This is a guest report from a geneticist

 

For more than two decades, the McGregor DNA project has been building a remarkable picture of shared heritage for Clan Gregor. We see that DNA is passed down through families in predictable but sometimes surprising ways, shaping how we are connected to one another. By understanding how this process works, we can better appreciate what our results are telling us and how they fit into the wider story of our family histories.

This post revisits the biological basis of genetic inheritance and explores it in the context of our DNA project. One thing quickly becomes apparent - understanding genetics is almost like learning a new language. There are many specialised terms and interpreting them is often the first step towards making sense of the bigger picture. Most of these terms will be written in italics when first introduced in this post.

Genetic inheritance

 All genetic inheritance begins with a biological process called meiosis, a specialised form of cell division that halves the number of parental chromosomes, producing sperm or egg cells a.k.a. gametes that carry a unique mixture of maternal and paternal DNA. This short video from the BBC explains the process. Human sperm cells carry 22 autosomes (chromosomes 1-22) and one of the sex chromosomes, either X or Y. Each chromosome on average carries approximately 2 metres of DNA made up of nucleotide bases which code for everything that makes us human.

 

Human eggs carry 22 autosomes and an X chromosome. Interestingly, eggs are the larger gamete containing mitochondria and other molecules to guide the developing embryo. Mitochondria contain their own DNA (mtDNA), which is inherited by both males and females from their mother, but only females pass it on to the next generation.

At fertilisation, male and female gametes combine to form a zygote and through multiple cell divisions during embryonic development, a new individual develops made up of trillions of cells, each containing 46 chromosomes. Males have 44 autosomes and XY sex chromosomes while females have 44 autosomes and XX.  All males inherit their Y chromosome paternally and their X maternally. Females inherit one X from each parent.

We don’t inherit an exact copy of our parents’ chromosomes

Along with halving of parental chromosomes, meiosis also involves recombination of chromosomes where DNA from the two copies of each chromosome is shuffled before being passed on. As a result, we don’t inherit an exact copy of our parents’ chromosomes, and they didn’t inherit an exact copy of their parents’ and so on. In just a few generations our DNA can be very different from our ancestors. Because of this, autosomal DNA (chromosomes 1-22) is limited to detecting relationships to approximately six generations, although linking these matches to specific surnames can be difficult without supporting genealogical information.

The X chromosome follows a different inheritance pattern, so X matches can provide additional clues about possible lines of descent, although they do not usually point to a single ancestral line.

The Y chromosome and paternal inheritance

The Y chromosome behaves differently during meiosis.

The majority of the Y chromosome does not recombine. Instead, it is passed from father to son largely intact, forming a continuous line of inheritance over many generations. This unusual stability is what makes Y‑DNA so powerful for tracing paternal ancestry. It also means that any changes that do occur are preserved and can be followed through time.

Mutation: a source of genetic variation

Mutations arise as part of the normal process of DNA replication, although their frequency can be influenced by factors such as age and environmental exposure. Mutations are often thought of as ‘bad’, but in most cases they are biologically neutral with no effect. Only mutations in gamete cells are passed on.

 

Despite its relative stability, the Y chromosome undergoes mutation events. These mutations are the source of variation along a paternal line. A DNA test does not measure ancestry directly, but the outcome of these mutational processes.

Two main types of mutation are relevant in Y‑DNA analysis:

1. single base changes in the DNA sequence, called single nucleotide polymorphisms (SNPs)
2. changes in the number of repeated DNA sequences at specific locations, called short tandem repeats (STRs)

These mutations occur at very different rates.

Haplogroups vs Haplotypes

SNPs arise relatively rarely. When they do occur, they tend to be inherited by all descendants of the individual in whom they first appeared. A haplogroup is a branch of the Y chromosome tree defined by shared SNP mutations. In practice, haplogroups are often experienced as large, overarching branches representing lineages that extend far back in time, although modern sequencing can now identify increasingly specific sub-branches within them, sometimes within genealogical timeframes. Haplogroups help answer the question, “Which major branch of the lineage does my paternal line belong to?”

If a haplogroup describes where a lineage sits on the tree, a haplotype describes how that lineage varies within its branches.

With STRs, which mutate more frequently, the number of repeats can increase or decrease from one generation to the next. This produces small, incremental differences between related individuals. A haplotype is a set of linked genetic markers that are inherited together. In Y‑DNA testing, this typically refers to a set of STR marker values (such as DYS393, DYS390, etc.) that together describe the genetic profile of a paternal lineage.

This is how many surname projects were built, by recognising patterns of similarity across STR profiles.

A model, not a record

One important point follows from this.

DNA testing does not reconstruct the past directly. Instead, it models how mutations accumulate over generations and interprets present-day variation in that context. These models are based on observed mutation rates, derived from large numbers of known relationships, such as father/son pairs and documented family lineages. By incorporating these rates, they estimate how likely it is that a given pattern of differences arose over a particular number of generations.

Matches, distances, and predicted relationships are therefore not exact measurements of ancestry, but statistically informed interpretations of biological processes. This does not make them unreliable, but it does place them into context.

From separate systems to integration

 Historically, STRs and SNPs were often treated as separate tools. STRs were used to compare individuals and detect relationships while SNPs were used to assign broad haplogroups. This led to a simplified view that STRs were linked to “recent ancestry” while SNPs were linked to “deep ancestry”. However, this distinction no longer holds in its original form.

 

Advances in technology: the shift to Y700

 Modern sequencing approaches, such as Big Y‑700, have transformed how Y‑DNA is analysed. Rather than testing a fixed set of markers, Y700 sequences large portions of the Y chromosome. both STR variation and thousands of SNPs, identifying both known and previously undiscovered mutations.

 

New SNPs discovered through testing can define previously unknown branches, so that the structure of the tree is continually being improved. In this sense, modern testing does not simply use the tree, it contributes to building it.

 

Ongoing surname projects, such as the MacGregor DNA project, illustrate how these principles play out in practice. Initial groupings are often based on similarities in STR markers, suggesting possible shared ancestry. However, as more detailed SNP data become available, particularly through tests such as Y700, some of these apparent relationships are confirmed, while others are shown to be misleading. This reflects the underlying biology - patterns of variation can suggest connections, but only the structure defined by shared mutations can fully resolve them.

