Fine Scale Population Structure in the British Population

[Angela]

This is the long awaited study by the POBI (People of the British Isles Project). Leslie et al 2015, Fine Scale Population Structure in the British Population:
http://www.nature.com/nature/journal...ture14230.html

Dienekes opines here at the link below. I've skimmed the paper, and if you don't have time to read the whole paper, Dienekes' review covers all the major points.
http://www.dienekes.blogspot.com/201...t-al-2015.html

My major take away is how little structure there actually is in the British population: "the average of the pairwise FST estimates between each of the 30 sample collection districts is 0.0007, with a maximum of 0.003 (Supplementary Table 1)"

Here are some other highlights as pointed out by Dienekes:

-One cluster, covering most of central and southern England accounts for almost half the people in the study. That cluster seems to be heavily influenced by the Saxon migration.

-However, " Two separate analyses (ancestry profiles and GLOBETROTTER) show clear evidence in modern England of the Saxon migration, but each limits the proportion of Saxon ancestry, clearly excluding the possibility of long-term Saxon replacement. We estimate the proportion of Saxon ancestry in Cent./S England as very likely to be under 50%, and most likely in the range of 10–40%."

-"In particular, we see no clear genetic evidence of the Danish Viking occupation and control of a large part of England, either in separate UK clusters in that region, or in estimated ancestry profiles, suggesting a relatively limited input of DNA from the Danish Vikings and subsequent mixing with nearby regions, and clear evidence for only a minority Norse contribution (about 25%) to the current Orkney population."

-"We saw no evidence of a general ‘Celtic’ population in non-Saxon parts of the UK. Instead there were many distinct genetic clusters in these regions, some amongst the most different in our study, in the sense of being most separated in the hierarchical clustering tree in Fig. 1."


From a personal point of view I find it interesting that N.Wales shows a connection not only to France, but to the border area between France and Italy in Piemonte, and in Liguria and down toward Lucca as well. This isn't the first time such a link has appeared; it showed up in the recent autosomal study of European populations by Hellenthal et al 2014:
http://www.sciencemag.org/content/343/6172/747

The Dienekes analysis of that paper:
http://dienekes.blogspot.com/2014/02...n-history.html

Interesting to consider whether it might have something to do with the "Isles type" R1b that shows up in Liguria in Boattini et al.

From 23andme, I know that people of British Isles descent have many devoted geneaologists among them. I'm sorry for their sake that the data won't be available to the public.
-

[Greying Wanderer]

The data thing (edit: keeping it private i mean) has spoiled this to an extent.

I've always thought that many of the various waves into Britain followed a repeating pattern of people being pushed to the west so i was looking forward to North Wales dna possibly being a repository of these waves showing them like the sediment layers in rocks.

Very annoying. Oh well.

The distinct Celtic clusters are no surprise as it's visibly noticeable if you've ever spent a lot of time in the remoter mountainous parts of north and west Britain and again (imo) speaks to the different invasions over time and their different source directions.

At least the Iberian/Ligurian connection to North Wales shows up - one of the earliest layers imo originally connected to Atlantic Megalith maybe.

[Aberdeen]

A very interesting paper, although I won't get a chance to read it in detail for a day or two. The stuff about Wales is quite interesting.

[sparkey]

My first impression based on the summary is that this is about as expected based on the bits and pieces they've been throwing out there for the past few years, and I'm glad that they've come to very sensible conclusions based on the data. I was wondering what they would make of their rather clear results that the pre-Anglo-Saxon population was not homogenous. Placing the Anglo-Saxons as more genetically homogenous than the Romano-Britons in light of the data makes sense to me. An alternative explanation for certain affinities, like Cornwall's greater affinity to the big red Southeastern English cluster than northern England, could be explained by saying that Cornwall is more Anglo-Saxon than northern England. It seems that they're instead suggesting that a better explanation is a differences in the substrata.

