Ancient genomes from present-day France unveil 7,000 years of its demographic history

[Regio X]

https://www.pnas.org/content/early/2020/05/20/1918034117

Ancient genomes from present-day France unveil 7,000 years of its demographic history

Abstract

Genomic studies conducted on ancient individuals across Europe have revealed how migrations have contributed to its present genetic landscape, but the territory of present-day France has yet to be connected to the broader European picture. We generated a large dataset comprising the complete mitochondrial genomes, Y-chromosome markers, and genotypes of a number of nuclear loci of interest of 243 individuals sampled across present-day France over a period spanning 7,000 y, complemented with a partially overlapping dataset of 58 low-coverage genomes. This panel provides a high-resolution transect of the dynamics of maternal and paternal lineages in France as well as of autosomal genotypes. Parental lineages and genomic data both revealed demographic patterns in France for the Neolithic and Bronze Age transitions consistent with neighboring regions, first with a migration wave of Anatolian farmers followed by varying degrees of admixture with autochthonous hunter-gatherers, and then substantial gene flow from individuals deriving part of their ancestry from the Pontic steppe at the onset of the Bronze Age. Our data have also highlighted the persistence of Magdalenian-associated ancestry in hunter-gatherer populations outside of Spain and thus provide arguments for an expansion of these populations at the end of the Paleolithic Period more northerly than what has been described so far. Finally, no major demographic changes were detected during the transition between the Bronze and Iron Ages.

Supplementary Information: https://www.pnas.org/content/pnas/su...34117.sapp.pdf

[Regio X]

[bigsnake49]

Huge changes between Neolithic and Iron Age. Talking about Y-haplogroup replacement!

[Maciamo]

Wow! That's one of the most surprising study in the last few years. I haven't checked the paper yet. (its behind a paywall - does anybody have it?) From what you posted there are a few things that stick out immediately:

1) Y-haplogroup E1b1b is almost as high as G2a in Neolithic France. That would be the very first place in Europe where E1b1b is more than a minor haplogroup.

2) Y-haplogroup I1 is present in Neolithic and Iron Age France! Now is that pre-I1 like in Neolithic Hungary or actual I1. The latter would be revolutionary as modern samples of I1 have a TMRCA of only 4600 years. This would either mean that French I1 is ancestral to modern Scandinavian/Germanic I1 (unlikely) or that all Neolithic I1 from France got wiped out. What about Iron Age I1? That would mean that the Gauls also carried I1. Are we talking about north-eastern tribe who possibly mixed with Germanic tribes? If so, that's not all that surprising. But if it's elsewhere (unlikely) that would have us rethink the whole I1 history.

3) A substantial number of C1a2 in Bronze Age France. I wonder if they survived in one specific region.

[Ailchu]

you can access the supplementals by clicking on "Figures&Si" then "download appendix".

[Regio X]

Wow! That's one of the most surprising study in the last few years. I haven't checked the paper yet. (its behind a paywall - does anybody have it?) From what you posted there are a few things that stick out immediately:

1) Y-haplogroup E1b1b is almost as high as G2a in Neolithic France. That would be the very first place in Europe where E1b1b is more than a minor haplogroup.

2) Y-haplogroup I1 is present in Neolithic and Iron Age France! Now is that pre-I1 like in Neolithic Hungary or actual I1. The latter would be revolutionary as modern samples of I1 have a TMRCA of only 4600 years. This would either mean that French I1 is ancestral to modern Scandinavian/Germanic I1 (unlikely) or that all Neolithic I1 from France got wiped out. What about Iron Age I1? That would mean that the Gauls also carried I1. Are we talking about north-eastern tribe who possibly mixed with Germanic tribes? If so, that's not all that surprising. But if it's elsewhere (unlikely) that would have us rethink the whole I1 history.

3) A substantial number of C1a2 in Bronze Age France. I wonder if they survived in one specific region.

Also interesting this relative growth of mtDNA U5 from EN to LN in North France.

[Maciamo]

you can access the supplementals by clicking on "Figures&Si" then "download appendix".

Yes, I saw that. But there is no data about the Y-DNA. It's usually in an Excel file listing the samples with their location, age and detailed haplogroup clade. It would be great if someone had access to it.

