PDA

View Full Version : DNA study of Britons has shown that genetically there is not a unique Celtic group



rock hunter
06-04-2016, 07:56 AM
According to the data, those of Celtic ancestry in Scotland and Cornwall are more similar to the English than they are to other Celtic groups.
The study also describes distinct genetic differences across the UK, which reflect regional identities.
And it shows that the invading Anglo Saxons did not wipe out the Britons of 1,500 years ago, but mixed with them.
Published in the Journal Nature, the findings emerge from a detailed DNA analysis of 2,000 mostly middle-aged Caucasian people living across the UK.
The individuals included had all four of their grandparents living close to each other in a rural area.
This selection criterion enabled the researchers, led from Oxford University, to filter out 20th-Century immigration and to peer back to migration patterns more than 1,000 years ago.
Striking similarities
According to Prof Peter Donnelly who co-led the study, the results show that although there is not a single Celtic group, there is a genetic basis for regional identities in the UK.
"Many of the genetic clusters we see in the west and north are similar to the tribal groupings and kingdoms around, and just after, the time of the Saxon invasion, suggesting these kingdoms maintained a regional identity for many years," he told BBC News.
Prof Donnelly and his colleagues compared genetic patterns now with the map of Britain in about AD 600, after the Anglo Saxons had arrived from what is now southern Denmark and Northern Germany. By then, they occupied much of central and southern England.
"We see striking similarities between the genetic patterns we see now and some of these regional identities and kingdoms we see in AD 600, and we think some of that may well be remnants of the groupings that existed then," he explained.
A map of different genetic groupings reveals subtle but distinct differences between those sampled in West Yorkshire and the rest of the country.
9611
Genetic Map of the UK
Image captionWho do you think you are? This genetic map might tell you. Each colour represents a different genetic group. Many correspond very closely to county borders, indicating a genetic basis for regional identities.

There is also a marked division between the people of Cornwall and Devon that almost exactly matches the county border. And the People of Devon are distinct again to those from neighbouring Dorset.
The Wellcome Trust-funded study, which is part of the People of the British Isles Research Project, also found that people in the north of England are genetically more similar to people in Scotland than they are to those in the south of England.
It also finds that people in North and South Wales are more different from each other than the English are from the Scots; and that there are two genetic groupings in Northern Ireland.
Prof Mark Robinson, an archaeologist who works with Prof Donnelly at Oxford University, said he was "very surprised" that Celtic groups in Cornwall, Wales, Northern Ireland and Scotland had such different genetic patterns.
"I had assumed at the very early stages of the project that there was going to be this uniform Celtic fringe extending from Cornwall through to Wales into Scotland. And this has very definitely not been the case," he told BBC News.
The researchers did see distinct genetic groups within those regions but those groups were quite different from each other, according to Prof Donnelly.
"Although people from Cornwall have a Celtic heritage, genetically they are much, much more similar to the people elsewhere in England than they are to the Welsh for example," said Prof Donnelly.
"People in South Wales are also quite different genetically to people in north Wales, who are both different in turn to the Scots. We did not find a single genetic group corresponding to the Celtic traditions in the western fringes of Britain."
Into the Dark Ages
The finding is the first genetic evidence to confirm what some archaeologists have long been arguing: that Celts represent a tradition or culture rather than a genetic or racial grouping.

Prof Robinson noted that the results also shed light on what happened during Britain's Dark Ages, in the years between AD 400 and AD 600, after the Romans left.
Towns were abandoned; the language over much of what became England changed (to Anglo Saxon, which became English); pottery styles altered; so too even the cereals that were grown, following the arrival of people from the base of the southwest Danish peninsula and northwestern Germany (the Anglo Saxons).
Some historians and archaeologists had wondered whether these changes occurred as a result of the Saxons entirely replacing the existing population as they moved westwards. That might have happened if the Saxons introduced disease, for example.
Others researchers suggested that the existing population simply dropped their old ways and adopted the Saxon way of life.
The new analysis shows a modest level of Saxon DNA, suggesting that the native British populations lived alongside each other and intermingled with the Anglo Saxons to become the English.
There is some evidence in the study that intermingling did not happen immediately following the Saxons' arrival, but occurred at least 100 years later. This suggests that Britons and Saxons had separate communities to begin with, and then over time they began to merge.
Northern Irish groupings
This may well be one of the first instances where genetics has been used to clear up historical controversy.
The study seems to confirm the view that Celts retained their identity in western and northern areas of England where the regions were incorporated into Anglo Saxon territory by conquest.
But what could account for the variation in the DNA of those of Celtic ancestry in Cornwall, Wales and Scotland? Time would be one possibility, according to Prof Donnelly.
"If groups have been separated for a period of time, they will diverge genetically so some of the differences we see genetically are the result of those kinds of effects," he said.
The study also notes that there are two genetic groupings in Northern Ireland: one of which also contains individuals across the sea in western Scotland and the Highlands; the other contains individuals in southern Scotland and southern England.
The former appears to reflect the kingdom of Dalriada 1,500 years ago; the other probably represents the settlers of the Ulster Plantations.

