h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

I came across a reference recently on "This is Cornwall" indicating that an Oxford Uni study has found differences between typical DNA between Devon and Cornwall.

The article is one of a few stressing the fact that the Welsh and Cornish have the most Brythonic DNA.

http://www.cornwall24.co.uk/discussion/genetics-and-the-cornish-t6698-30.html
http://wessexregionalists.blogspot.com.au/2012/06/devons-difference.html

I have attempted to follow this back to its source, and I found the following site

http://www.peopleofthebritishisles.org/

and then to this more interesting site

http://sse.royalsociety.org/2012/exhibits/genetic-maps/

Interestingly this does show that Cornish DNA can be differentiated from that of Devon, but also shows that Devon DNA differs from that of Dorset, or any compass points further east - which itself shows a great degree of conformity from Dorest to the Humber.

There is a smattering of Devon "green" in eastern cornwall and western somerset.

It also shows that Wales as two distinct types, with a third in the Welsh Marches. The English North East, English North West and Scotland also shown as separate colours.

I do not know what the underlying data says, or what the colours represent - at least not yet.

Taking these colours at face value it does appear that Cornwall is different, but then so is Devon. Further information is clearly needed here.

What the maps do indicate is the Wessex Regionalists may have jumped the gun, and that Devon is (still) different from Wessex. perhaps we need to talk again to our Somerset neighbours.

Ozzie Exile,

These reports looked very odd to me as both Sykes and Oppenheimer (also from Oxford University) examined the Y-chromosome and Mitochondrial DNA from thousands of people across Britain, and came to the conclusion that the people of Cornwall, Devon and West Somerset (i.e. what was Dumnonia) could be grouped together as a genetic group, which was genetically different from those further east in England, but was closely grouped with Tayside and Fife, Borders, and North Wales.

Close examination of the Royal Society genetic maps web site showed questions earlier than your question (which are no longer being displayed), one of which asked whether Y-chromosome and mtDNA were included in the analysis, which resulted in the map. The answer given was that they were not included. This means that only Autosomal DNA and X-chromosome DNA were analysed. As both of these types of DNA are subject to a process known as recombination, they tell you a lot about your close ancestors, but very little about your direct line ancestors going back a thousand or more years. This is explained very well in this link:-

http://learn.genetics.utah.edu/content/extras/molgen/index.html

Only the NRY (or non-recombining part of the Y-chromosome), which traces direct male line ancestry, and mtDNA, which traces direct female line ancestry, are reliable indicators of deep ancestry along these lines. As neither of these types of DNA were analysed to produce the map, the results are unreliable to say the least in determining which parts of Britain have the highest densities of original post-Ice-Age migrants as ancestors of the current population.

Some Autosomal DNA genes do seem to be more prominent in regions where the descendents of early post-Ice-Age migrants are most in evidence, such as the R151C and R160W variants of the MC1R gene, which correlate with a tendency to have ginger hair. These variants reach a maximum in the British Isles in Ireland (31%). In Orkney the figure is 26%, in Wales 21%, in Cornwall 16%, in Devon 23%, while the highest level in the rest of England was found in Sussex and Kent at 13%. Therefore for this particular marker Devon correlates best with Wales and Orkney, and less so with Cornwall and the rest of England.

PlymouthExile,

It appears that the Royal Society is posting answers periodically. A couple were posted recently.

They are not all that helpful however. The last posting simply says that the colours represent common genetic clusters. The red is mostly found in southern and eastern England, the yellow Scotland and Northern Ireland.

Heigh ho! Why did we need a map?

I am curious what they measured and what the interpretation is.

Prof Peter Donnely is attributed in some newspaper articles as saying that the oldest/most celtic samples are found in Wales and (on some sites) in Cornwall, with some pre-celtic elements. This last part may tie into your response.

But which of the Welsh groups? And what does all that mean for Devon?

What is evident is the clear demarkation between Devon (almost all green) and Dorset (red). The only Somerset samples are a few from the west, and they are all green.

Also the map shows pink for NW Wales and the Shetlands, and yellow both in SW Wales and NE England. Are these different genetic clusters and they simply ran out of imaginative colours?

Also, it is difficult to see the symbols. Some boxes seem to have a "handle" on top, some below etc... I have no idea what they mean as yet ( or whether it is just my eyes).


I cannot find the exact source of the Donnely statements, and so I am unsure what he really said (assuming he did).

Hopefully more will become evident in July.

Gere is the Royal Society response to my first question, which was (essentially) what do the colours mean?