 

Types of DNA testing: strengths and limitations

 

Different types of DNA tests provide different kinds of information, and each has strengths and limitations.

Autosomal DNA
Identifies relatives across all ancestral lines but within recent generations only and often cannot specify which ancestral line a match comes from without supporting records.

Y-DNA testing
Traces the direct paternal line and is particularly useful for surname studies but only follows paternal ancestry.

mtDNA testing
Traces the direct maternal line, but like Y-DNA represents only one ancestral pathway.

 

For this reason, DNA results are most powerful when combined with genealogical research rather than interpreted in isolation.

 

Final thoughts

DNA testing provides powerful insights into shared ancestry, but those insights sit within a biological system shaped by mutation and a statistical framework used to interpret it. By recognising how haplogroups define the broader structure of lineages, and how haplotypes capture variation within them, we can read our results with greater clarity. For a project that has been building evidence over more than 20 years, this perspective allows us to move beyond simply observing patterns to understanding what those patterns mean, and where their limits lie.

 

 MacGregor DNA project blog update 2026 

If DNA genealogy has taught us anything it is that surnames have many origins, and the adoption of a surname does not indicate a direct relationship with another individual simply on the basis of that surname.  In this version of the blog, I am going to concentrate on one group only to show it is possible to look for similarities but that what STRs show us can sometimes be misleading because of the fact that mutations are random. In other words, it is possible for two individuals to have the same mutations but not share a common ancestor and this has proved that SNP testing is vital in order to find and verify connections which are suggested by the STRs. But, we have to be prepared for these apparent connections to be disproved, or thrown into doubt, by the more detailed results that SNPs give. Also the time frame within which shared SNP results operates can vary from 200 to perhaps 1600 years.

 

I am going to concentrate on the MACGREGOR distant group – partly because with so many people in it one would expect that there will be some shared origins and, because of dating predictions given by the SNP analysis, we can attempt to date when two people shared the common ancestor.

 

Before I begin, I just want to reassert some facts about the nature of clanship – that is the idea of belonging to a clan. There is a misconception that all members of a clan descend from a single individual – and more precisely an individual of the name. On the contrary, the thinking behind clanship is that people with similar names are united as a single clan - so individuals named Gregor and Gregory for example, despite having multiple genetic origins are all considered part of Clan Gregor. There is also an idea current that all people in a clan MUST descend from a common ancestor, which in Clan Gregor's case would be the Argyllshire branch. This is not the case. For example, the Perthshire Gregors/Grigors who became McGregors are genetically different in origin from, and therefore not related to, the Argyllshire MacGregors. They are members of Clan Gregor as much as the Argyllshire MacGregors are. If name origins are correct the name Gregor might well  mean something like 'watcher', or ‘watchful, vigilant’ and is believed to come from the Greek ‘gregorein’ 'to be watchful’.  

 

A M(a)cGregor who has Viking genetics is still clearly a member of the clan – and this genetic origin could have come about through adoption, illegitimacy or the fact that the surname was simply assumed at some point in the past. In Highland society the 'Common Man’ for want of a better way of putting it, might well have been living on the land as a tenant of some landed individual, and was only called by what is known as patronymics – so John whose father was Donald whose grandfather was Peter would be known as John MacDonald VicPatrick  or in Gaelic Iain mac Dhòmhnaill mhic Pheadair. No need for a surname. However, as populations expanded and individuals became more settled on the land, it was important to the landowners to be able to rely on tenant’s support – in the best case this was a symbiotic relationship – but not always. One way to ensure loyalty was for individuals to adopt the landowner’s surname.

 

The MACGREGOR distant group includes individuals from a wide variety of backgrounds.  This statement could be applied to any surname group in the DNA project.  Because of the limitations that STRs have, I have had to omit any results which had 25 or less STRs tested. There are therefore 101 others who remain, of whom 38 have done either Y500 or Y700 and surprisingly out of those only four pairs of individuals match each other closely as far as is indicated by terminal SNPs.  Of the 101 people with 37 markers or above 77 have 67 markers and 59 have 111 markers.  Using Dean McGee’s Y Comparison Utility and Splitstree I have created two spider charts, one with the 37 results, and one with the 111.  I have underlined some of the same  kit numbers in the 37 as are in the 111, and you will see that some results look to be closer in 37 then they actually are when you look at 111.  In 37, kit 1006596 seems to be the origin point of a number of different branches, whereas in 111 it is just another branch with a possible common origin (on both charts it is on the left).

 

These charts are based on probability of 75%, with the FTDNA mutation rate 0.004 etc, and 30 years per generation.

                                   MacGregor Distant 37 markers                                                               

                                   MacGregor Distant 111 markers                                                                    

Using Dean’s comparison grid the group in the bottom left-hand corner containing 98616 to 126138 show as being possibly related to each other (that is, have a common ancestor) between 120 and 210 years  ago. The only other group which suggests even a degree of recent relationship is the one containing 941169 and M135332 but the estimated time distances vary between 600 and 1710 years [this is the group at the very top of the chart].  Occasional pairs show some relationship: so, 448450 and IN22242 are suggested to have a common ancestor 450 years before, IN132077 and B908902 within 390 years, I99742 and 395944 the same, and 430122 and 245917 of 300 years. A closer link of 120 years is suggested for 404828 and 170627, between 60 and 120 years for 941518, 153532 and 164088, and finally 270 years between 981710 and IN124384.

 

All other shared ancestors are suggested to be between 1440 and 5850 years ago.

 

What we will do now is look at these possible relationships in detail to see if the SNP results confirm or contradict what is suggested by STRs.

 

The above groups and pairs will now be dealt with in sequence.

 

1)     185487, 173181, 126138, 667833, 200914, and 981616

In this group unfortunately only one person has undertaken Y700 and shows an ancestral split around 1400CE . One other participant suggests that this group may all connect to William McGregor the early 18^th century preacher in America. I would expect that all the others would be related to that person – either ancestral or descendant from, but we would need at least another Y700 result to confirm this.

