig/Grieg/Gregg etc surname 37 marker phylogenetic tree

The second example therefore shows the chart in graphic form, generated by Splitstree (acknowledgement is given at the bottom of this blog). I have labelled them as the chart is labelled so perhaps the easiest way to make comparisons is to print off both examples and cross-refer between them. In this graphic representation some relationships become more obvious but there also are some surprises. The first, and most important, point to make is that we are seeing at least 11 distinct family groups who each shared a common ancestor many thousands of years ago long before surnames became common. So we see at least 11 different genetic origins for people called Greig, Grigg, Gragg and Gregor. In general, if results are closely clustered together on the graphic then they probably share a recent common ancestor (and by recent I mean since the acquisition of surnames).

1)    There are 7 individuals (kit 20673 is one) who share a common ancestor who could be the William Gregg born in 1616 or his immediate forebears - an early emigrant to the New World. This family have no spelling variations – always Gregg – so may have been literate from the earliest emigrant

2)    There is a group of 3 (kit 214992 is one), who are related and may have a connection to Tipperary in Ireland: these families are Gregg or Gragg

3)    There is another group of 3 (kit 239449 is one) - these connect to a common ancestor but there is no indication of who this might be in the genealogies submitted – these used spellings Grieg, Greig and Gregg

4)    Another group of 4 (kit 158127 is one) who seem to connect to Antrim in Ireland (using Gragg and Gregg)

5)    There is a group of 3 (45360 is one) who connect to Edinburgh and Pathhead (a nearby village)(surnames are Greig and Gregg)

6)    There is a group of individuals who are genetically related and with geographical links to North East Scotland with very different versions of the name: Griggs, Greig and Grigor (9690 is one of these)

7)    A group of 2 (kit 6979 is one) using Gregg (no locations available)

All other individuals appear to belong to separate unrelated families although the distant possible connection between Gregor(y) 476609 and Charles Greig 585177 would be worth further investigation. Robert Gregor 239031 belongs to a completely different genetic haplogroup.

Example 3: Greig/Grieg/Gregg etc surname 67 marker grid

This example [Examples 3 and 4] shows the smaller group of individuals who have tested 67 markers (they are all in the 37 group discussed above). What we are interested in is whether the greater number of markers gives any further information on genetic connections. Since 50% of the sample shown in Example 1 and 2 is now not present, the graphic representation is much sparser.

Example 4: Greig/Grieg/Gregg etc surname 67 marker phylogenetic tree

The problem that is immediately apparent [in Example 4] is that without a larger number of individuals testing to 67 (or more markers) family groups do not break down significantly further.  In Example 4 the only possible confirmation seen is that Kits 7489 Gregg, 214992 Gragg and 9690 Greig may share a common ancestor in the relatively distant past but it is possible that all three share a geographical origin –  North East Scotland as suggested in point 6 above.

                                                            ***

By way of comparison I have used the same processes on the Gregory group who also have diverse origins, but what is particularly interesting with this group - given that membership of a DNA surname testing project is essentially based on random participation - is to see just how many individuals descend from the same ancestor: almost certainly, given that there are forebears in common (for example Gideon Gregory – kits 58711 and 179683) then this group probably had a common emigrant ancestor in the United States.

Example 5: – Gregory surname 37 marker phylogenetic tree

 Apart from that group of related individuals there are only 4 other, much smaller groups, whose individual ancestry lies close to each other – their ‘earliest known ancestor’ as given by each participant,  suggesting a range of possible genetic origins (see the grid on Example 6)

Example 6: Gregory surname 37 marker grid

Dean McGee’s DNA Utility allows an estimate of time to most recent common ancestor.

Example 7: Gregory surname 37 marker Time to Most Recent Common Ancestor grid (partial)

This is only an estimate and the number of years suggested always depends on the confidence level chosen for the program – choosing 100% confidence would give a different result from choosing 10% confidence. In this example from the Gregory charts we see an estimate of the possible time to the shared ancestor for the each individuals in the group with each other person. In order to see a good number of results I have had to remove the labels from the top grid but they can simply be put in by hand - going from left to right on the grid top line in the same order as reading top to bottom. Notice, for example, that comparing the first two individuals ‘Peter R. Gregory’ 275887 and Gregory 37140 suggests that they share a common ancestor 5220 years ago.

All of these analyses benefit greatly – and benefit other genealogists – if testers indicate the name of their earliest known male ancestor with the surname – no matter how recent that might be.

If we now look at the phylogenetic tree created when we use only those kits that have tested to 67 marker level the only real difference is that some of the genetic distances seems to be clearer.  

Example 8 Gregory DNA 67 marker grid

However, this particular program only separates by mutation – so if we look at the Gideon Gregory results again it looks like they come from different lines of the same family, rather than from the common ancestor Gideon. This is a limitation not of the program but of an ability to input into the program that two results come from the same ancestor. We have to remember that programs such as this were not designed primarily for family history but for comparing genetic markers in species and not just the human species. After all, it would be next to impossible to say whether two turtles shared the same great great grandfather …

The development of SNP analysis

For several years now there has been an increasing focus on the testing and analysis of SNPs (single nucleotide polymorphisms). The difference between these and the more commonly tested STRs was given in my 2014 blog (opening paragraphs). To put it simply, SNPs are markers in time: as far as is known if these mutate they stay mutated in subsequent generations. What that means is that once enough SNPs are identified an element of dating can be applied to when the mutation happened. For family historians this fact is becoming hugely important. Dating SNPs to the time before surnames is of limited use to family historians but to have dates, even approximate, from the time after the adoption of surnames means that family surname groups can be split down into smaller and more recent family subgroups.  On the Greig and Gregor grids I show the SNP information which is assigned to each individual in the leftmost column. In most cases this is simply M269, a SNP that happened thousands of years ago. Some individuals have had some SNP testing done but very few people have had the ‘Big Y’ test done which takes results forward in time towards the present day, and identifies SNPs which may have happened between 500 and 800 years ago.

