Case Report: The British & Sardinian connection

[Riverman]

Case Report: The British & Sardinian connection

With better sampling we can better reconstruct the migration and expansion of E-V13 lineages. The best NGS sampled areas up to now are the British Isles, Sardinia (because of one major study done on the island population) and Albanians in moderns and Early Medieval Hungary.

Looking at the data on FTDNA and YFull, there is a non random relationship between Hungarian samples from the early Avar and early Hungarian period, with Western people and a particularly close one of Sardinians and British E-V13 carriers, to the, largely, exclusion of the Southern Central Balkans in a specific time window.

Like we can reconstruct with time windows: When did these E-V13 carriers migrate out of Hungary, when did they reach Sardinia and when, sometimes even how, did they reach the British Isles.

There are many examples of this peculiar pattern of British lineages sharing their most recent ancestor with Sardinian ones, nearly always to the exclusion of the Balkans, sometimes, but not always, with the inclusion of other Italians, Germans and French.
It is also noteworthy that the Sardinians are likely not the primary source on the Westward route, but these were North Western Italian people, mostly from the area of Liguria, Genua, from where the majority of these Sardinian lineage is supposed to come from, but from which we have simply not as many samples.

Here some of the definining branches in question (TMRCA is always just the average, range must be considered; also, I used the just a relevant subclade identification for the FTDNA Discovery browser, Time Tree view, not necessarily the final subclade in question) - everyone can easily check on: https://discover.familytreedna.com/) :

E-BY6162 (TMRCA of 476 BC between the Hungarian-Avar and Italian)
E-FTB70721 (TMRCA of 60 AD between British and Sardinian)
E-Z21367 (TMRCA of 500 BC between British/Irish/German and Hungarian-Avar and Sardinian, Sardinian diversify 200 AD)
E-Z21291 (TMRCA of 624 BC between French, British and Sardinian, with Sardinian diversification 200 BC)
E-BY4991 (TMRCA of 143 AD between German, Hungarian Arpad and British)
E-FT92232 (TMRCA of 417 BC between British and Italian)
E-BY159978 (TMRCA of 398 BC between British and Italian)
E-BY6263 (TMRCA of 317 BC between British and Swiss)
E-FTA40200 (TMRCA of 456 BC between British and modern Hungarian)
E-BY4404 sticks out as one of the few Eastern Balkan connections with (TMRCA of 153 BC between British and Bulgarian, Turkish, not sure if its an ethnic English lineage, because of an unidentified recent branch member)
E-FT106938 (TMRCA of 760 BC between Irish and modern Hungarian)

E-Z16988 is particularly interesting, because it combines in different of its subclades ancient Hungarian, Sardinian-Italian, German, Serbian and British lineages:
E-Z21350 (TMRCA of 576/552 BC between German and English, which start to diversify 826 AD, and Irish, Polish and Sardinian, with Sardinian starting to branch 50 AD)
E-A11837 (TMRCA of 601/448 BC between Serbian and Sardinian which diversify 500 AD, and Serbian and Italian)

What are the big takeaways from this data?
- There is very little to no overlap with the more Southern and Eastern Balkans and these British, Sardinian-Italian branches after the initial E-V13 dispersion in the Transitional Period (younger than 900 BC).
- Most of the later overlap with other macro-regoins, in ancients and moderns, is with Hungarians and Serbs dating to about 500-400 BC. Therefore it looks that the source group for these E-V13 branches was in the area of Hungary-Western Romania-Northern Serbia about 500-400 BC.
- The branching in North West Italia started likely earlier, but the proven diversification in situ starts in Sardinia/North Western Italia between about 300 BC-100 AD.
- The British branched off from these Sardinian-Italian branches about 100 BC-100 AD.

This means it looks to me like a good portion of these British E-V13 ancestors lived in or close to North Western Italy 300 BC-100 AD. The start of the local founding events in North Western Italy overlaps with the last common ancestors of the British and North Western Italians.