 

2)    94589, IN109971,135116, 28296, 941169, 395944, 43065, MI35332

This group shows a variety of M(a)cGregor ancestors, not obviously related, and one McGown. Five of these have done Y700 and the terminal SNPs are BY69722, FT213382, FT47815, FT371661 and S7361.  Almost all of these are connected with at least one other result at  between 400 and 450CE  or 1600 years ago according to the FTDNA dating. 

 

Kit number 94589 McGregor actually shares a common ancestor at c550CE, by FTDNA reckonings, with 38516 McLaren – again this connection would have happened pre surnames. The same applies to 135116 Peddie with 28296 McComas McGregor who both have later SNP splits, but who share a common ancestor at S744 around about 400CE to 450CE.

 

3)    448450 and IN122242

Both these individuals share SNP FTF57555. This mutation arose around 1500CE but since both have ancestors called (christian name)  Angus with slightly different dates of birth (and apparently a close geographical connection) it is clearly probable that they are related to the same ancestor one or two generations back from what they currently have found genealogically. The 111 mutation chart shows these two lines as joining to form one.

 

4)    IN132077 and B908902 (and 138485 not shown on the chart)

 

IN132077 and 138485 share a common ancestor (or, rather, share a mutation FTF403 borne by a common ancestor) at c1450CE while B908902 descends from that original by a mutation which happened about 1600CE (SNP FTE86294).  One of these McGregors comes from a group associated with Gregor/McGregor families round Perth (138485) so it would be a good idea for IN132077 and B908902 to investigate this area for a genetic connection. I have now moved both these results from the McGregor distant section to the (Mc)Gregor  group which heads the results grid.

 

5)    IN99742 and 395944

Kit 395944 actually shares an origin at 550CE with IN99742. The Splitstree programme puts these results in separate areas  which suggests that both individuals, although having the McGregor name are accidentally related through surname choice only (since their shared SNP is c550CE). However, it would be worth doing more genealogical searching since looking at their earliest ancestors, these could be in fact be brothers or direct cousins and they clearly would share the same ancestral SNP no matter how distant. The only key difference in their STR results is the fact that DY385 (one of the early DNA results) has a significant and unusual one step difference, but that has been enough to put the two results into separate lines. If the two individuals involved would make contact, we can try to find if the connection is much closer than suggested by DNA.

 

6)    430122 and 245917

Neither of these individuals has done a Y700 test so it is not possible to compare them – in addition 245917 has not detailed earliest known ancestor. The 111 chart suggests that they are connected genetically at some point in the more recent past. More DNA information and genealogical information is needed.

 

7)    404828 and 170627

As shown by their earliest known ancestor entries these two individuals have a common ancestor born in 1802. Not surprisingly they have SNP FT87169

 

8)    941518, 153532 and 164088

These three are all descended from the same individual – a William McGregor who died in 1815 in North Carolina. They have SNP FTF1629 – this SNP was estimated to have split from MF409 around 1800 which suggests that it was actually William himself who acquired the mutation from his father whose mutation MP409 arose around 1600CE

 

9)    981710 and IN124384

These two individuals are probably quite closely related but one (981710) has not tested to Y700.  The common ancestor dates back to 950CE but from that ancestor the SNPs go back in time to M222 (arose c50BCE) which probably has its origins in Ireland.

 

It seems that, for the most part, when the the 111 SNP grid is used for comparisons it produces the correct connections between kits, but not100% of the time.

 

The process I have used here could be used with any of the surname groups with diverse entries so long as sufficient numbers of individuals have done 111 markers and preferably Y700. As usual, if anyone would like me to run a similar exercise with entries from a name group then please provide 10-12 entries you want to compare with. Please remember that it needs ideally  to be individuals who have done the Y700 test.

 

Contact me at richardmcgregor1ATyahoo.co.uk, substituting @ for AT.

 

With thanks as always to the authors of Splitstree, and Dean McGee.

 

              MacGregor DNA Project  Blog update 2025 

It has been a rather quiet year for DNA developments. In terms of the tests which have been available over the past couple of years there is nothing new. Y700 still remains really the best Y chromosome test around but only males can do it, and it relates directly to male surnames. What has been developed by FamilytreeDNA are tools for understanding a person’s male genetic past on the Y chromosome using SNPs (rather than the ‘usual’ STR results). STRs or ‘short tandem repeats’ are necessarily less accurate for genealogical matching because they mutate randomly – in some individuals hardly at all, while in others, mutations occur with greater frequency (reasons evinced have been age of father, diet, exposure to radiation etc).  

    However, as I will explore in a moment with an example in this blog, it seems that the relatively small genetic pool of results from those with male ancestry in the Highlands of Scotland means that some individuals have a large number of ‘private’ SNPs in their results – in other words nobody tested so far has a SNP that is shared which would make it no longer ‘private’. Familytreedna has developed a tool to use when two or more individuals share a common SNP, which gives an approximate date for when that SNP was shared. Last year I explored how this gives a dating sequence for family groups of those called MacGregor (or alias names) who descended from the Argyllshire MacGregors. SNPs (or Single Nucleotide Polymorphism) are points on the Y chromosome that occur and stay fixed (as is believed). 

    This year I was contacted separately by two individuals, named MacGregor, who currently share the same terminal SNP, R-BY3023. They are not related to the Argyllshire MacGregors, or indeed to the other main MacGregor groups (that is, from Rosshire, from Ireland and from near Perth). They do share that SNP with quite a few other individuals called Campbell, MacDonald, Cameron but most particularly, MacKenzie. 

     While one of the two MacGregors has a terminal at BY3023 the other is currently waiting for results of BigY which should prove interesting to see if he comes out with a closer match to the MacKenzies. 

    But that is where the problem arises and in part interpretation is severely handicapped by the incomplete (and sometimes non-existent parish registers). And that’s also where it gets interesting:

    Taking one of the family genealogies I was sent, the nineteenth century line, thanks to census and statutory records, is well established, so that there is direct evidence for Hugh born 1822/3 – he gives his age consistently in the census records and when he dies in 1881 his age is given as 58. And, he is also consistent in claiming that he was born in Kilmallie parish, Invernesshire. However, there is no record of a Hugh being born in 1822/3 at all let alone to Angus McGregor and Catherine Cameron who are named on his death certificate. 