[Maleth]

Its on bbc news too....be patient, of course you NEED to watch the amazing ad before, THANK HEAVENS WE HAVE THE RIGHT TO CHOOSE, HALLELUJAH.

http://www.bbc.com/news/science-environment-31905764

[Greying Wanderer]

Annoyance over data aside looking at the list of components h/t http://dienekes.blogspot.co.uk/2015/...t-al-2015.html

http://1.bp.blogspot.com/---rAR5GpVQ...talSources.jpg

and assuming for the sake of argument that the refuge zone follows the physical geography so North Wales (and probably parts of western Scotland) is the last refuge and so assume that components which are at a higher frequency in North Wales than England derive from earlier population layers being pushed to the west (it could also be the route taken but for now i'll assume they represent earlier waves just to see what comes out) we get

more in north wales than england
--------------------------------------
FRA12 (only scotland and wales)
SFS21 (med. coastal connection?)
FRA14 (all over, more in wales and scotland)
GER6 (all over, more in wales and scotland)

(nb gap between wales/scotland frequency and england frequency seems slightly bigger for FRA14 over GER6 which according to the refuge premise might imply FRA14 was earlier)

more in england than north wales
--------------------------------------
BEL11 (everywhere but less in wales)
DEN18 (less in wales and west scotland)

little or none in wales
-------------------------
GER3 (mostly missing in wales and scotland)
FRA17 (missing in wales but not scotland)

sweden/norway
------------------
hard to say just by looking but to me the sum total of the various components looks about the same - will ignore that for now

then thinking of the different migrations/invasions

1. Doggerland
2. Atlantic Megalith
3. BB / pre-celtic / early celtic (?)
4. La Tene / Belgae (?)
5. Romans
6. Saxon/Viking/Norman
7. later?

1. Doggerland: would either be a very widespread base component or be the most refuge-like. i'm gonna guess the latter so maybe FRA12
2. Atlantic Megalith: second most refuge-like and with a connection to the Mediterranean - SFS21

5/6/7. jumping ahead to those with little or none in north wales so assumed to be later we have
DEN18 (everywhere but less in wales and west scotland)
GER3 (mostly missing in scotland and wales)
FRA17 (mostly missing in wales but not scotland)
maybe saxon/viking/norman? or FRA17 is later but seems too massive in England for Huguenots, earlier? Romans?

(maybe Germany used to be more French-like and everything shifted anti-clockwise i.e. the "Germanic" expansion involved French-like Germans being pushed into France by more German-like Germans?)

anyway too confusing for now

.

the main aim of this was to try and see if you could pin down components from earlier and later to see how many possible population slices there were left over for the confusing "Celtic" era in the middle:
FRA14
GER6
BEL11

.

Anyway, I doubt any of this is actually correct (except maybe the Atlantic Megalith slice) but the intention is to show that maybe the data can be squeezed to show a sequence, especially if any of the data can be used to indicate time depth, or at least a partial sequence if at least the earliest and latest can be identified.

[Angela]

I think that all this talk of the differences between the "clusters" obscures the fact that all the people of the British Isles are incredibly homogenous. Those are amazingly small Fst differences.

That said, it's still interesting.

Another publication has covered it, understandably. There are some quotes from the authors.
http://www.dailymail.co.uk/sciencete...#ixzz3UmozzND9

In explaining the lack of traceable Dane Viking, "Roman", and Norman ancestry, they have this to say:
"Sir Walter Bodmer, from Oxford University, said the lack of Viking DNA may have largely been to do with numbers.

He said: 'It's important to emphasise that when you get that mixture it's very much a question of the ratio of the people who come in and the indigenous population."


I think it's important to keep this in mind whenever people are talking about migrations of conquering male elites; they can change the culture and the language, but their genetic heritage can disappear from the autosomes, if not the yDna, no matter how many wives they took. You need large numbers of men and women, a "folk migration" to make lasting change. The only exception might be if they brought diseases with them, diseases which wiped out large numbers of the indigenous inhabitants. That's why I'm always very skeptical in studies of my own country about claims that the Normans left an autosomal impact on western Sicily, or some quartering of troops around a "Lombard" castle in the south changed the genetic make-up to some large degree.