When these E-V13 lineages entered Britain is unknown, but the latest date for their arrival is definitely before 800 AD for many, with potential stop in or around Germany and France in between. That's not known.

This leaves a variety of options of how these modern British E-V13 lineages migrated, but it makes a stop of their ancestors in or close to North Western Italy-Switzerland highly likely, which is a curious thing in itself I'd say. Possible explanations are Late Hallstatt-Vekerzug into La Tene migrations (backflow from the Carpathian Basin and North Balkan?), to Romans assimilating those North Italians or transfering people from the Carpatho-Balkan/Middle Danubian zone towards North Western Italy in that time frame, which later migrated onwards to Britain.

Because of the scope of this, we see it in many British lineages which have sufficient testing and can be put into a network of non-British testers, I would rather suggest a Celtic La Tene backflow event, after Late Hallstatt-Vekerzug, Illyrian and Thracian areas integration into the La Tene Celtic sphere.

Also noteworthy: For these lineages is more overlap with areas like Poland other than any other Balkan country than Serbia, which is closest ot the Danubian region. There is no other noticeable, recent Balkan overlap for these branches. Probably somebody knows one, but it won't change he overall impression for the great majority.

This also shows how far the reconstruction of pathways can go with enough modern and ancient DNA testing combined. Because things got really narrowed down for some of these branches, like fairly recent (AD) TMRCA dates for the English-Irish vs. Sardinian-Italian lineages.

Now it would be interesting to explore whether we can associate a specific event(s) and movement(s) of people from North Western Italy with an arrival in Britain or alternative scenarios to explain the observable pattern.

[Riverman]

To cross check other haplogroups, I looked up R-L2 a bit, and while many subclades don't go in the same direction, there might be some which do. One example being
R-FT40813, which doesn't show such a clear more recent overlap, but still is a branch which includes Hungarian ancient, Irish, German, Czech, Italian and Sardinian. The TMRCA of 1.187 BC is of course too high. Czech, German and Sardinian split around 814 BC, which is a Hallstatt timing.
R-FGC36273, which has again a too high TMRCA of 1.925 BC, but curiously brings together Sardinian, English and Romanian, which might be no ethnic Romanian.
R-S1480 is too high as well, TMRCA of 1.449 BC, with a big Sardinian subclade founder effect starting 850 BC.


As far as I can see, R-L2 doesn't show the same pattern to the same extend like E-V13 at all. There is just a much smaller, but significant, earlier influx into North Western Italy-Sardinia at the start of the Hallstatt period. So we see the association with Hallstatt-La Tene as well, but rather independent, earlier and much lower level migrations than for E-V13.

For J-L283:
J-YP29 shows an overlap in the same direction, but no such recent one for all branches. There is however a Sardinian founder lineage dated to 450 BC.
J-YP157 is an older local Sardinian founder lineage, that ha nothing to do with Iron Age migrations.

Curiously, even though J-L283 has a much older presence in Sardinia, its much less well-represented in modern Sardinians and there is just one case for an Iron Age entering point. No such overlap of Hungary-Italy-British Isles at all. It is missing and not comparable to E-V13.

To cross check other haplogroups, I looked up R-L2 a bit, and while many subclades don't go in the same direction, there might be some which do. One example being
R-FT40813, which doesn't show such a clear more recent overlap, but still is a branch which includes Hungarian ancient, Irish, German, Czech, Italian and Sardinian. The TMRCA of 1.187 BC is of course too high. Czech, German and Sardinian split around 814 BC, which is a Hallstatt timing.
R-FGC36273, which has again a too high TMRCA of 1.925 BC, but curiously brings together Sardinian, English and Romanian, which might be no ethnic Romanian.
R-S1480 is too high as well, TMRCA of 1.449 BC, with a big Sardinian subclade founder effect starting 850 BC.