And in Orkney, the study finds clear evidence of Norwegian DNA, as might be expected from the Viking settlement of the Islands.
Interestingly, it persists at fairly low levels, suggesting that the Vikings and the existing populations coexisted and intermingled more than people had expected - in the way that occurred with the Anglo Saxons.
The Viking armies that laid waste to parts of England, and for a while ruled what became known as the Danelaw, left little if any genetic trace, confirming that their success was due to their military prowess rather than large-scale population movement.
Likewise, the Norman conquest of England did not leave any genetic evidence.
Source: http://www.bbc.com

A Norfolk L-M20
06-04-2016, 09:24 AM
POBI. Last year's news, but you've put a different angle on it. Before any debate starts - my bias is that I'm a Celtic reductionist, although accepting a general Atlantic seaboard group of cultures, that might tie up with increased Steppe ancestry.

The POBI report might support my bias, in the sense that it's arguing against any homogeneous Celtic population in the British Isles / N. Ireland (Republic of Ireland was not included in the study).

Dubhthach
06-04-2016, 09:53 AM
Doesn't help that the media ignore the fact that all the PoBI clusters share components in comparison but with different proportion,

In case of below pic (taken from before they removed their Irish eg. 26 county sample)

http://compsoc.nuigalway.ie/~dubhthach/DNA/POBI/POBI-16-17.png

Cluster 17 is termed "French" due to high levels in Northwest France, of course what might have been more obvious is that 17 represents a Brythonic component.

rms2
06-04-2016, 01:25 PM
I'm not sure why anyone would expect there to be a unique Celtic group based on autosomal dna, which is recombinant and reflects generational churning and change. Naturally, whoever the male ancestors of the Celts were, they undoubtedly married and bred with local women. And local women varied with the locality.

Since the Celts were an Indo-European, patriarchal people, why not take a look at y-dna instead, especially since it is passed pretty much intact from father to son? There I think you find a distinct Celtic pattern, with the areas of the British Isles and Ireland where the Celts held out the longest and preserved their languages the longest predominantly R1b-P312 and especially R1b-L21.

How could anyone miss that? Really.

9619

avalon
06-05-2016, 08:34 AM
Doesn't help that the media ignore the fact that all the PoBI clusters share components in comparison but with different proportion,


Cluster 17 is termed "French" due to high levels in Northwest France, of course what might have been more obvious is that 17 represents a Brythonic component.

Although in the final POBI paper there was a northern France component (FRA17) which was present in all the UK clusters, even as far north as Orkney, but which was totally absent from the Welsh clusters.

In terms of the whole celtic uniformity thing, the fine scale genetics isn't really surprising. All you have to to do is look at the basic history and geographical separation of places like Cornwall, Wales and Scotland to know they were going to differ genetically somewhat. Even within Wales, the genetic separation between North and South Wales could have been predicted from a cursory reading of the history, ie. the emergence of early Welsh kingdoms such as Gwynedd in the north, Dyfed in the SW, etc

9636

Dubhthach
06-05-2016, 10:36 AM
Although in the final POBI paper there was a northern France component (FRA17) which was present in all the UK clusters, even as far north as Orkney, but which was totally absent from the Welsh clusters.

In terms of the whole celtic uniformity thing, the fine scale genetics isn't really surprising. All you have to to do is look at the basic history and geographical separation of places like Cornwall, Wales and Scotland to know they were going to differ genetically somewhat. Even within Wales, the genetic separation between North and South Wales could have been predicted from a cursory reading of the history, ie. the emergence of early Welsh kingdoms such as Gwynedd in the north, Dyfed in the SW, etc

9636

That's cluster 16 in my earlier photo, as you note in original analysis completely absent in Wales as well.