I found this intriguing, by the way:
"But even within Wales there are two distinct tribes, with those in the north and south of the principality less similar genetically than the Scots are to the inhabitants of Kent."

What could explain this? Religion is the same, yes? Topography? Language? The same applies to the differences between Cornwall and Devon, although there you have a river as a dividing line. Did both areas retain the use of "Celtic" languages for a longer period?

[bicicleur]

IMO the Britons were Celtic tribes R1b-L21 which are still the majority (60%) in Wales, Cornwall, Scotland, Ireland and Bretagne.
So once these tribes had settled, and absorbed the local indogenous people they must have remained sedentary and rather isolated without many intermarriages?
Or is there another explanation?

[Sile]

as I have been saying for a long time,,,,,there is no celtic ethnicity

from link P#5

A DNA study of Britons has shown that genetically there is not a unique Celtic group of people in the UK.

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.

[sparkey]

What could explain this? Religion is the same, yes? Topography? Language? The same applies to the differences between Cornwall and Devon, although there you have a river as a dividing line. Did both areas retain the use of "Celtic" languages for a longer period?

In Wales, there has really always been a North/South divide in history, so perhaps it is also an ancient divide. All of the old kingdoms were divided North vs. South. Between the Roman period and annexation to England, unification of Wales happened for a grand total of 7 years under Gruffydd ap Llywelyn, the rest of the time there was a dominant northern kingdom (usually Gwynedd) and a dominant southern kingdom or two (usually Deheubarth and sometimes Morgannwg or Gwent). A similar division seems to have existed prior to Roman arrival, with the Ordovices and Deceangli in the North, and the Demetae and Silures in the South.

One surprise to me is how the Welsh Marches cluster is nowhere near the South Wales cluster, I thought that they would be similar. Instead, the Welsh Marches cluster shows some Anglo-Saxon affinity (although less than most of England) and a shared substratum with Cornwall+Devon+Southeast England.

As for Cornwall vs. Devon: In additional to a likely ancient divide, there is a large gap in time when they were absorbed into England and when they started speaking English. Devon was completely absorbed into Wessex in 710, when King Ine of Wessex defeated King Geraint of Dyvnon (that is, Dumnonia, or Cornwall and Devon combined). Devon began speaking English very soon after. The Cornish, meanwhile, maintained their independence after their victory at Hehil in 722, and we don't find evidence of the Cornish submitting to the English until over a century later, and we don't find it being considered a part of England until about a century after that. That allows the Cornish to maintain some cultural distinctiveness through the years of Anglo-Saxon cultural conquest, and they continued to speak Cornish until about 1700 (depending on which part of Cornwall we're talking about). So western Cornwall spoke a Celtic language for literally a millennium longer than Devon.

[Sile]

In Wales, there has really always been a North/South divide in history, so perhaps it is also an ancient divide. All of the old kingdoms were divided North vs. South. Between the Roman period and annexation to England, unification of Wales happened for a grand total of 7 years under Gruffydd ap Llywelyn, the rest of the time there was a dominant northern kingdom (usually Gwynedd) and a dominant southern kingdom or two (usually Deheubarth and sometimes Morgannwg or Gwent). A similar division seems to have existed prior to Roman arrival, with the Ordovices and Deceangli in the North, and the Demetae and Silures in the South.

One surprise to me is how the Welsh Marches cluster is nowhere near the South Wales cluster, I thought that they would be similar. Instead, the Welsh Marches cluster shows some Anglo-Saxon affinity (although less than most of England) and a shared substratum with Cornwall+Devon+Southeast England.

As for Cornwall vs. Devon: In additional to a likely ancient divide, there is a large gap in time when they were absorbed into England and when they started speaking English. Devon was completely absorbed into Wessex in 710, when King Ine of Wessex defeated King Geraint of Dyvnon (that is, Dumnonia, or Cornwall and Devon combined). Devon began speaking English very soon after. The Cornish, meanwhile, maintained their independence after their victory at Hehil in 722, and we don't find evidence of the Cornish submitting to the English until over a century later, and we don't find it being considered a part of England until about a century after that. That allows the Cornish to maintain some cultural distinctiveness through the years of Anglo-Saxon cultural conquest, and they continued to speak Cornish until about 1700 (depending on which part of Cornwall we're talking about). So western Cornwall spoke a Celtic language for literally a millennium longer than Devon.

the

http://en.wikipedia.org/wiki/Prayer_Book_Rebellion

was basically the start of the end of these "non-angleized" devon and cornish people

[Greying Wanderer]

...