"The colours do not mean anything in particular. They just denote a particular genetic cluster. A cluster is a group of individuals that are more similar to each other genetically than to other individuals. There is no indication on this map as to how the clusters are related to each other. When we plot the individual samples on the map by the central position of where their grandparents were born, we get the map we have here. This shows a very high level of concordance of the distribution of the cluster with geography."

I sense that we are dealing with a technician here rather than the organ grinder - or perhaps the scrutiny is yet to be completed.

It simply states that the colours are based on similar genetics.So, this may indicate that Devonians are more closely related to other Devonians (as well as some eastern Cornish folk and western Somerset people) than anywhere else, and (presumably) the difference to other clusters is statistically significant.

Yes, so.....

There seems little connection to the statements that the Welsh and Cornish are more purely Celtic than elsewhere.

We will have to wait I guess, although half of the nations newspapers seem to have jumped at the simple headlines.....

I noticed that the boundary between the green and red markers on the Royal Society website matched very closely to the Devon Dorset border.

It struck me that they matched even closer to the Axe river - the ancient boundary of Dumnonia. There were green markers in Somerset as well, but all in western Somerset. Again - ancient Dumnonia.

There are no markers in eastern Somerset and so it is unclear exactly where this boundary lies, but it could be the Parrett.

This boundary appears even more clearly delineated than the "Cornish/ Devon" mark, as there are a few green markers in eastern Cornwall, and a handful of Cornish brown east of the Tamar.

So I asked the Royal Society what the differences were between the Devon and Cornwall clusters, and pointed out the above similarity to the Dumnonian border on our east - asking that they include it in their analysis.

This is their response

"One way to think about Cornwall, Devon and the rest of England is that Cornwall is more representative of the peopling of the British Isles from the Atlantic facing region of Europe, whilst the extensive red English cluster has a large signature from the Belgium/Denmark/North Germany area. Devon can then be thought of as a zone of mixture between the two. The tight boundaries are very interesting and may well be a result of political spheres of influence"

This does not really address my questions, but for what its worth..

My current view is that the groupings are based on some measure or genetic connection, presumably based on some statistical measure of validity. The Royal Society have not yet loked at the Y-chromosone or mDNA data yet and so it seems the analysis is based on a comparison of 50,000 DNA positions - no doubt a very complex analysis. If statistical measures have driven the anaysis to date then the analysis may well be only starting.

What is interesting is that the groupings do not apply between Dorset and/or points east. If the difference dates back to the period when the Dumnoians eyed the Durotigesacross the Axe, they why aren't there similar boundaries scattered across England. Perhaps with more analysis they will find some, but they are clearly less obvious if they exist at all.

But the difference between Devon and points east are clearly more significant. Why then is Devon (and Cornwall) different? I might conjecture that the Dumnonian border is where the Saxon advance stopped or diluted (from a migration point of view) - and that Saxon influence beyond that boundary was much reduced.

This is only conjecture for now and we must await more detailed findings. However the fact that Devon and Cornwall differ may be similar to the differences between Northand South Wales. Maybe....

Ozzie Exile,

I have always regarded the Bodmer survey (upon which this latest genetic map is based) and his conclusions as being seriously flawed, in that he didn’t start from a clean slate. Instead he made a number of basic assumptions, unlike Oppenheimer and Sykes. Bodmer’s analysis was broadcast as a Channel 4 TV series in 2006 with an accompanying book “Face of Britain”. In the book, it is made clear that he assumed that the Welsh, Cornish and Devonians represented pure Celts and that the East Anglians represented pure Anglo-Saxons, and that he was able to determine admixtures of Celts and Anglo-Saxons elsewhere in Britain by comparing their DNA markers with those in the (supposed) pure Celtic and pure Anglo-Saxon areas.

He did not attempt to date any of the DNA Haplotypes (as Oppenheimer and Sykes did using Y-chromosome and mtDNA samples), so he also assumed that the Celts were the original (post-Ice-Age) inhabitants of Britain, and that there were no other major migrations to Britain until the Anglo-Saxons arrived in the 5th century AD. On the other hand Oppenheimer was able to show, using dating techniques, that most of the migrational influx from northwest Europe occurred long before the Roman period, and that the 5th/6th century Anglo-Saxon influx accounted for no more than 5.5% of the current English population. Furthermore, both Oppenheimer and Sykes concluded that, even in East Anglia, the descendents of the post-Ice-Age migrants still represent a majority (51% according to Sykes). If Oppenheimer’s and Sykes’s analyses are to be believed, then one of Bodmer’s basic assumptions, i.e. that the East Anglians represent pure Anglo-Saxons, falls at the first hurdle.