Greig SNPs

In the Greig grid confirmed SNPs are in green. M269 is too early in date to be considered so the only other SNPs to be taken into account are kit 476609 R-L066; 363402 R-FGC10125; 259416 R-U152; 585177 R-Z253; 195430 R-U106; 295321 R-FGC5494; 404866 R-FGC37100; 110496 R-L21; B196295 R-ZP77; and 239031 T-M70

Of these U152, R-Z253, R-U106, R-L21 are well known early SNPs which happened before surnames, sometimes by thousands of years and most have further testing options available to bring the results further forward in time. FGC in the results indicates that the SNP was identified by the Full Genome Corp (as indeed the other letters identify the source lab or individual who identified the SNP in question in the first place). Of the other SNPs:

L1066 is more recent but still before surnames.

[see https://www.familytreedna.com/groups/l1066-plus/about/background].

 FGC10125 [see https://www.geni.com/projects/R-FGC10125-Y-DNA/36784]

is the next SNP in the sequence for some individuals after L1065 [not the same as L1066] which is said to identify the Scots modal group

[see http://www.ytree.net/DisplayTree.php?blockID=160] Since L1065 roughly dates to 1750 years before the present, FGC10125 may have happened before surnames.

FGC5494 is European in origin but again is somewhat earlier than surnames, and has SNPs which descend from it towards the present time

[see https://www.familytreedna.com/groups/r-fgc5494/about/background].

FGC37100 is a descendant, or technically, ‘downstream’ of L151 – that SNP is again an much earlier one and found in England as well as other places.

ZP77 is the same as FGC6562 and is found in concentrations in Ireland and to a lesser degree in Scotland: it also has numerous downstream markers

[see https://www.yfull.com/branch-info/R-ZP77/]

Finally T-M70 is an very early SNP with a distribution over southern Europe, the Middle East and East Africa. It is comparatively rare among tested individuals [see www.yfull.com/branch-info/T-M70/ which dates it to 16,000 years before the present].

Similar discussion on identified SNPs could be done for all the subgroups in the MacGregor DNA project. The Gregory group, for example, has the following identified SNPs (see example 6):

R-S16906; DF21; R-CTS7678; L48; Z343; R-P312; R-L1336; R-BY15955;R-S691

MacGregor DNA – current SNP analysis

The work which Neil McGregor in Australia has been doing in analysing MacGregor SNPs is not concerned with the earlier SNPs. We already knew from Jim Wilson’s work that most MacGregors in the main line group carried the SNPs S690+ and S697+, both probably dating from after 1200AD (though no absolute dating is yet available). In his analyses Neil has begun to break down the test results of MacGregor participants into individual family groups – which the clan has known about for generations and which are referred to in older documents as the ‘houses’ or ‘sleik’ [of] Clan Gregor, the main ‘houses’ being Glenstrae, Roro, Gregor McIan (or Brackley) Dugall Keir and more recently ‘of Glencarnock’ the Chief’s line (Glencarnock is the area they held from the mid 18^th century).

Neil’s current identification of family groups is given in Example 9:

Example 9: current predictions of SNPs associated with MacGregor family groups.

Neil’s email to me allowing this to be included in the blog suggests that MacGregors from the main line should seriously now consider doing the BigY DNA test (rather than FGC – Full Genome Corps – with whom we have also undertaken testing). He says:

“The best recommendation is that people get BigY as everybody seems to have between 3 and 8 separate SNPs which will allow them to be separated from everybody [else], other than from their own immediate family or first cousins. Some of them [those who have tested under BigY] appear to have a cluster of SNPs which appear to have mutated together and may represent one mutation. A mutation seems to be as low as once per generation through to once every 4-5 generations – seems related to the number of STR [the marker scores than participants start with] mutations as well.

The clan seems to be divided into two major clusters and this would appear to be early on. The section I am in has at least 3-5 sub-branches as does the other major group. The dividing SNP appears to be BY28714”.

Just to repeat that I can do comparisons of STR results for individuals – comparing with up to 10 to 12 others. I would repeat Neil’s encouragement to do BigY if you can – please ask me for further information if needed [richardmcgregor1ATyahoo.co.uk substituting @ for AT]

I have just had a comment from EMC which is worth repeating here in case folks miss it:
It is important to note that the results Neil has recently shown are also due to FTDNA reprocessing BigY kits under a new genome reference called HG38. Prior, under HG19, many of the SNP's used now were heretofore unknown. This SNP refinement is important.

Charts were constructed using Dee McGee’s Utility at http://www.mymcgee.com/tools/yutility.html?mode=ftdna_mode, using a 75% level of confidence, on Doug MacDonald’s mutation rate, an average of 30 years per generation and with no modal results assigned. The graphic representations of phylogenetic trees are made by Splitstree:

-->

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