[bicicleur]

Wow! That's one of the most surprising study in the last few years. I haven't checked the paper yet. (its behind a paywall - does anybody have it?) From what you posted there are a few things that stick out immediately:

2) Y-haplogroup I1 is present in Neolithic and Iron Age France! Now is that pre-I1 like in Neolithic Hungary or actual I1. The latter would be revolutionary as modern samples of I1 have a TMRCA of only 4600 years. This would either mean that French I1 is ancestral to modern Scandinavian/Germanic I1 (unlikely) or that all Neolithic I1 from France got wiped out. What about Iron Age I1? That would mean that the Gauls also carried I1. Are we talking about north-eastern tribe who possibly mixed with Germanic tribes? If so, that's not all that surprising. But if it's elsewhere (unlikely) that would have us rethink the whole I1 history.

We know that the Belgae were not 100 % Celtic, they were already a mix of Celtic and Germanic tribes.
Some Germanic tribes may have infiltrated further south too, like Ariovistus did in Ceasar's time.
The first Germanic tribes we know of are the Cimbri and the Teutones, because the Romans new them.
Many other tribes may have been wandering around without us knowing it.

[Angela]

What they said in regard to MtDna N1a could perhaps be applied to all uniparentals, i.e. any lineage which was less than 25% of the total has a 95% chance of not being picked up.

Differences between northern and southern France go back to the Neolithic. That’s sometimes obscured in their graph covering broad areas and time periods.For example, look at all the U5b in Michelsberg culture.

I think they’re right to question the origin of these mitochondrial lineages. Perhaps, along with some “local” resurgence, we also have people moving in from further north.

So perhaps we could say that pre-Bronze Age, the south was fairly homogeneous in terms of mtDna but the north was heterogeneous with quite a few areas showing more U5b.

I think this fits in with a scenario where there just weren’t as many Mesolithic hunter-gatherers in Southern Europe, and so the Neolithic “take over” was pretty complete.

With the Bronze Age come I, U2, and U5a. So, some women came to France with the steppe admixed groups, unlike what seems to have happened in Spain.
Interesting also in this regard are the Bell Beakers. Perhaps it’s a function of small sample sizes, but there’s very little U of any kind; it’s almost all H and K.

Then the U5a and U5b go up in the Bronze Age (continuing migration?) to go back down in the Iron Age.

The Iron Age is interesting in another way in that H and J go way up. They point out the similarly high levels of H on the Pontic steppe and the J in central Europe Neolithic. Is that a hint for them of another population movement from more Central Europe?

[Angela]

I wish they had included Tuscans on their analysis. Oh well. :)

The variation in Albanians is interesting given how early studies said they were so homogeneous.

I’m assuming the French samples which plot so close to Northern Italy are from southern France. I’d like to look at the location for each sample, however. Are they from southeast France, south Central?

Note how one Bell Beaker sample is very far north indeed, plotting near Icelandic and Finnish people, two are down around Sardinians and one near western Spaniards, and a few in modern France. So, not a lot of homogenization. This reminds me of the Beaker “phenomenon” in Italy, i.e very variable genetic change in more southern regions.

Only with the Bronze Age do the southern French move a bit north and quite a bit east into the areas of the modern Spanish and the Basques and even some modern French. The change in the Basque areas seems to be Iron Age. Could this be when the Basques acquired their distinctive admixture? We’ll have to look at the ydna.

Interesting how much variety there is also in the Eastern French Bronze Age 1: some samples plot north and east of modern France, and some right within what I am assuming is Southern France.

I say that because there is a group of French samples plotting north of the areas of France where Bronze and Iron Age samples plot. Is that northern France showing the impact of the Germanic invasions and/or Brittany showing the Britons back migrating?

As to northern Italy, some of them are very close to France Iron Age and even approaching France Bronze Age. If it’s just Val D’Aosta it doesn’t really count imo, however.

[bicicleur]

[Angela]

Maybe the Iron Age had slightly less Yamnaya? Or is it just that it stabilized?

Where is the Ydna E?

[kingjohn]

Regio thanks for sharing👍
Mtdna h3 versions :
H3q1 in early neolithic alsace
H3c2 in bronze age france

[bigsnake49]

I wish they had included Tuscans on their analysis. Oh well. :)

The variation in Albanians is interesting given how early studies said they were so homogeneous.

I’m assuming the French samples which plot so close to Northern Italy are from southern France. I’d like to look at the location for each sample, however. Are they from southeast France, south Central?