To cap it all, it now appears that Bodmer did not analyse Y-chromosome and mtDNA samples, which are the only types of DNA that are not subject to recombination at each generation. For this reason, they are the only types of DNA which have essentially remained unchanged (apart from minor but dateable mutations), and are therefore able to provide clear details of who our ancient ancestors along the direct male and female lines were.

Furthermore, Bodmer’s (or was it Donnely’s) reference to “some pre-Celtic elements” is meaningless, without a definition of what is meant by ‘Celtic’. The only meaningful definition of ‘Celt’ is someone who speaks (or whose ancestors spoke) a Celtic language. The post-Ice-Age settlers did not speak a Celtic language, but their direct descendents did speak a Celtic language thousands of years later, so genetically they were the same people. Therefore to refer to pre-Celtic and Celtic separately in a genetic sense is meaningless.

As for genetic boundaries, Sykes shows a clear boundary at the Devon/Dorset border and the River Parrett, i.e. what was the eastern boundary of Dumnonia. West of this boundary the post-Ice-Age genetic element dominates at 78%, whereas east of this boundary this figure drops dramatically to 58%. In the rest of England only the Nothumbrian/Cumbrian figure comes anywhere near the ‘Dumnonia’ figure, at 68%.

My conclusion is that the Bodmer/Donnely analysis is basically a mess, being based on unsubstantiated assumptions (shown by Oppenheimer and Sykes to be incorrect), and by the non-use of Y-chromosome and mtDNA data (the only types of DNA that can accurately reflect a population’s deep ancestry).

Plymouth Exile,

Interesting.

lt is useful to note that the intent is to look to Y-chromosone and mtDNA data in future analysis. Perhaps this can be reconciled in time.

One common factor is the Devon (Dumnonian) border. It is significant under either analysis.

I found the following relevant links which maybe of interest

http://www.peopleofthebritishisles.org/pobipaper.pdf
http://www.peopleofthebritishisles.org/nl5.pdf

The former site includes numericalogical data that may suite PE's expertise more than my one (I think this loses me fairly easily)

However, the latter inludes the following interesting statement

"We are working on final analyses and writing up of a really exciting paper. It is based on 2,031 of our samples on which we have examined about 500,000 genetic markers across the genome. We have used a new approach to group individual samples together by genetic similarity into a number of clusters. These are then colour coded and placed on a map of the UK at the average position of where the grandparents were born. This new method of analysis, combined with our careful collection strategy, meansthat we can detect amazing fine scale population structure within the UK(see map). Indeed the detail of the geographic differences astonished us.

For example, the genetic boundaries between Cornwall, Devon and the
rest of England remarkably fall on the County boundaries, whilst in
Orkney there are obvious differences between Westray and Mainland.
This level of detail is unprecedented in human population genetics - untilnow it has been difficult to even differentiate reliably between North and South Europe. Furthermore, by collaborating with a couple of other large projects (within the Wellcome Trust Case Control Consortium 2), we have access to genetic data from about 6,300 more samples from throughout Europe. With these, we are currently looking to see if we can detect those European
regions that contributed to the clusters that we observe in the UK. The results are looking promising and will be an important part of our paper."

This clearly states that the researchers acknowledge that there are disctinct boundaries on both Devon's western AND eastern boundaries.

Something not stated elsewhereto my knowledge

Why was this second element not mentioned in the WMN??

I wonders why?

I also found the following referenceon the Wikipedia site for the Parrett river

"The origin of the name Parrett is unclear, but several derivations from the Celtic languages used in Wales have been suggested. Priestley-Evans suggests "Parrett has been said to be a form of the Welsh pared, a partition, and that it was the name which the Welsh people of Somerset and Devon gave to that river because it was at one time the dividing line between themselves and the Saxons". Another spelling parwydydd is also translated as partition. Another explanation from Welsh Peraidd meaning the sweet or delicious river, has also been suggested. An alternative explanation, based on Old English, is a derivation from Pedair or Pedride from pedr meaning four and the Old Cornish Rit meaning flow, which in this case would relate to the four flows or streams: the Tone, Yeo, Isle and Parrett. This is based on the explanation given in Ekwall's 1928 book English River Names. Whichever derivation is correct, the name Parrett and its spelling variations have been in use since the Anglo-Saxon era, as evidenced by the addition of -tun onto river names as seen in the local towns North Petherton and South Petherton"

The Parrett therefore maybe more significant as a boundary than I thought!!??