    What records there are for Angus in Kilmallie are as follows (I have included some other entries for what comes next):

1773 Angus/Mary Cameron John, Kinlochleven, 8 June 

1774 Duncan/Florence Cameron Ann, Achintore, 5 July 

1776 Donald/Sarah Cameron Marion, Inchrie, 28 Apr 

1782 Ewen/Sarah Boyd Janet, Blarachern, 17 Mar 

1783 Ewen/Ann Boyd Rachel, Blarachern, 16 Nov 

1785 Donald/Sarah Cameron Donald, Inchrie, 20 Jan 

1785 Angus/Sarah McMillan Flory, Blarmacfeltach, 19 Apr 

1794 Harry/Betty Wilson Elspet, Maryburgh, 30 Aug 

1797 Harry/Beatrice Wilson Robert, Maryburgh, 17 June

1800 Harry/Beatrice Wilson Beatrice, Maryburgh, 13 Jul 

1803 Harry/Beatrice Wilson Archibald, Maryburgh, 30 May 

1806 Harry/Eliza Wilson Catharine, Maryburgh, 9 Mar 

1810 Ewen/Rachel McMillan Catharine, Blaraclurin, 23 Jun 

1813 Ewen/Rachel McMillan Ann, Blarnaclerach [Blarnaderach] 28 Aug 

1818 Angus/Catherine Cameron Ewen, Blarmacfoldach, f 8 Nov 

1823 Angus/Catharine McGregor,[ - ], John, 5 Dec 

1825 Angus/Catharine Robertson, Rachel, Blachorin [dist Fort William], 25/26 Jun 

1827 Angus/Catharine Robertson Sarah, Badacheanan Corran, 18 Aug/8 Sep 

1829 Ewen/Ann Margaret 19/19 May 

1829 Angus/Catharine Robertson Ewen, Blarchoachan, [Corran Dist] 16 Apr/17 May 

1834 Angus/Catharine Robertson Catharine, Blarmachfailan , 21 Mar/6 Apr 

1839 Angus/Catharine Robertson Susan, Blarmacfoldach, 23 Jul/8 Sep 

 Key: father/mother child, place, one date = baptism, 2 dates = birth/baptism, f = in fornication

There is a tombstone inscription in the Fort William Episcopal Churchyard: In loving memory of Angus McGregor and wife Catherine Cameron, this tombstone is erected by direction of their son Donald McGregor Achnacarry who died Banavie June 11th and was buried here June 13th, 1892 

 

FORT WILLIAM [gravestone from GD50/233 in National Archives of Scotland] Henry (McGregor) imo son Charles died 25 May 1811 age 12y 9mo 

 

In the census record there are some entries that show an Angus: 

 1. 1841 

 Angus McGregor born c 1786 (possibly 1782-86 because of census rounding down), wife Catherine (40), son Donald, 14, daughter Sarah 8, at Blarmacfoldich 

2. And: in Fort Wiliam [Maryburgh] 

Angus McGregor age born 1791 (1788-91), wife Ann age 60, grandson? Peter a 5 

3. And: 

Angus MacGregor age 85 (1752-56), Ann (daughter?) age 30 

4. And: 

Catherine MacGregor age 35 (born 1802-06), Ewen age 10, daughter Annabell age 8, daughter Catherine age 6, son Donald age 4 at Blarmacfoldich 

5.In the 1851 census 

Angus MacGregor born c1783 woollen weaver Fort William, wife Ann born c1776 

6. And: 

Angus McGregor born c1779 widower, farmer Muirsheorlich, children, Donald 23, daughter Sarah, 20, son James 13 

7. And: Angus McGrigor born c1791, wife Catherine age 49 (born 1802), son Ewen age 21, son Donald age 14 and son Alexander age 6 at Blarmacfoldich 

1 and 6 go together, 2 and 5, and 4 and 7 (3 is clearly dead by 1851). But there is no Hugh aged c17/18 (rounded down to 15) in 1841 or aged 27/28 in 1851 

 There is a death record for Angus in 1863, a farmer? aged 82 (therefore born c1781) with parents given as Donald McGregor and Sarah Cameron, and he is the widower of Catherine Cameron – the informant being Donald McGregor, son at Auchintore. So, that ties up with the tombstone inscription, and therefore 1 and 6 in the census. Is this the same Donald who had a son Hugh in 1823? It doesn’t look like the the Ewen aged 10 is the same as Hugh as heshould be given as at least 15. 

   However, if the wife in 4 and 7 is Catherine Cameron that suggests that the wife for 1 and 6 is Catherine Robertson. 

   Look though at the names in the parish register – Sarah Boyd becomes Ann Boyd, Sarah Boyd becomes Ann Boyd and Betty Wilson becomes Beatrice: you can’t but get the feeling that whoever recorded wife’s names wasn’t too accurate. So, you wonder if John son of Angus and Catherine McGregor in 1823 was actually a child Hugh by Catherine Robertson. OR is that Catherine Cameron again?? OR is it Ewen because that is interchangeable, apparently, with Hugh? 

    I have included Harry in the list above because according to Joseph Foss he is related in some way (but again it all depends on the DNA connection! 

    I should say that there is a letter from Duncan in Blarmacfoldich in 1825 in the PD60 Central Archives deposit saying that his brother Angus is bearing a letter of recommendation from John Gregorson of Ardtornish and that the family had been farming in that area since Prescription - so that suggests that they were there a while before Rob Roy and his son James Mhor. The significance for this is that it is claimed that Donald father of Angus was the last child of James Mhor son of Rob Roy. That would be doubtful if they had been farming there before Rob Roy. And where is Hugh – before he turns up in 1851 in Ballantrae Ayrshire aged 29 married with wife Elizabeth aged 22, son Angus 2 and son John aged 1 month? 