Going by your image, it looks like they removed the samples from Italy, Poland and Finland and than regenerated new admixture components.
http://compsoc.nuigalway.ie/~dubhthach/DNA/POBI/POBI-original-clusters.png

What's of note in the original, is that when you look at the three major French groups. What's evident is that "Component 16" is dominant in all three, the North/Central group been prime example, the other two clusters appear to have high admixture with another cluster. In case of "NW France" this is with "Component 17", which if you ask me is a Brythonic component as it's at it's highest level among the Welsh. The "SW France" cluster obviously is admixed with "Component 15" (eg. modal in Spain). Makes sense from a geographical point of view.

What seems to be obvious is that higher levels of "17" and "24" (Ireland 26 counties) in British sub-clusters actually reflects linguistic history, eg areas that are still or were still "Celtic speaking" in the last 1,000 years have higher levels of both these components.

http://compsoc.nuigalway.ie/~dubhthach/DNA/Breakdown_of_UK_DNA.gif

Leaving that aside, I agree that geography is a key thing here, especially when you consider that areas where Brythonic hung on longer were all seperated from each other, one reasons why we see "intermediate clusters", the "Devon cluster" is a good example of that compare to Cornish to west and "Red cluster" to east. We see other "intermediated clusters" in welsh borders for example.

Anyways if you look at the Irish DNA atlas slide that Gerard posted before, it's fairly obvious that geography plays a big part in way people cluster.

IF you invert the image you get this:
http://compsoc.nuigalway.ie/~dubhthach/DNA/Atlas02.jpg
9638

Almost reflects geography.

avalon
06-05-2016, 08:53 PM
That's cluster 16 in my earlier photo, as you note in original analysis completely absent in Wales as well.

Going by your image, it looks like they removed the samples from Italy, Poland and Finland and than regenerated new admixture components.

What's of note in the original, is that when you look at the three major French groups. What's evident is that "Component 16" is dominant in all three, the North/Central group been prime example, the other two clusters appear to have high admixture with another cluster. In case of "NW France" this is with "Component 17", which if you ask me is a Brythonic component as it's at it's highest level among the Welsh. The "SW France" cluster obviously is admixed with "Component 15" (eg. modal in Spain). Makes sense from a geographical point of view.

What seems to be obvious is that higher levels of "17" and "24" (Ireland 26 counties) in British sub-clusters actually reflects linguistic history, eg areas that are still or were still "Celtic speaking" in the last 1,000 years have higher levels of both these components.

Leaving that aside, I agree that geography is a key thing here, especially when you consider that areas where Brythonic hung on longer were all seperated from each other, one reasons why we see "intermediate clusters", the "Devon cluster" is a good example of that compare to Cornish to west and "Red cluster" to east. We see other "intermediated clusters" in welsh borders for example.

Anyways if you look at the Irish DNA atlas slide that Gerard posted before, it's fairly obvious that geography plays a big part in way people cluster.


I agree, what was cluster 17 in the original analysis and in the final paper is Fra14 does look like a Brythonic component. From a Welsh perspective, the big mystery for me is the absence of the north/central France component in Wales. This component is also low in Southern Scotland/N Ireland so that may have some meaning.

fridurich
06-05-2016, 10:20 PM
Doesn't help that the media ignore the fact that all the PoBI clusters share components in comparison but with different proportion,

In case of below pic (taken from before they removed their Irish eg. 26 county sample)

http://compsoc.nuigalway.ie/~dubhthach/DNA/POBI/POBI-16-17.png

Cluster 17 is termed "French" due to high levels in Northwest France, of course what might have been more obvious is that 17 represents a Brythonic component.

Very good point Dubhthach! It would also have been nice if the POBI project had left their chart showing the proportions of the Irish component in the rest of the British Isles regions. As you know, it was a chart similar to the one showing a French component.

The Irish component was shown in purple and all or almost all areas of the British Isles they tested showed this component. Some areas such as the one in Galloway showed a fairly high Irish component (about 45 percent in Galloway).

Way too high I think, to explain it only coming from Irish immigration to Galloway in the 19th century and after.

fridurich
06-05-2016, 10:38 PM
I'm not sure why anyone would expect there to be a unique Celtic group based on autosomal dna, which is recombinant and reflects generational churning and change. Naturally, whoever the male ancestors of the Celts were, they undoubtedly married and bred with local women. And local women varied with the locality.