Ozzie Exile,

I note that the “People of the British Isles” (PoBI) paper does actually contain some Y-chromosome data, but only at the haplogroup level. However, some comparisons can be made with the Sykes data. Unfortunately the geographical regions used differ between the two studies. Also Sykes gives the haplogroups names, i.e. R1xR1a1 = Oisin, I = Wodan, and R1a1 = Sigurd.

For SW (Cornwall, Devon and Pembrokeshire), PoBI gives:-
R1xR1a1 at 75%, I at 15%, and R1a1 at 1.3%
For South West (Cornwall, Devon and West Somerset), Sykes gives:-
Oisin at 78.2%, Wodan at 12.6%, and Sigurd at 4.2%

Sykes’s sample size (at 239) was much larger than the PoBI sample size (at 80), but also bearing in mind the geographical differences, there is quite good correlation between the two sets of figures.

The PoBI study also used HLA (Human Leukocyte Antigen) as a measure, concerning which Jobling, Hurles and Tyler-Smith, “Human Evolutionary Genetics” state:-
“Environmental selection can be expected to skew the frequencies. Consequently, anthropological studies of HLA variation have become less popular in recent years.”

Other autosomal DNA types included in the PoBI study were MC1R for the R151C and R160W variants (associated with ginger hair). In this comparison, Devon (at 23%) correlates better with the ‘Celtic’ areas (Wales at 21%, Ireland at 31%) than Cornwall does at 16%.

The last comparison in the PoBI paper is of blood groups A, B, and O. Here (according to data presented by Oppenheimer), West Cornwall correlates well with Pembrokeshire, and Devon less well. However, as Oppenheimer points out, blood groups and their ABO SNPs (Single nucleotide polymorphisms) are poor indicators of human origins, e.g. Cornwall correlates far better with Norfolk and Lincolnshire than with most of Wales.

The intention of Bodmer’s group to use DNA data from Europe in order to identify European contributions to the British gene pool could introduce yet more errors of interpretation unless they are careful. Weale et al (“Y Chromosome Evidence for Anglo-Saxon Mass Migration”) compared Y-chromosome haplogroup frequencies from Frisia with those from Eastern England, and as they were very similar, concluded that there was a mass Dark Age migration from Frisia to Eastern England. They did not consider the possibility that both regions received similar migrations from Northern Iberia at the end of the Ice-Age, and later smaller scale migrations in the pre-Roman period. Oppenheimer, using the same data, but analysing at the haplotype level and using dating techniques concluded that there was no significant migration from Frisia to Eastern England.

The River Parrett information is interesting, as it has always been regarded as marking the eastern boundary of Dumnonia.

I note that the Cornwall24 brigade, in their desperate attempt to show how different they are from Devonians, have come up with their own reason why Devonians show as being a genetically different group from both the Cornish, and the ‘Wessex’ people on the PoBI map. They are suggesting that Wessex conquered Devon and deliberately filled it with pure Saxons in order to stress the differences at the Tamar border with Cornwall.

This is totally absurd, as the Sykes Y-chromosome survey clearly shows Cornwall, Devon, and West Somerset at an R1b haplogroup frequency of 78.2% (similar to North Wales, Borders, and Tayside and Fife), whereas the rest of Southern England (including what was core Wessex) is at an R1b frequency of 57.7%. If their hypothesis was correct, Devon would have shown an R1b frequency even lower than the 57.7% for the rest of Southern England.

I have seen a couple of postings to that effect.

Clearly the theory would not hold water. Actually, it is laughable.

The "Devon green" symbols spread somewhat across the border into Cornwall. Cornwall is not a pure brown. Do they suggest that parts of eastern Cornwall was also an area of "clearance".

I am more interested in the question whether Devon differs from Cornwall because it is an interim zone (with mixed influences) or whether the differences actually existed in pre-Saxon times. There has been the hypothesis that the Cornish (Cornovii) were somewhat distinct even though a part of Dumnonia. This could suggest an ancient split- but not unlike North and South Wales.

We will have to wait.

As you say, the clearance and repopulation hypothesis is laughable.

I have yet to find any real evidence that the Cornish were a separate tribe from the Devonians within Dumnonia during the Roman period. There were certainly well recorded tribes named ‘Cornovii’ in what is now the Shropshire/Cheshire region and also in what is now Caithness, but the name ‘Cornovii’ doesn’t appear in any Roman (or earlier) records pertaining to the far South West. This hasn’t stopped the Cornish nationalist brigade assuming that the term ‘Corn’ (meaning ‘horn’), found in Cornwall, must mean that there was a separate tribe with the name ‘Cornovii’.