   The second family who are related through the early SNP BY-3023 (estimated to have arisen about 250CE [used to be labelled AD] may provide a clue. They also descend from an Angus – this Angus died in 1866 at the age of 68 according to his son, but on his marriage certificate (2nd marriage) in 1855 he says he is 50 and then in the censuses he gives his age as 45 in 1851 and 59 in 1861 so clearly didn’t actually know his age. But, presumably, when he said his father was Duncan, (a mason) and mother Margaret Fraser he knew that for a fact. We can find Duncan in the parish registers. Duncan was actually for a time in “Locheil’s Highland Regiment of Foot” and there are children born to them: Catherine in Paisley (however we cant be sure if this the same couple as he is a ‘threadmaker’ according to the record), a daughter (unnamed) born in 1799 in Boleskine and Abertarff , Invernesshire [bottom end of Loch Ness] (the father was ‘not near’ so the child required a sponsor), and then in 1805 the same couple (he is still a solder) had a son Duncan in Edinburgh. Angus’s son knew that his father’s father had been a soldier, but he said that Angus’s father was Angus not Duncan (however he also says the regiment was the 78th so that could be checked). Duncan was no doubt from the Highlands and since Lochiel’s lands include the Fort William area it seems likely that Duncan came from the same parish and family as the family of Hugh that was discussed above, but the parish records are incomplete, and as you can see what someone’s children thought their grandparents were called was not always accurate. 

    I know this was a long diversion but I wanted to show what the steps were to trying to find out the facts and how DNA can be used to confirm or deny a story, If you have followed the thread so far you will have guessed that Angus who married Catherine Cameron (1 and 6) was the son of Donald McGregor and Sarah Cameron but not entered in the parish record. The significance of this, as I noted earlier, is that Joseph Foss maintained that Donald was a son of James Mohr MacGregor, son of Rob Roy born in 1753 – more or less at the time James fled to France where he died in 1754. So, the DNA for Hugh son of Angus McGregor and Catherine Cameron would not necessarily show any connection to the Argyllshire MacGregors (always assuming Rob Roy’s line was legitimate) and the confusion of males with the same christian name creates a problem. In other words Donald born 1753?was not a child of James Mohr (unless Rob Roy’s line did not descend from the Argyllshire MacGregors. Of course, there is one little fly in the ointment – that Hugh (not entered in the parish record) had a sister called Annabell (also not entered) – which of course was the name of James Mohr’s wife (Annabell McNicol). Wouldn’t it be good if genealogy was straightforward.? So we are still looking for a probable male MacGregor descendant of Rob Roy!!

For the second part of this blog, I wanted to deal briefly with the Greig, Gregor, and Gregory families (no matter how spelt) to see if there was any overlap now that a number of BigY SNP results are available. This process involved downloading and combining all the Y chromosome 67 STR marker data from the FamilytreeDNA results page for the several surnames and editing it so that only the kit numbers appeared. I then ran the number sequence through Dean McGee’s Y DNA tool to produce a grid which shows the apparent relative closeness of some family groups by the colour matching – note the relationship time is not necessarily correct as I asked for a high percentage of probability. The resulting grid is Fig, 1. I say ‘apparent’ because this seems to show the relative closeness of families with different surnames, but the dates start to go back into the past when the closest relatives of the same name are ignored. There would have been less colour if I had opted for a lower percentage of probable time when individuals were related.

Fig 1 name relationship grid. From these, 6 more related groups can be seen – the largest being Gregories.

A well as this, Dean’s programme produced a Phylip grid – effectively the above in another form  -  and  that was uploaded into Daniel Huson and David Bryant’s program Splitstree (as usual)  which produced a spider diagram of  results as below (Fig. 2):

Fig 2:  raw chart of kits for the various surnames

At this point I inserted the surnames attached to each kit number and then added in red the BigY terminal SNPs (Fig 3).

Fig 3: the grid name and Big Y terminal SNP data

From this you can now see the various family groups and more particularly the high incidence of relationship among some members who share a surname, BUT, also  that others of that same surname are much less closely related. The SNP results suggest the common ancestor date, but again this needs to be treated with caution because the connection date could be a lot further back in  historic time than it appears. It does seem that ARE some branches which produced different surnames (for example Gregg and Gregory), but their common root is far back along the line, potentially pre surnames. Other than this there does not seem presently to be a direct link between or among these different surnames.

     As always, this is just my interpretation, but do feel free to contact me at richardmcgregor1ATyahoo.co.uk (substitute @ for AT). My thanks to Dean McGee, and the creators of Splitstree.

MacGregor DNA Project update 2024

This year’s DNA update is rather later than usual, partly because it’s not been a year where very much new has happened so it is difficult to know what might be of interest in a blog like this. I have previously discussed aspects of Scottish clanship and how what we see in the DNA reflects the diversity of origins for clans and their associated names. 

     I will begin by repeating an important fact about clans and kinship. Clans were never derived from a single surname; they were always collections of related names and septs. That means we should not expect everybody who is associated with the clan to have the same genetic profile. In many cases, we have to remember that surnames were adopted long after they first appeared, particularly in Scotland because individuals who were less well placed in society, and therefore did not sign documents or make legal agreements, often were unable to write their names. 

      It is true that in Scotland there was a move in the 18th century to provide at least an elementary education for all children, and in some cases children from very poor backgrounds who had particular abilities were not then confined to their social class. Much of the early work in education was done by the SSCPK (Scottish Society for the Propagation of Christian Knowledge)

The Society began to establish schools in the Highlands with the aim of reducing Jacobitism and resisting the rise of Roman Catholicism. The first school was opened on St Kilda in 1711. By the end of that year, the SSPCK had five schools, by 1715 twenty-five, by 1758 176, and by 1808 189, by which time 13,000 children were attending the schools [Wikipedia]

 In the Highlands particularly, surnames were not so commonly used, particularly where individuals were simply known by their patronymics - for males this was Christian name followed by son of (Mc) father’s Christian name followed by grandson of (vic) grandfather’s Christian name. It’s not surprising, then that when it came to adopting a surname, individuals chose surnames that were local, influential, or had some connection to them – it might even be the landowner’s name. In the next section of this blog, I will discuss a new dating mechanism which is attached to Big Y results. I am using the SNP data from the main MacGregor bloodline because there are sufficient results within it to begin to try and begin to make sense of family connections going back into historic time. This therefore is a principle which can be applied to other groups of surnames both within and beyond the project.