As far as I can see, R-L2 doesn't show the same pattern to the same extend like E-V13 at all. There is just a much smaller, but significant, earlier influx into North Western Italy-Sardinia at the start of the Hallstatt period. So we see the association with Hallstatt-La Tene as well, but rather independent, earlier and much lower level migrations than for E-V13.

For J-L283:
J-YP29 shows an overlap in the same direction, but no such recent one for all branches. There is however a Sardinian founder lineage dated to 450 BC.
J-YP157 is an older local Sardinian founder lineage, that ha nothing to do with Iron Age migrations.

Curiously, even though J-L283 has a much older presence in Sardinia, its much less well-represented in modern Sardinians and there is just one case for an Iron Age entering point. No such overlap of Hungary-Italy-British Isles at all. It is missing and not comparable to E-V13.

For J2a I checked e.g.
J-Z467 = No
J-Z7671 (the Kyjatice branch) = No
J-S23154 (Minoan) = No
J-Z6064 = Generally no, but one Sardinian founder J-PF7415 from 350 BC. Upstream are French, German and Spanish, with two ancient samples from Italy (Himera and Isola Sacra).


Therefore going by these three examples with a presence in the Carpatho-Balkan region, this Iron Age-Early Roman expansion into North Western Italy-Alpine Central Europe and from there to the British Isles is pretty E-V13 specific, especially in this magnitude.

[Riverman]

There are multiple big R-U152 founder effects dating to the Hallstatt period, like below: R-PF6658
Very big Late Urnfield and into Hallstatt and La Tene period founder events in Sardinia, very big!

Actually R-U152, and R-PF6658 specifically, shows the closest relationship of the so far checked haplogroups (R-L2, J-L283, J2a branches).

Therefore so far it looks like E-V13 came into R-U152 territory of Alpine Europe-Northern Italy, not just a couple of people, but a greater number, and along them, within networks the R-U152 people already had established since Urnfield, they expanded too?

Very curious pattern, very curious. Makes the Sardinian case all the more interesting. Any other suggestions for a haplogroup which might share a similar pattern in Northern Italy-Sardinia and Britain?

[rafc]

I think looking in detail at the modern distribution of V13-branches, or to what is found in ancient DNA, could really help in understanding how V13 dispersed. Unfortunately any attempt to do so will immediately result in people trying to explain the results through the lens of their personal pet-theory, so I have largely given up on that.
I'm happy you launched another try, but I think there is a flaw in the analysis. You see overlaps with the Avar-age Hungarians, and that might well be true. However, this area is very well sampled in this period, while others are not. How can we say that British samples are closer to Avar age Hungarians than other parts of the Balkans, if we don't have samples from other regions in this period? It's very well possible some of them were closer ...

[Bruzmi]

I think looking in detail at the modern distribution of V13-branches, or to what is found in ancient DNA, could really help in understanding how V13 dispersed. Unfortunately any attempt to do so will immediately result in people trying to explain the results through the lens of their personal pet-theory, so I have largely given up on that.
I'm happy you launched another try, but I think there is a flaw in the analysis. You see overlaps with the Avar-age Hungarians, and that might well be true. However, this area is very well sampled in this period, while others are not. How can we say that British samples are closer to Avar age Hungarians than other parts of the Balkans, if we don't have samples from other regions in this period? It's very well possible some of them were closer ...

This sums up my thoughts as well.

There are many haplogroups where such connections among modern samples exist, but projecting them to the past without aDNA evidence leads to faulty scenarions. Even at a basic level the statement:

E-FT92232 (TMRCA of 417 BC between British and Italian)

doesn't mean that in 417 BCE this clade split between a British and Italian branch. It just means that a modern Italian and a modern Brit descend from a common ancestor who lived in 417 BCE. Where this ancestor lived and where the ancestors of the Italian and the Brit lived is an entirely different discussion which can't be determined without aDNA or very extensive testing of this clade.

IMO, the starting point for Britain is that no E-V13 has been found so far in the IA and for the rest we can wait for aDNA studies.