It is possible that the differences between the Autosomal DNA found in Devonians and Cornish could be due to differences in the origins of some of the 19.8% of males who were not in the R1b Y-haplogroup. For instance, if there was a small genetic input (male and female) from Wessex into Devon, and a small Norwegian Viking genetic input into Cornwall, then the genetic lottery that is the recombining process in Autosomal DNA could easily lead to a distinct difference, without having any effect on the R1b Y-haplogroup figures in either area. Also, as Y-chromosome and mtDNA were not included in the PoBI analysis, any such Autosomal difference would tend to look much bigger than they really were.

Sorry, that should be 21.8% of males who were not in the R1b Y-haplogroup, not 19.8%.

I found the following website which shows Donnelly and Bodmer discussing the genetic map survey.

http://www.ox.ac.uk/media/news_stories/2012/120703.html

Donnelly states that the Cornish are genetically distinct from Devonians, and that Devonians are genetically distinct from areas to the east.

There is nothing here talking about what makes up these differences, or how this relates to Celtic genes - or otherwise. There is only conjecture that these different groupings may come from the way Britain was repopulated after the ice-age, or was influenced by the Anglo Saxon invasion. It seems to be only at the conjectural stage so far.

I have posted on the "thisisdevon" forum (see tab) about the Oxford Uni survey - if you want to have a look.

Sadly this forum is not all that popular based on posting frequency, but still

Ozzie Exile,

Having studied what autosomal DNA can reveal about ancestry, I would say that what this survey tells us is that there has been very little interbreeding between the Cornish and the Devonians during the last couple of centuries, and that similarly there has been very little interbreeding between Devonians and people from the South of England (to the east of Devon).

It tells us virtually nothing about the ancient ancestors of Cornish, Devonians, or the rest of the people in Southern England. That would require haplotype level analysis of Y-chromosome and mtDNA samples, such as that which has already been done by Oppenheimer and Sykes. It will be interesting to see the results of Bodmer's analysis of Y-chromosome and mtDNA from the samples that he has collected, but is apparently still to examine.

PE,

That is interesting, and fits the view that I was drifting to over the last couple of weeks.

What is interesting is that (for Devon & Cornwall) the peninsular was divided (roughly but clearly) along traditional County boundaries.

This survey looked only at those who have all four grandparents born within the local area, and so this might bias towards those who are the "stay at homes" in the community - but still, it sheds insights.

Why this "local" pattern does not extend to England's South, South- East, or Midlands, is not clear to me.

Perhaps this represents some sort of social boundaries but I am not sure. Why this should fit so tightly the the ancient Dumnonian boundary I cannot say. The Cornish/Devon border is not as distinct by the way (at least not based on what has been disclosed graphically) as it appears that eastern Cornwall and Western Devon are a "similar mix" of both Cornish and Devonian genotypes - despite what has been said elsewhere. The Devon/Dorset border appears to be much more distinct - and closer to traditional boundaries.

Certainly it does nothing to suport the olde Wessex Regionalism crap. No Wessex conformity evident!



I believe we still need to wait further analysis.

Ozzie Exile,

I get your point about the Devon/Cornwall border not appearing to be as distinct as the Devon/Dorset border in terms of the autosomal DNA genotypes. I suppose it is only natural to expect to see some evidence of interbreeding between (for instance) people in Plymouth and people in Saltash, or between people in Tavistock and people in Callington, but very little between (for instance) people in Honiton and people in Truro.

The apparent lack of evidence for interbreeding between people of East Devon and West Dorset could well be due to some sort of social boundary, which doesn’t seem to occur in the rest of Southern England. Unfortunately there doesn’t appear to be any autosomal DNA data for East Somerset, to see if it conforms more with the Devon genotype (which is also apparent in West Somerset), or with the Southern England genotype. If the boundary in this area is more a social one, then one might expect to see a fairly distinct genotype boundary coinciding with the River Parrett, as this marked the eastern boundary of Dumnonia. Certainly Sykes shows such a boundary at the Parrett for the Y-chromosome at both the haplogroup and the haplotype levels. It will be interesting to see what Bodmer’s Y-chromosome and mtDNA analyses show, and whether it correlates with the Sykes analysis.

As for the Wessex Regionalists’ fantasies about Devonians being West Saxons, that notion was blown out of the water by the Sykes analysis, and this is more or less confirmed by the Bodmer autosomal DNA genotype cluster map.