     What Big Y has shown is that clan groups are made up of many many lines which split off from each other over thousands of years. One of the latest developments which FtDNA offer is ‘Discover (Your Haplogroup story)’. This option gives you the ability to track when your terminal SNP (given by Big Y) [terminal here means the SNP which is identified as most recent in historical time to you] is estimated to have occurred, and then to track SNPs back in time to see roughly when the line split to form a new line and then back from that into more distant history. 

     I will say more about the dating offered by FtDNA shortly but for the moment here is an example. This individual’s terminal SNP is BY74650 estimated to have split off from S690 in 1300CE and the individual’s ancestor was ‘born’ around 1400CE (common era – that is 600+ years ago) and two individuals on current testing have that terminal SNP. The sequence back in time then goes as follows: S690 split from BY144 around 1050 with 50 individuals from that line. [comment: this is possibly misleading as it suggests that people with S690 should be identifiable who are NOT called MacGregor, [Gregor the genetic origin of the clan bloodline probably WAS born about 1300], but to date NO other individuals with S690 and different surnames [other than those adopted as aliases during proscription] have been identified. Also misleading is the ‘50 individuals’ bearing S690 whereas in fact in the DNA project everybody called MacGregor who is in the main bloodline group has that SNP, and that is over 200 individuals. It suggests that the 50 individuals noted by FtDNA are those who still have S690 as their terminal SNP). Going back in time from S690: 

BY144 split from S703 at around 750CE (55 individuals) 

S703 split from FGC 59322 around 550CE (55 individuals) 

FGC59322 split from S695 around 500CE (56 individuals) 

S695 split from S691 around 450CE (148 individuals) 

S691 split from S744 around 400CE (251 individuals) 

S744 split from Z-16325 around 350CE (802 individuals) 

Z-16325 split from L1065 in around 150CE 

Without going into detail, it’s clear that there are anomalies and approximations here. Another company, Yfull, also does dating estimates – but not for all the later SNPs – their date for S695 occurring is about 600CE and most recent ancestor 650CE, while for BY144 occurring about 650CE and most recent ancestor 1150CE. If you examine the detail on BY144 which these estimates are made at YFull you see that the range is given as between 1220 and 1447 years ago therefore the estimate is given as 1333 +30 years = 1363 years ago (there is a formula which is given on the website, so my explanation is somewhat simplistic). Anyway, 1363 years from 2023 gives 660CE as the presumed date of origin. The map given is strongly UK located with much fainter occurrence in Scandinavia [https://www.yfull.com/tree/R-S695/].  

     The Yfull method for estimating time to most recent ancestor on the other hand is much more straightforward and is based on three estimates added together and divided by 3 – however again the formula for achieving the estimates is much more detailed. However, the estimate for TMRCA is 851 years before present – that is, 1172CE. Whereas the mechanism in FtDNA is not apparent, Yfull is much clearer as to how approximations have been calculated. 

     For the SNPs they do cover (and S690 is not one of them) the mapping component is helpful, but not recent enough. A website which maps more recent SNPs is www.genetichomeland.com. To view maps for a SNP you must register but registration is free, and you can store up to 10 maps. Note that Google Chrome seems throw up an ‘unavailable for legal reasons’ problem so a non-Google browser is better (Firefox worked best for me). 

     You will notice that the time estimates given by FtDNA are always rounded up or down to 50 or 100 years. The point is that these are truly estimates of when a split occurred. So, it’s important NOT to take the dates as absolute, and particularly when trying to relate two or more individuals together. A gap of 50 years between SNPs could therefore be anything from 1 to 3 generations (or more). 

     I have drawn out a terminal SNP chart using the approximated years given by FtDNA to show how some of the DNA lines are connected to each other. This is particularly true of the line:

 

FGC17836 to FGC17835 to  FT203761 to MF133765 

which, using the earliest ancestor information given looks as if it might relate to the area north from the top of Loch Tay – those 2 testers who claim descent from John who died in Ontario might look to that area for origins.

It is interesting to note how many of the different genetic lines have an alias associated with them, reflecting the clan’s turbulent history. It might be obvious to say this but the point where the SNP divides is not, particularly in the historic SNPs, to be thought of as a place where historic divisions of MacGregor families (such as Roro and Glengyle [Dugal Ciar] took place. They are simply the point where a DNA mutation took place. This is well illustrated with the only division that produced two subsequent branches here, namely S696 and FT140264 whose shared ancestor had mutation A14374 but currently no one in the project has retained that mutation as their terminal SNP. In fact, this happens all the time in different family groups - whenever a division occurs, that indicates a shared ancestor. And in some cases, like the Stirlings, that is apparently a very recent shared ancestor. 

     No significance should be drawn from mutation happening very early on historically, nor the kit tester having that alias (e.g. Pressley), since almost all the aliases will have been taken post proscription of the MacGregor name. In towns before that there is evidence of individuals dropping the Highland ‘Mac’ because Highlanders were not particularly welcome – so we have Johnsons as an alias before 1600 as in these two entries: 

 “1575. June 25. Advocatis against McGregor and others, George McGregour alias Johnstoun, Johne and George McGregor his sones, Duncan McGregor & Burgesses of Perth to have broken Lawborrowis5. Delayed till 15 Jan. next. “1576. June. 2. Charter and infeftment given to Gregour McGregour alias Johnstoune, burgis of Perth, by Dene Adam Forman Prior of the Charterhouse of Perth. is rescinded by the Court of session in so far as regards the sunny half of three quarters of St. Leonards. Lee. - Decreets of Court of Session. 

     It is worth noting that the alias Drummond is in 2 different generations (see BY39563 line) – one might hope it was the same family, but Drummond was rather a common alias for different branches of the MacGregors to take whereas Pressley came as rather a surprise. 