[Riverman]

I think looking in detail at the modern distribution of V13-branches, or to what is found in ancient DNA, could really help in understanding how V13 dispersed. Unfortunately any attempt to do so will immediately result in people trying to explain the results through the lens of their personal pet-theory, so I have largely given up on that.
I'm happy you launched another try, but I think there is a flaw in the analysis. You see overlaps with the Avar-age Hungarians, and that might well be true. However, this area is very well sampled in this period, while others are not. How can we say that British samples are closer to Avar age Hungarians than other parts of the Balkans, if we don't have samples from other regions in this period? It's very well possible some of them were closer ...

There are many arguments in this respect:
- Hungarians are not better sampled than the Balkans if considering FTDNA and YFULL combined, but even in moderns the Hungarians predominate very clearly for these branches among the next relatives East of the Alps. There is no reason why these subclades, out of so many Balkan ones, did survive only or at least primarily in Hungary. If such major branches would be common in the Balkans, they would appear. Apparently there is one branch with a fairly recent TMRCA, towards the South Eastern Thracians, but this branch is clearly an outlier among the British, especially among the British with a Sardinian-Italian connection. At this point of sampling density, overlooking major branches becomes less and less likely, but rather we have to ask if single individuals or small recent branches appear, whether they migrated South late too.
- The ancient distribution suggests not necessarily Hungary alone, if putting the regional groups into context, but also areas like Western Romania and Northern Serbia, which is something I said from the beginning, it looks Carpatho-Balkan. So it might be present in say Northern Serbia, but it just doesn't look like it was in e.g. Albania, Macedonia, Bulgaria, Greece etc. primarily, because there is no ancient nor modern overlap of significance.
- The respective upstream branches are also known from areas like the Caucasus, Central Asia and Northern China, so they seem to have had a position close to the steppe
- Also, the TMRCA between the ancient and modern samples and the branches they form is too close for large scale additional jumps. Like if they had to move to Hungary from the Balkans first, it makes no sense. I'm not even saying they can't have been from say Bulgaria in the Early Iron Age, but in the time before the split, they were in or around Hungary. The most likely candidates for these groups and their distribution is very clearly Basarabi, Eastern Hallstatt and Vekerzug.

This sums up my thoughts as well.

There many haplogroups where such connections among modern samples exist, but projecting them to the past without aDNA evidence leads to faulty scenarions. Even at a basic level the statement:

doesn't mean that in 417 BCE this clade split between a British and Italian branch. It just means that a modern Italian and a modern Brit descend from a common ancestor who lived in 417 BCE. Where this ancestor lived and where the ancestors of the Italian and the Brit lived is an entirely different discussion which can't be determined without aDNA or very extensive testing of this clade.

IMO, the starting point for Britain is that no E-V13 has been found so far in the IA and for the rest we can wait for aDNA studies.

The question is whether there is a local branching event and diversity of significance. This is the case for a couple of Sardinian branches, they have too many branches and diversity, for it to be explained away. Like if we find a branch with 2-3 members since 400 BC, it could still mean they migrated into Sardinia later, but seriously, what are the chances this split happened in the Balkans and out of all of European sampling, those two closely related lineages end up exclusively in Sardinia? They could very well be present in say North Western Italy around Genua, where there are many related branches, of which many didn't test for the BigY unfortunately, or from Switzerland, but surely not Bulgaria or Southern Serbia.

We see the same for other haplogroups as well, if they have multiple branching events in the same time frame, we can, usually, assume they appeared in that region around that time the latest. Such branching events, usually, mark a terminus ante quem. Like for the British we have such branching events which safely put e.g. English lineages in a local context mostly from around 800 AD. There are many going back much further, even the Urnfield period, but they don't show that pattern of exclusively regional branches which are very close in time to each other.