I found this intriguing, by the way:
"But even within Wales there are two distinct tribes, with those in the north and south of the principality less similar genetically than the Scots are to the inhabitants of Kent."

What could explain this? Religion is the same, yes? Topography? Language? The same applies to the differences between Cornwall and Devon, although there you have a river as a dividing line. Did both areas retain the use of "Celtic" languages for a longer period?

What I wonder is if you have a backdrop of a mostly very homogenous population which contains a small but highly divergent component within it e.g. a component with an African element, whether the genetic sorting software might cause that component to show disproportionately. If so then even very small traces of that component might be trackable.

I think Atlantic Megalith might be trackable this way because of their connection with North Africa.

[Aberdeen]

Still thinking about the paper, but I do believe, based on things I've read about the history of Wales, that the divide between north and south is very ancient and may be pre-Celtic. I know that the Y DNA of Wales is 74% R1b and 16% I (I1, I2, I2a and I1b) and a bit of R1a so very little Neolithic and possibly Neolithic. It would be interesting to so how that breaks down locally. Although I suppose more detailed subclade information would be needed to get an idea of how long the various bits of R1b and I had been there, since some of the I could have been brought into the area in the post-Roman period by various invaders. But some of the I could be really old and I wonder if there's more of it in the north, or more of one kind of I in the north compared to the south. However, it really does seem as if the Y DNA profile may not correspond very well to the autosomal makeup.

I'm dubious about the supposed lack of Danish DNA in the area that was the Danelaw. Is this conclusion based on an assumption that the DNA of Scandinavia during the Viking period was the same as it is today? I haven't had time to really read carefully yet.

[Greying Wanderer]

Still thinking about the paper, but I do believe, based on things I've read about the history of Wales, that the divide between north and south is very ancient and may be pre-Celtic. I know that the Y DNA of Wales is 74% R1b and 16% I (I1, I2, I2a and I1b) and a bit of R1a so very little Neolithic and possibly Neolithic. It would be interesting to so how that breaks down locally. Although I suppose more detailed subclade information would be needed to get an idea of how long the various bits of R1b and I had been there, since some of the I could have been brought into the area in the post-Roman period by various invaders. But some of the I could be really old and I wonder if there's more of it in the north, or more of one kind of I in the north compared to the south. However, it really does seem as if the Y DNA profile may not correspond very well to the autosomal makeup.

I'm dubious about the supposed lack of Danish DNA in the area that was the Danelaw. Is this conclusion based on an assumption that the DNA of Scandinavia during the Viking period was the same as it is today? I haven't had time to really read carefully yet.

Some parts of North Wales are up to 30% E1b.

http://dienekes.blogspot.co.uk/2011/...marker-in.html

Which is what makes me wonder if they're the remnant of Atlantic Megalith - so maybe neolithic but not the standard neolithic.

[Expredel]

I'm dubious about the supposed lack of Danish DNA in the area that was the Danelaw. Is this conclusion based on an assumption that the DNA of Scandinavia during the Viking period was the same as it is today? I haven't had time to really read carefully yet.

https://en.wikipedia.org/wiki/List_o...rthern_Ireland

Danes were cleared from the English region in 1002, some more cleansing in 1070. Later conflicts in Scotland may have been genocidal as well. Populations have a tendency to vanish mysteriously.

[Aberdeen]

Some parts of North Wales are up to 30% E1b.

http://dienekes.blogspot.co.uk/2011/...marker-in.html

Which is what makes me wonder if they're the remnant of Atlantic Megalith - so maybe neolithic but not the standard neolithic.

Someone found a spot or two where there were high levels of E1b1b, but that haplotype makes up only 4% of the total Welsh Y DNA. I wonder how much of that 4% is found in northern Wales.