Note how one Bell Beaker sample is very far north indeed, plotting near Icelandic and Finnish people, two are down around Sardinians and one near western Spaniards, and a few in modern France. So, not a lot of homogenization. This reminds me of the Beaker “phenomenon” in Italy, i.e very variable genetic change in more southern regions.

Only with the Bronze Age do the southern French move a bit north and quite a bit east into the areas of the modern Spanish and the Basques and even some modern French. The change in the Basque areas seems to be Iron Age. Could this be when the Basques acquired their distinctive admixture? We’ll have to look at the ydna.

Interesting how much variety there is also in the Eastern French Bronze Age 1: some samples plot north and east of modern France, and some right within what I am assuming is Southern France.

I say that because there is a group of French samples plotting north of the areas of France where Bronze and Iron Age samples plot. Is that northern France showing the impact of the Germanic invasions and/or Brittany showing the Britons back migrating?

As to northern Italy, some of them are very close to France Iron Age and even approaching France Bronze Age. If it’s just Val D’Aosta it doesn’t really count imo, however.

Where do you see these observations about the Basques and Albanians?

[Duarte]

Where do you see these observations about the Basques and Albanians?

@bigsnake49. I think it was here:

[Angela]

Yes, it's page 36 of the Supplement. All the good stuff is always in the Supplement and the Figures. :)

The list of mtDna per sample is in Table 5.
https://www.pnas.org/content/early/2...b-figures-data

I only skimmed it but I don't see any H in northern France until the Iron Age. Is that right?

Well, well, my U2e2 (one mutation difference) shows up in Bronze Age France in BIS382.

The sample carrying the "C" ydna is from SCHW-72-15.

Table S9 is very interesting. It gives proportions of the three population mix using QpAdm. Unfortunately, they just use the sample numbers, not the country/culture designation, so it's going to be a pain going back and forth.

S10 finally the E1b1b samples. They're from BERG. I checked out the numbers and they all seem to be EN Michelsberg. How interesting. Maybe they lasted longer in the Tyrol and southern Germany. So, a REALLY long presence there perhaps if the hints are pointing me in the right direction; not from the Bronze Age.

BERG157-4	E1b1b
BERG157-3	E1b1b1a1
BERG157-9	E1b1b1a1
BERG79	E1b1b1a1

I also don't see any incoming J2b, but maybe the migration into the Balkans was different? Maybe it was picked up in "Old Europe" eastern settlements?

According to their PCA of modern populations, there is quite a bit of variation in the French samples, too. One group is close to what is probably the Spanish Basques, so perhaps Gascony? Some are close to Northern Italian populations. The bulk stretches north/south over quite some distance.

I have to check their list of modern Italian samples, because Tuscany might be in there, but then the "gap" is perhaps Central Italy?

This is all from a quick first run through. Feel free to correct where I'm going wrong.

[bigsnake49]

Yes, it's page 36 of the Supplement. All the good stuff is always in the Supplement and the Figures. :)

The list of mtDna per sample is in Table 5.
https://www.pnas.org/content/early/2...b-figures-data

I only skimmed it but I don't see any H in northern France until the Iron Age. Is that right?

Well, well, my U2e2 (one mutation difference) shows up in Bronze Age France in BIS382.

The sample carrying the "C" ydna is from SCHW-72-15.

Table S9 is very interesting. It gives proportions of the three population mix using QpAdm. Unfortunately, they just use the sample numbers, not the country/culture designation, so it's going to be a pain going back and forth.

S10 finally the E1b1b samples. They're from BERG. I checked out the numbers and they all seem to be EN Michelsberg. How interesting. Maybe they lasted longer in the Tyrol and southern Germany. So, a REALLY long presence there perhaps if the hints are pointing me in the right direction; not from the Bronze Age.

BERG157-4	E1b1b
BERG157-3	E1b1b1a1
BERG157-9	E1b1b1a1
BERG79	E1b1b1a1

I also don't see any incoming J2b, but maybe the migration into the Balkans was different? Maybe it was picked up in "Old Europe" eastern settlements?

According to their PCA of modern populations, there is quite a bit of variation in the French samples, too. One group is close to what is probably the Spanish Basques, so perhaps Gascony? Some are close to Northern Italian populations. The bulk stretches north/south over quite some distance.