Since the Celts were an Indo-European, patriarchal people, why not take a look at y-dna instead, especially since it is passed pretty much intact from father to son? There I think you find a distinct Celtic pattern, with the areas of the British Isles and Ireland where the Celts held out the longest and preserved their languages the longest predominantly R1b-P312 and especially R1b-L21.

How could anyone miss that? Really.

9619

Excellent point!

It is kind of like if I, and a large group of my male relatives who have the same YDNA haplogroup and terminal snp, decided to migrate to 3 parts of the world: Japan, Uganda, and Norway, without taking a lot of women with us who were from the same areas as us. So, the majority of us wind up marrying local women in Japan, Uganda, and Norway.

So after 500 (maybe only 200 years are needed) or 1000 years, scientists, who have no knowledge of where our direct paternal line ancestors came from, or what their YDNA haplogroup is, do autosomal dna testing on our descendants in the 3 groups.

After exhaustive tests and also doing FineStructure analysis, they show a chart where the original 3 immigrant groups are far away from each other, showing great distance. So, the scientists, state "There is absolutely no genetic relationship between these 3 groups as can be attested by the distances from each other shown on the chart, and the other evidence."

lol

JMcB
06-06-2016, 02:59 AM
I'm not sure why anyone would expect there to be a unique Celtic group based on autosomal dna, which is recombinant and reflects generational churning and change. Naturally, whoever the male ancestors of the Celts were, they undoubtedly married and bred with local women. And local women varied with the locality.

Since the Celts were an Indo-European, patriarchal people, why not take a look at y-dna instead, especially since it is passed pretty much intact from father to son? There I think you find a distinct Celtic pattern, with the areas of the British Isles and Ireland where the Celts held out the longest and preserved their languages the longest predominantly R1b-P312 and especially R1b-L21.

How could anyone miss that? Really.

9619

Wouldn't that also apply to the Danish Vikings whose autosomal DNA would have soon been subsumed by intermarriage with the indigenous population or perhaps reinforced the preexisting Anglo Saxon signature by intermarrying with them?

From what I remember reading, it's virtually impossible to genetically distinguish between the Anglo Saxons and the Danes because they both came from the same general area within a short time span.

Which makes me wonder about their statement that:

"The [Danish] Viking armies that laid waste to parts of England, and for a while ruled what became known as the Danelaw, left little if any genetic trace, confirming that their success was due to their military prowess rather than large-scale population movement."

I sometimes wonder if the genetic contributions of the Danes, isn't mistakenly being attributed to the Anglo Saxons. Be that as it may, I would imagine that only SNPs will tell.

rms2
06-06-2016, 11:13 AM
Wouldn't that also apply to the Danish Vikings whose autosomal DNA would have soon been subsumed by intermarriage with the indigenous population or perhaps reinforced the preexisting Anglo Saxon signature by intermarrying with them?

From what I remember reading, it's virtually impossible to genetically distinguish between the Anglo Saxons and the Danes because they both came from the same general area within a short time span.

Which makes me wonder about their statement that:

"The [Danish] Viking armies that laid waste to parts of England, and for a while ruled what became known as the Danelaw, left little if any genetic trace, confirming that their success was due to their military prowess rather than large-scale population movement."

I sometimes wonder if the genetic contributions of the Danes, isn't mistakenly being attributed to the Anglo Saxons. Be that as it may, I would imagine that only SNPs will tell.

True. Of course, the problem is that the Danish Vikings and the Anglo-Saxons were pretty much the same people separated by an interval of 300 years or so.

avalon
06-06-2016, 04:33 PM
It is kind of like if I, and a large group of my male relatives who have the same YDNA haplogroup and terminal snp, decided to migrate to 3 parts of the world: Japan, Uganda, and Norway, without taking a lot of women with us who were from the same areas as us. So, the majority of us wind up marrying local women in Japan, Uganda, and Norway.

So after 500 (maybe only 200 years are needed) or 1000 years, scientists, who have no knowledge of where our direct paternal line ancestors came from, or what their YDNA haplogroup is, do autosomal dna testing on our descendants in the 3 groups.

After exhaustive tests and also doing FineStructure analysis, they show a chart where the original 3 immigrant groups are far away from each other, showing great distance. So, the scientists, state "There is absolutely no genetic relationship between these 3 groups as can be attested by the distances from each other shown on the chart, and the other evidence."