But we see that here in Sardinia which suggest, very strongly so, that these lineages, before the split with the British lineages ancestors, did live in or close to Sardinia, presumably in North Western Italy, at least in Alpine Central Europe. And the branching event form the Hungarians, both modern and ancient, is again so close in time, also between the Hungarians themselves sometimes, that we have to assume a departure from a region close to Hungary, not far removed from it, but very close (like Western Romania and Northern Serbia included).

The data is no dense enough to allow such conclusions and make alternatives extremely unlikely.

Beside, from that time frame, just to repeat it once more, we have very little sampling and the local population nearly exclusively cremating. For some Dacian branches, later ancient samples and modern phylogeny might be ALWAYS the only thing we might get, because they continuously cremated. They won't appear in sufficient numbers and diversity, to account for the E-V13 population diversity and size, at any point with the possible exception of short time windows like Mezocsat and Basarabi groups which which switched to inhumation.

[Aspar]

Let's also keep in mind that the 'Avar' individuals from Hungary under E-V13 have mostly been placed in EU_Core1 and EU_Core2 clusters on the PCA comprising populations from IA Thrace and Imperial Romans to IA Illyrians.

In order to make a local origin for these samples provable you would need to give us as many as possible older samples from Hungary.
We only have two such, from La Tene period and one of them was as Balkan IA as he could be.

[Riverman]

Let's also keep in mind that the 'Avar' individuals from Hungary under E-V13 have mostly been placed in EU_Core1 and EU_Core2 clusters on the PCA comprising populations from IA Thrace and Imperial Romans to IA Illyrians.

In order to make a local origin for these samples provable you would need to give us as many as possible older samples from Hungary.
We only have two such, from La Tene period and one of them was as Balkan IA as he could be.

Again the sampling is already pretty dense if combining both ancient and modern Hungarians. There is no other region or group which comes even remotely close to this source. And nobody can claim that Hungary is so much better sampled than say Serbia or Albania.

You also know that Basarabi was influenced by Psenichevo-Insula Banului, with an origin in Southern Romania and the Central Balkans, but they expanded upwards while assimilating or replacing locals. If assuming a Basarabi origin, which is the most likely thing, I don't expect them to be particularly Northern and rather Balkan by that time, after continuous exchange with the South East. Not as South Eastern as the Southern Thracians, but likely as much or more South Eastern than the average Illyrian.

By the time of the Avar samples, we deal with different clusters, very clearly so, and the most Southern one seems to have incorporated generally speaking South Eastern Thracian and "Roman era" influences.

But all of this is really of secondary importance, because we know that large portions of the locals were cremating, we have zero (male) samples from Mezocsat locals, Basarabi, Eastern Vekerzug, any clearly Daco-Thracian related context, not even locals from the La Tene context, just typical early, arriving Celts. So there is a huge sampling gap, but the ancient and modern data combined is clear about the association.

We just need to combine some very close branches:
E-FTB70721 TMRCA of British and Sardinian of around 60 AD
E-BY193951 TMRCA of Irish, ancient Hungarian and Sardinian of 500 BC with 3 (!) branches/individuals among ancient Hungarians, 3! The Sardinians have again 3 independent branches which date to around 200 AD! Upstream of this subclade we have Belgian, German, English and Polish (likely German) as well, going back to 758 BC. Chances that this upstream branch E-Z21365 was related to an Hallstatt-La Tene expansion is just extremely likely, extremely likely.

E-S2978 shows a similar pattern of early Sardinian split, otherwise Central European distribution.

Its not like the well sampled Balkan branches are that far away, if you look at e.g. E-BY166132, but they have no depth with a TMRCA to to these British-North Italian branches no earlier than the Transitional Period.

The Serbian branch is a clear exception and shows the North Balkan-Danubian proximity, probably even deeper origin, but by Hallstatt there is no doubt that these lineages should have lived closer to where they were found later, namely in or close to Hungary-Western Romania-Northern Serbia.
We got samples from other regions which don't reflect the same frequencies and diversity like the Eastern Hungarian sites in the Avar period, its not like its just sampling bias only, can't be if looking at the data.