[Aberdeen]

https://en.wikipedia.org/wiki/List_o...rthern_Ireland

Danes were cleared from the English region in 1002, some more cleansing in 1070. Later conflicts in Scotland may have been genocidal as well. Populations have a tendency to vanish mysteriously.

Massacres seldom result in a total clearing of populations. I suspect Danish DNA was similar enough to Anglo-Saxon DNA that it's difficult to tell the two apart.

[sparkey]

The thing about Danish DNA is that there is a clear British Isles group with a lot of affinity to Scandinavia: Orcadians. And the Orcadians, as a result, are the greatest outlier among British populations. Unfortunately, I'm not seeing that they used Danes as a reference population, only Norwegians and Swedes, but nonetheless, it seems that they're drawing their conclusions about the effects of the Danelaw based on reference population affinity.

Naturally, drawing specific conclusions about much of anything based on modern reference populations is highly problematic. There's no real temporal aspect in doing so, so there may be other explanations for why Orcadians are much closer to Scandinavians than the rest of the British. But at least this study can count as evidence.

[Greying Wanderer]

Someone found a spot or two where there were high levels of E1b1b, but that haplotype makes up only 4% of the total Welsh Y DNA. I wonder how much of that 4% is found in northern Wales.

Thing is there are two small components: SFS31 and FRA12 which show the clearest links to Iberia and the Mediterranean. It would have been interesting to see if the E1b was connected to one or both of those.

http://1.bp.blogspot.com/---rAR5GpVQ...talSources.jpg (h/t dionekes)

[MOESAN]

Massacres seldom result in a total clearing of populations. I suspect Danish DNA was similar enough to Anglo-Saxon DNA that it's difficult to tell the two apart.

good remark:
1- today Danes are maybe not exactly the same ones as the historic ones, even if the difference would be light -
2- even today (say 1930/1950!), Denmark is/was not completely homogenous by regions, neiher for pigmentation, nor stature nor cephalic index - the Western Jutland shew strong trend towards 'borreby', as do Western Norway (what Y-DNA? what autosomes component according to what poolings?)
3- the Angles/Engles occupied Southern Denmark-Schlesvig so were as 'danish' than 'anglo-saxon' and surely enough were of the same stock than the Anglo-Saxons roughly said - the Jutes remnants examined shew same whole tendancies for crania (I think it is not too precise) but already shew broader mandibules (inf. jaw) than the common means in Northmen of the germanic periods: I see here, without to much doubt (because it is always present among today genuine Danes on whom we can observe physical traits) a 'cromagnon' imput - nevertheless I think all these ancient Danes passed to Britain were not the westernmost typical 'borreby' of West Jutland (pre-I-Ean/pre-Neolithic remnants of fishers by ascendance, at these times maybe not completely germanized nor celticized?) so the comparisons with modern Danes autosomes in not fully accurate???

[giuseppe rossi]

Thing is there are two small components: SFS31 and FRA12 which show the clearest links to Iberia and the Mediterranean. It would have been interesting to see if the E1b was connected to one or both of those.

http://1.bp.blogspot.com/---rAR5GpVQ...talSources.jpg (h/t dionekes)

The Iberian link is only with Catalans and Basques who are not proper Iberians per se.

[MOESAN]

Thing is there are two small components: SFS31 and FRA12 which show the clearest links to Iberia and the Mediterranean. It would have been interesting to see if the E1b was connected to one or both of those.

http://1.bp.blogspot.com/---rAR5GpVQ...talSources.jpg (h/t dionekes)

your hypothesis for Y-E1b is very possible but FRA12 ? I'm not sure I understand well the map but this "component" doesn't seem 'mediterranean' by force? could you explain in a deeper way?

[Vallicanus]

your hypothesis for Y-E1b is very possible but FRA12 ? I'm not sure I understand well the map but this "component" doesn't seem 'mediterranean' by force? could you explain in a deeper way?

Indeed component FRA12 might be called "Core French" or "Core Gallic" as it is found in many parts of France with an overspill into NW Italy down to NW Tuscany.