    Also, I have noticed that Ireland comes up several times when considering aliases – most notably the Boyds. What is rather unusual about the Boyds is that they split into three distinct family groups (split from S690 that is) at about the same time – presumably between 1600 and 1700 as a result of proscription and relocating to Ireland. In the book Illustrations, Historical and Genealogical of King James’s Army Lists there is the following rather fascinating entry which may give some clues to the Boyds who are MacGregors: 

 ‘…about this time an Alexander McDonnell was appointed by Lord Tyrconnel Governor of Galway; he, in the progress of the war, became a Brigadier-General. Colonel O'Kelly, in his Excidium Macario, says he was a "soldier of fortune, raised by merit from the ranks;" and Croker, in his brief notes on that little work, adds that he was otherwise called. 'McGregor,' and was of Drumsna, County of Leitrim. He married in 1685 the Lady Jane Nugent, a sister of Thomas Nugent, afterwards created Lord Riverston. In December 1690, he was removed from the Government of Galway. It is remarkable that in the Outlawries of 1691 he is styled Alexander McDonnell, alias Gregor, alias Boyde, of Clonin, County of Westmeath. At the same time were attainted in Antrim, Daniel Mc Donnell of Dunluce…’ 

   A similar process to the one done here can be applied to other surname groups, as long as there are sufficient entries from BigY tests. For example, many of the Greigs share a common Y line when the SNPs are related to each other. Anyone can check the SNP sequence of mutations to find out if the DNA results point to a common ancestor. Simply put the SNP label and ‘DNA’ into Google and select the resulting ‘FtDNA Discover’ link. The most significant Greig line mentioned is: 

BY135575 [c.900CE] to FT108241 [c.1550CE] to BY212591 [c.1600]  to FT106941 [c.1750] to BY166266 [c.1900]. 

 As always if you find a problem with doing this process you can contact me at richardmcgregor1Atyahoo.co.uk [substitute @ for AT]

MacGregor DNA Project 2023

Welcome to the MacGregor DNA Project blog for 2023. I began last year’s DNA blog looking at how SNP results could be used to understand historical connections between surname groups and how different family groupings are suggested by subgroupings of the results. Over the past year, there has not been considerable development in terms of DNA testing other than the fact that individuals have realised the usefulness of Y700 for identifying closer family relationships. I am going to use the initial part of this blog to explore further this idea based on the fact that family tree DNA have introduced a new feature which uses scientific tools to explore potential connections between participants. I am taking examples taken from the MacGregor project and based on the main Argyllshire line as we already have some traditional sources which suggest when family groups split off from the main branch.

     I want to begin with an example which appears to show a family connection within the last 250 years, and which suggested that individuals that appeared in two family trees could in fact be brothers – which in turn led to a search in parish registers for an appropriate family living in the area from where it was believed these individuals lived.  In this ftdna example our 2 participants are suggested to have close familial connection dating to approximately 1802. As it happens that is slightly too late as we appear to be dealing with two brothers but given the proximity to 1800 of the averaging process it was determined that these two participants ancestors were more than likely related. 

 Fig 1: two closely genetically related individuals – with SNP BY23340.

 

It seems that BY23340 is a relatively modern terminal SNP as can be seen in this graphic analysis:

Fig 2a: graphic representation of the DNA relationship between two individuals and Fig 2b the Mean which suggests a date of 1802.

 

In Fig 3. we see how this is projected forward:

 Fig 3: How this translates into two related lines today.

 

This parish register research suggested that two brothers James and Robert were the children of Alexander who lived in Stevenston, Ayrshire  in the second half of the 18^th century.

          SNP – R-S696

In the MacGregor results one SNP, which was identified almost a decade ago by Jim Walker,  has significance as it is borne by the Chief’s family and two others (both related to emigrants to America whose earlier histories are not recorded).  The SNP is labelled R-S696.

  Fig 4: The genetic descendant to S696

 

The graphic in Fig 4 is slightly misleading as it only relates two SNPs whereas in fact R-A14374 is a descendant of S690 and we already know that S690 arose about 1300-1350. The suggested split of A14374 sometime in the early to mid 1500s seems very likely, and that S696 did indeed happen not too long after the banning (proscription) of the MacGregor name following the Massacre of Glenfruin in 1603.  The fact that two of the participants with S696 ended up in America but with no known genealogies available suggests that they may have gone there during the time of proscription to escape the personal danger and legal difficulties that were in force at that time. Proscription lasted from 1603 to 1660 when Charles II became King after the Commonwealth collapsed. Proscription was reimposed in 1692 following the unsuccessful attempt to restore the catholic King James II to the throne since the MacGregors were largely supporters of the Stuart cause. The second period of proscription was principally legal in nature but lasted through to November 1774. 

 Fig 5:  other lines related to S696. 

 

You can see from this diagram that there is a younger line (FT140264) from the same descent. One of these participants has a genealogy suggesting a Glengyle MacGregor connection which has been discussed in the last three blogs (Rob Roy MacGregor was a Glengyle). In summary: there was a gap in the documentary evidence at c1800 which threw  doubt on that family connection and that instead this participant comes from the main MacGregor family (that is, of Brackley)

     One final example will show, I hope, how accurate this scientific predicting could turn out to be. It is well known within MacGregor genealogies that the main line split into 4 main houses – Glengyle or more correctly Dugall Keir (as later epitomised by Rob Roy), as mentioned, was one, MacGregor (main line – Brackley) another, Roro (based on a location in Glenlyon near Fortingall in Perthshire) was a third and the family (Sleik [sliochd = tribe]).  of Gregor vcIan.

     The family based in Roro were known to be there from about 1460, so this Time Tree for SNP R-BY54364, carried by a descendant who has a good claim to have Roro descent, is therefore suggestive that the claim is accurate.

 Figs 6a/b: the genetic split that suggests the origin of the Roro family of MacGregors

 

The above discussion of SNP connections is clearly still somewhat speculative, but, since we know that S690 split off around 1300, or slightly later, leading to the establishment of the MacGregor main bloodline, then the results do begin to help us identify family groupings post 1300. We can see that so far there are seventeen family lines currently identified, some reaching back to the late Medieval/early Renaissance, and others (as in Fig 1) much later. This shows the importance of as many individuals undertaking the Y700 test, particularly those who have more comprehensive genealogies, so that more inter-family relationships can be established.  To end this section here is the current family descent tree for S690 as revealed by Y700.

Fig 7:  MacGregor Y700 test results as at December 2022

 

For the second half of this blog I want to go back to questions I am often asked:

 

1)    What is a clan?