The only thing I would say is that although R1b-P312 is clearly a strong Celtic y-dna signal, modern Celtic fringe populations differ somewhat at the autosomal level (which is the whole of an individual's ancestry).

Take for example Cornwall. A Celtic language was still widely spoken there only a few centuries ago, R1b-P312 is likely at high frequencies there, but it's important to note that bound by geography, the modern Cornish are most closely related to people from Devon and then from southern England, not to people from Wales or Scotland.

Dubhthach
06-06-2016, 05:52 PM
The only thing I would say is that although R1b-P312 is clearly a strong Celtic y-dna signal, modern Celtic fringe populations differ somewhat at the autosomal level (which is the whole of an individual's ancestry).

Take for example Cornwall. A Celtic language was still widely spoken there only a few centuries ago, R1b-P312 is likely at high frequencies there, but it's important to note that bound by geography, the modern Cornish are most closely related to people from Devon and then from southern England, not to people from Wales or Scotland.

Sure but Cornwall has been under some sorta English overlordship right back to Kingdom of Wessex, as a result you've got obvious admixture from the components found in higher levels in the "Red Cluster" than the Welsh.

Same goes with the "Forest of Dean" cluster which is basically the English/Welsh borderlands. The cornish pop out at K=6 (in published paper) as distinct cluster, if you look at PCA space, they basically occupy similar space as Northern Britain cluster at that K level.

Basically due to admixture probably with extension of English language (and gradual population movement over last 1,200 years) they've probably shifted more towards the general English cluster. One could argue that looking at PCA that technically speaking the Cornish cluster falls in between most extreme samples of Red Cluster and North Wales.

What would be interesting is if ancient-DNA samples were included in the analysis, after all I'm assuming the specific SNP list that they used on their chip has been published? You could cull those SNP's from ancient samples (Hinxton, Yorkshire Iron age, Yorkshire Romano-British, Yorkshire AS) and do a reanalysis. It wouldn't surprise me that if you took aDNA samples from Cornwall (say on order of 1,000 years old) that they would cluster differently to modern Cornish (given history of movement into Cornwall in last 3-400 years), the modern Cornish would probably cluster intermediate between any ancient samples and the Red Cluster.

Dimanto
06-06-2016, 07:29 PM
A bunch of pseudohistory is what they want to sell us, but the study of genetics saves the day by refuting all this nonsense one by one.

avalon
06-06-2016, 08:30 PM
Sure but Cornwall has been under some sorta English overlordship right back to Kingdom of Wessex, as a result you've got obvious admixture from the components found in higher levels in the "Red Cluster" than the Welsh.

Same goes with the "Forest of Dean" cluster which is basically the English/Welsh borderlands. The cornish pop out at K=6 (in published paper) as distinct cluster, if you look at PCA space, they basically occupy similar space as Northern Britain cluster at that K level.

Basically due to admixture probably with extension of English language (and gradual population movement over last 1,200 years) they've probably shifted more towards the general English cluster. One could argue that looking at PCA that technically speaking the Cornish cluster falls in between most extreme samples of Red Cluster and North Wales.

What would be interesting is if ancient-DNA samples were included in the analysis, after all I'm assuming the specific SNP list that they used on their chip has been published? You could cull those SNP's from ancient samples (Hinxton, Yorkshire Iron age, Yorkshire Romano-British, Yorkshire AS) and do a reanalysis. It wouldn't surprise me that if you took aDNA samples from Cornwall (say on order of 1,000 years old) that they would cluster differently to modern Cornish (given history of movement into Cornwall in last 3-400 years), the modern Cornish would probably cluster intermediate between any ancient samples and the Red Cluster.

For sure. I have said before on this forum about English admixture in Cornwall that probably goes back centuries. A lot of it may actually have occurred fairly recently, linked perhaps to the decline of the Cornish language, large scale emigration of Cornish miners in the 18th/19th centuries and in-migration of English speakers. All of this happened in the last 400 hundred years so it probably wasn't that long ago that the Cornish were more Welsh-like in terms of autosomalDNA.

I imagine we'd have to go back even further in time though to see close genetic affinities between Cornish and Highland Scots for instance.

wrt to ancient dna there has been a paper recently with some interesting stuff on steppe ancestry, in which I believe they projected PoBI data on to ancient samples.