I have to check their list of modern Italian samples, because Tuscany might be in there, but then the "gap" is perhaps Central Italy?

This is all from a quick first run through. Feel free to correct where I'm going wrong.

There is a lot of variation in the French, Italian and Greek samples. I wonder where the Greek sample that's way out there by itself is from?

[kingjohn]

Yes, it's page 36 of the Supplement. All the good stuff is always in the Supplement and the Figures. :)
The list of mtDna per sample is in Table 5.
https://www.pnas.org/content/early/2...b-figures-data
I only skimmed it but I don't see any H in northern France until the Iron Age. Is that right?
Well, well, my U2e2 (one mutation difference) shows up in Bronze Age France in BIS382.
The sample carrying the "C" ydna is from SCHW-72-15.
Table S9 is very interesting. It gives proportions of the three population mix using QpAdm. Unfortunately, they just use the sample numbers, not the country/culture designation, so it's going to be a pain going back and forth.
S10 finally the E1b1b samples. They're from BERG. I checked out the numbers and they all seem to be EN Michelsberg. How interesting. Maybe they lasted longer in the Tyrol and southern Germany. So, a REALLY long presence there perhaps if the hints are pointing me in the right direction; not from the Bronze Age.

BERG157-4	E1b1b
BERG157-3	E1b1b1a1
BERG157-9	E1b1b1a1
BERG79	E1b1b1a1

I also don't see any incoming J2b, but maybe the migration into the Balkans was different? Maybe it was picked up in "Old Europe" eastern settlements?
According to their PCA of modern populations, there is quite a bit of variation in the French samples, too. One group is close to what is probably the Spanish Basques, so perhaps Gascony? Some are close to Northern Italian populations. The bulk stretches north/south over quite some distance.
I have to check their list of modern Italian samples, because Tuscany might be in there, but then the "gap" is perhaps Central Italy?
This is all from a quick first run through. Feel free to correct where I'm going wrong.

Yes angela
Also in table S1:
They were anlaysed in capture method
3 E- m78*
1 E-m215* ( which is wierd probably another
E-m78 just low resulution)

[Regio X]

Our data have also highlighted the persistence of Magdalenian-associated ancestry in hunter-gatherer populations outside of Spain and thus provide arguments for an expansion of these populations at the end of the Paleolithic Period more northerly than what has been described so far.

@Bicicleur
This is interesting.

@Angela
I see mtDNA H in other phases, yes.

Also here they show up (apparently only the samples that have a decent coverage for Autosomal are listed below, which would explain why some Y-DNAs - and likely mtDNAs - are lacking):




Am I wrong or we're seeing two possible "replacements"? Firstly, R1b1a1a (R1b-M73) Bell Beakers replacing I2, G2a, E1 etc., then R1b1a1b (R1b-M269) replacing R1b1a1a in Bronze Age. Curiously, the Bell Beakers R1b1a1a-M73 shows up with a G2a2b2b-PF3359 (this one has no WHG ancestry), and both are uncommon nowadays (the more common type of G2a in Europe now is G2a2b2a-P303).
It seems the bulk of the R1b-M269 was R1b-P312. There's also a confirmed R1b-DF27 and a R1b-Z290.

Is C1a2 the same clade of the Epipaleolithic Anatolian Hunter-Gatherer? This French C1a2 from Early Neolithic has only Anatolian Neolithic ancestry.

Concerning Autosomal, there is indeed a "resurgence" of WHG in Middle Neolithic.

And "if" the samples are classified purely by age in each phase and area, it seems to be a tendency for Steppe ancestry to increase along Bell Beaker period, decrease along Bronze Age and increase again along the Iron Age.

[Angela]

@Bicicleur
This is interesting.

@Angela
I see mtDNA H in other phases, yes.

Also here they show up (apparently only the samples that have a decent coverage for Autosomal are listed below, which would explain why some Y-DNAs - and likely mtDNAs - are lacking):




Am I wrong or we're seeing two possible "replacements"? Firstly, R1b1a1a (R1b-M73) Bell Beakers replacing I2, G2a, E1 etc., then R1b1a1b (R1b-M269) replacing R1b1a1a (https://www.yfull.com/tree/R-M478/) in Bronze Age. Curiously, the Bell Beakers R1b1a1a-M73 shows up with a G2a2b2b-PF3359 (this one has no WHG ancestry), and both are rare nowadays (the more common type of G2a in Europe now is G2a2b2a-P303).