Six hundred years ago this question was quite simple to answer. You would have been born with the name or you would have adopted it, rather than being called by a patronymic (e.g. John MacPatrick vicDonald = John son of Patrick son of Donald).  The name itself then becomes complicated with non-standardised spelling and attempts to render Gaelic into English which meant that, in the MacGregors’ case, the word might be expressed as Gregor, Grigor, MacGregor, McGrigor, McGregor with a whole range of alternative spellings. There were also accepted variants, such as Grierson, or Grier, Greig/Grieg/Grig etc. These were understood to be shortened (or ‘anglicised’) 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 the people bearing these accepted names were genetically related to the main family line was not the point, since the clan was composed of members using a collection of related surnames who recognised, as Chief, the head of the main line (the Chief of the MacGregors for example).

     The time of Proscription (when using the name MacGregor was banned), also caused a great many different surnames to be associated with Clan Gregor, many as the result of the adoption of an alias (such as Stirling, Drummond, Bain, Black, Campbell – hence Rob Roy’s alias of Robert Campbell).

 

2)    What is a sept?

Other clan members bearing different names were from septs or associated groups of the clan. Sometimes the same name can appear on several lists of accepted septs for different clans.  This is the case for the surname King, for example: for MacGregors this surname was certainly thought to connect to the clan motto ‘Royal is my Race’.

 

3)    We have male MacGregors in our tree, how can they test?

Go to www.familytreedna.com, use the surname search box for ‘MacGregor’ and follow hotlinks on that name until you get to the project join page then select Y chromosome testing for one of the males – do at least 37 markers or any number of markers above this number.

 

4)    Why are there no family grids/list of results for Ancestry/Family Finder results and can I not prove my genetic connection to MacGregors that way and appear on the list?

 

From here I am largely repeating some text from my 2017 blog.

 

There has been a noticeable increase in the number of individuals taking the test known as ‘Family Finder’, or something similar, rather than Y chromosome or mtDNA tests. This has probably been as a result of quite aggressive marketing by Ancestry.com [it has a variety of website endings depending on where it is based] in particular. This has promoted the equal use of DNA testing for both males and females, and tied it into the submission of family trees which individual testers can use to identify the same family name(s) with others who have tested and submitted their genealogies. What has perhaps been rather glossed over in this is the fact, firstly, that DNA gets ‘lost’ over time – if it didn’t, we would have the DNA of billions of ancestors in our bodies, and, secondly, that it is only a tiny portion of our DNA which is currently being examined for genealogical purposes.

     You do not inherit 25% of your autosomal ancestry from each of your 4 grandparents. This is because your autosomal DNA is randomly recombined, and not in equal proportions from each parent, and so the more you go back in time the percentage inherited from people in a particular generation becomes smaller and smaller and therefore the more distant the ancestor is the more difficult it becomes to identify what you received from that person. What then are the chances of that same bit of DNA being preserved from a specific ancestor in yourself and someone else? For example, if your name is, say, Smith, and your male MacGregor ancestor lived 10 generations ago on your mother’s side it is really not feasible with today’s technology to identify that ancestor’s DNA by looking at your DNA today. The tests which are offered by Ancestry, Family Tree DNA etc. only try to identify links to 5/6 generations back. The key thing to remember is that if you and someone else have, say, people called Brown in your trees it does not necessarily mean that you have a recent ancestor in common, or indeed that you have any ancestor called Brown in common at all. For these tests of connection to work properly, you, and the person you are comparing with, need to have as much genealogical information as possible on every ancestral line in your respective trees, going back 5 or 6 generations (and that you both have a significant shared portion of DNA). This test works best for up to third cousin. You have 4 grandparents, 8 great grandparents and 16 great great grandparents and beyond that it is usually very difficult to have all ancestral lines documented.

     So, you and the person you are comparing with both have to put all your family information into one of the computer genealogy programs (like Family Tree Maker, Reunion etc.), save it as a GEDCOM file, and upload it to whichever DNA company with which you have both tested your autosomal DNA. The possible links between the two family trees are then highlighted in some way which allows a comparison of ancestry to be made in order to explore if there is a family match on some surname. As /I said in the previous paragraph the fact that there is a match on surname does not necessarily mean that it is the same family, only that there is a surname in common. Clearly the more unusual the surname, the more likely that the match will be with the same family.

    I give an example from my own ancestry. Using the compare options on the top of the page I find this result:

   Fig 8:  comparison results for FMS (full mitochondrial sequence) and X-match.

 

For x-match the x means ‘not a match’ and the ‘-‘ means no result. The 9 matches that I do have all have 3 variants except for two individuals who match me either exactly or with one variant.  For obvious reasons I have edited out the names if those who match. This is My mother’s side (and her mtDNA) and unfortunately her family history knowledge was non-existent. 

 Fig 9: a search using Family Finder

 

I have 171 pages of matches, and this is fairly common, but almost all are 3^rd to 5^th cousin or distant, and only one is described as a first cousin to second cousin. In fact I am his second cousin. Interestingly his son is described as my second to fourth cousin, and he is actually my second cousin once removed. None of the other participants show surnames I can relate to, and about half of the participants have given no family genealogy background at all which means that it is impossible to work out the relationship, although in most cases a contact email is available. My experience here is quite typical of what happens with Ancestry-based tests and it would be good if more people who do DNA tests for genealogy could be encouraged to upload trees – or even just a list of ancestral surnames.

     As a project administrator I am very aware of the need to keep participants’ personal details secure and in familytreedna I can only undertake changes if given ‘Advanced Access’ (and the default for familytreedna is ‘Limited’ so I am unable to update most participants’ earliest known ancestor tab which means that the results grid has gaps). I would encourage everyone who reads this to update their information – it helps both the participant and others.

     To end with, I repeat my final paragraph from last year’s blog, which can easily be accessed from the side menu:

     You will have noticed that again I have not used spider diagrams. I now prefer to offer individuals the possibility of seeing a spider diagram (Y chromosome only) for their own result, generated for them alone. You can see the sort of spider diagrams I mean by looking at previous versions of this blog. If you would like me to generate one for you, ask me offline and suggest up to 12 individuals with whom you would wish to be compared by reference to their kit number (for me to be able to do that these 12 need to be in the MacGregor project). You can email me on chairmanATclangregor.org [change AT to @ - I just want to avoid spam!).