[Aberdeen]

The thing about Danish DNA is that there is a clear British Isles group with a lot of affinity to Scandinavia: Orcadians. And the Orcadians, as a result, are the greatest outlier among British populations. Unfortunately, I'm not seeing that they used Danes as a reference population, only Norwegians and Swedes, but nonetheless, it seems that they're drawing their conclusions about the effects of the Danelaw based on reference population affinity.

Naturally, drawing specific conclusions about much of anything based on modern reference populations is highly problematic. There's no real temporal aspect in doing so, so there may be other explanations for why Orcadians are much closer to Scandinavians than the rest of the British. But at least this study can count as evidence.

It was the Norwegians, not the Danes, who settled in Orcadia. It was the Danes who settled in the English Danelaw, and I suspect that their DNA was closer to that of the Anglo-Saxons than the Norwegians prior to the medieval political union between Denmark and Norway. If you look at the map, the red dots that supposedly represent Anglo-Saxons match very closely a combination of the Anglo-Saxon heartland and the Danelaw (which did overlap to some extent). There just doesn't seem to be any separation between the two and I suppose one could assume, as the authors seem to, that the Danes just disappeared, presumably after King Canute's last descendent was replaced by an Anglo-Saxon type in 1042. But, when one considers such things as the continuing influence of Old Danish on the accents of people in modern northeastern England, it seems possible that Danish DNA was more like Anglo-Saxon DNA than Norse 1000 years ago, especially given that the Angles appear to have come in part from what is now southern Denmark.

[Angela]

I'm having some trouble deciphering the below graphic. FRA17 has blue bars, but the arrow is pointing to teal colored circles. Or am I color blind?

Leslie et al-British ancestry.JPG

It's the teal colored circles which appear in Piemonte, Liguria and down the coast into Toscana. So is it FRA 17 which is in Italy or FRA12?

The teal circles also appear in what looks like Provence, the coastal Loire?, Orleans? and southern Poitou?
Is there something which ties those regions together, particularly?

This is what they have to say about these clusters in the Supplement:
The pattern in North Wales, North Pembrokeshire, South Pembrokeshire: "Absence of GER3 and FRA17, presence of FRA12, and relatively higher proportions of GER6 and FRA14."

A second general pattern is shared by a number of other UK clusters(from north to south: Northumbria; Cumbria; W Yorkshire, Cent./S England; Welsh Borders; Devon; Cornwall): Significant presence of GER3, absence of FRA12, relatively higher contributions from groups FRA17 and DEN18, and relatively lower contributions from FRA14.

As to FRA17, they have this to say:
"presence/absence pattern (notably its absence from Wales) strongly suggests that it results from a migration or migrations later than those of the earliest migrations which contributed DNA to the modern UK population (GER6, BEK11, FRA14: internal migration has spread the DNA from these early immigrants across the UK, so that if the migrations represented gy FRA17 were earlier than or contemporaneous with these, then the same migrations should also have spread the resulting(FRA17 like)DNA through out the UK, including Wales. We also argue that the FRA17 contribution is unlikely to reflect any of the known movements in histori(i.e.since the Roman invasion of Britain)


Their main argument for the last statement is that none of those invasions were large enough to effect such a wide spread of alleles.

This leads me to think FRA17 is not the one in Italy. FRA14 is, according to the authors, mostly found in northwest France (although it appears in other places on the French map.) Therefore, does it seem more likely that it is FRA12 which is present in Italy?

If that is correct, it's interesting that it is basically only present in Wales. I hesitate to ascribe it to Roman legionnaires (many of whom came from northern Italy and Gaul) as per the thread on E-V13. If they impacted the genetics of Wales, why not that of the area around Hadrian's Wall, or on the Saxon Shore? The authors maintain that the most widespread components are oldest. I'm not so sure that would always be the case. Couldn't the oldest component have been pushed into Wales by subsequent invasions?

As to the presence in Italy there are numerous explanations, from shared Neolithic ancestry to the documented Gallic migrations in the first millennium BC into Piemonte and Liguria specifically from the west.