Is C1a2 the same clade of the Anatolian Hunter-Gatherer? This French one from Early Neolithic has only Anatolian Neolithic ancestry.

Also, notice that, "if" the samples are classified by age, Steppe ancestry are increasing along Bell Beaker period, and then it decreases along Bronze Age.

I was talking strictly about northern and eastern mtDna from the Table where the mtDna is listed for every sample, not just the ones complete enough for autosomal analysis. I was still overstating. To rephrase, there's a paucity of mtDna "H" as a percentage of the total, except in southern France Middle Neolithic, until the Bronze and Iron Ages. MtDna "K" seems to be the biggest player, although there are others. Some of that "H" which then shows up is, according to the authors, Pontic Steppe in origin.

This points back to the controversy between the Reich Lab (Lazardis) and another researchers as to whether the steppe people brought women with them. According to these authors, that's a yes for France and a no for Spain.

I went back to look at the mtDna for those E1b1b samples:

3-K1a 195
4-X2b
9-K1a3a
79 -K1a2c

[Duarte]

Probably these are the 3(three) SF-BA2, the 2(two) SF-IA1 and 1(one) SF-BA1 that are closer to Iberians:



Comparatively with samples of same age, a little more of WHG and a little bit less of yamna. Am I wrong?

[Angela]

@bigsnake49. I think it was here:

Further south and far to the east. Pontic Greek?

It's hard to distinguish between the Albanians and Mainland Greeks too. Couldn't they have used contrasting colors??? Maybe the Albanians are indeed mostly in one spot: just east of Tuscans, along with northern Greeks, but running up toward Bulgarians?

What do you think? I'm slightly color blind they tell me. I know, odd for a woman, but there you have it. My Dad was severely color blind, so I blame him. No wonder he liked sculpting more than painting. :)

[Olympus Mons]

@Bicicleur
This is interesting.


And "if" the samples are classified purely by age in each phase and area, it seems to be a tendency for Steppe ancestry to increase along Bell Beaker period, decrease along Bronze Age and increase again along the Iron Age.

Yes, really interesting if image shows it "over time" (oldest to newest) and have those two R1b L151 and P312 steppe warriors have so little steppe in them, does it not???

[Angela]

Probably these are the 3(three) SF-BA2, the 2(two) SF-IA1 and 1(one) SF-BA1 that are closer to Iberians:



Comparatively with samples of same age, a little more of WHG and a little bit less of yamna. Am I wrong?

I think that's right as far as Yamnaya is concerned, but as to WHG I'm not sure. They're consistent in having it, and some of the Bell Beaker, Bronze and Iron Age samples don't have any at all. Of the ones who do have it, maybe it's approximately the same if we averaged it out? Could the difference be in the amount of "farmer" as well?

I really can't tell. Maybe we need a ruler.:)

In the Tables it will tell us precisely, but unfortunately the rest of the samples aren't labeled,so I don't think I can say anything definitive.

From Table S 9

It runs Yamnaya/Anatolia/Villabruna

Pir3037AB

0.37

0.409

0.222

Quin58

0.273

0.449

0.278

Quin234

0.345

0.503

0.152

Pir6

0.366

0.445

0.19

Now for Iron Age:

PECH8

0.412

0.426

0.163

PT2

0.429

0.419

0.152

Btw, any idea which samples are closest to Italy?

[bigsnake49]

Further south and far to the east. Pontic Greek?

It's hard to distinguish between the Albanians and Mainland Greeks too. Couldn't they have used contrasting colors??? Maybe the Albanians are indeed mostly in one spot: just east of Tuscans, along with northern Greeks, but running up toward Bulgarians?

What do you think? I'm slightly color blind they tell me. I know, odd for a woman, but there you have it. My Dad was severely color blind, so I blame him. No wonder he liked sculpting more than painting. :)

Yeah, the colors are too close together between Greeks & Albanians particularly since they overlap so much. I did not think that Albanians would be so close to Bulgarians, I was kind of surprised by that. What I was not surprised by was the overlap and closeness between Bulgarians and Romanians. I guess the Ancient Thracians & Dacians were pretty close. As far as that one way out there sample, maybe Cappadocian Greeks. The ones that are separate from the main cluster probably Pontic or Thracian Greeks.