search this blog

Tuesday, March 13, 2018

First real foray into Migration Period Europe: the Gepid, Roman, Ostrogoth and others

This is going to be our first meaningful look at the all important Migration Period, thanks to the recently published Veeramah et al. 2018 paper and accompanying dataset (see here). The Migration Period is generally regarded to have been the time when present-day Europe first began to take shape, in a rather sudden and violent way, with, you guessed it, a lot of migrations taking place.

Here's where most of the ancients from Veeramah et al. 2018 cluster in my Principal Component Analysis (PCA) of ancient West Eurasian genetic variation. Those East Germanics (the Gepid and Ostrogoth) are certainly very eastern, and indeed more exotic than I would've ever expected them to be. But I do love surprises like this. The relevant datasheet is available here.

Obviously, as per the paper, the ACD in about half of the labels stands for Artificial Cranial Deformation. I've also updated my Global25 datasheets with many of the same ancients. You can use these datasheets to plot them on 2D or 3D "genetic maps", and model their ancestry proportions. Feel free to share your findings in the comments below.

Global25 datasheet ancient scaled

Global25 pop averages ancient scaled

Global25 datasheet ancient

Global25 pop averages ancient

Here are a few of my own models for some of the more interesting of these individuals, using nMonte3 and based mainly on Iron Age (IA) reference samples. I used the same data file for all of the models; it includes scaled coordinates and is available for download here.

[1] distance%=3.7819



[1] distance%=3.6339




[1] distance%=2.5535




[1] distance%=2.9444



The Gepid and Ostrogoth show significant Scythian- and Armenian-related ancestry proportions, respectively. Should that be taken literally? Or do we have to wait for, say, Avar and Hunnic genomes to expect more realistic models?

Update 15/03/2018: This is where many of the Medieval German samples cluster in my PCA of modern-day Northern European genetic variation (see here). Obviously, I could only run the individuals with wholly or overwhelmingly North European genomes, and most of these turned out to be the males without any signs of ACD. They look very West Germanic. The relevant datasheet is available here.

See also...

Modeling genetic ancestry with Davidski: step by step


«Oldest   ‹Older   201 – 222 of 222
Richard Rocca said...

@David said... Maju and his rather limited cheer squad are at it again.

Maju, Olympus Mons and Gioiello in one place = The land of misfit toys.

Unknown said...

Ted Kandell of Open Genomes analyzed the BAM files of the Moroccan Hunter Gatherers circa 15000 ybp. They are spot on M78 partially on to the breakup of M78 via YFull estimates. They are posted in Facebook group

Ric Hern said...

@ Davidski

Regarding Maju. I can not see any 2018 discussions, only 2016 early 2017 regarding DF 27. So maybe this discussion died and early death due to new Bell Beaker samples ?

Davidski said...

Well I just linked to the 2018 DF27 discussion.

Anonymous said...


On your list of what you consider improbable:

1) There actually are pieces of evidence of a pre-proto-Celto-Italic language. The linguists that proposed the North-West Block considered the substrate they found related to Italic.

2) You project modern day distributions on pre-Celtic populations.

3) We see a clear cultural connection to other IE cultures: Tumuli and cruched burials.

4) We see a likely language shift at the same time as we see the arrival of steppe ancestry

5) We *know* absolutely for sure that there were lots of non-IE populations during the ascent of Hittite because Hattic was used by Hittites as liturgical language. If we found that Hittites represented a total Yamnaya population replacement, THAT would be remarkable.

Simon_W said...

@Anthro Survey

Some critical remarks re: your Italian_Bergamo run:

I would have used French instead of French_East, because of some likely Germanic admixture in the latter. And instead of using just one single North Italian Beaker I would have used the average. Because the average is probably closer to the population average of North Italian Beakers, whilst one single individual is bound to be more random. I'm not quite sure what your custom Germanic set includes, but the average Germany_Medieval without ACD should be a fine proxy for the original Germanic part of the Longobard migrants. Then again the Amorim et al. paper showed that people with different ancestry migrated with the Germanic-like Longobards to northern Italy. And finally, what regards the usage of Minoans... They may be a good proxy for pre-Greek Aegeans indeed. But then again there is hardly any archaeological evidence for Minoan influence in Italy. The large majority of Bronze Age Aegean influence in Italy is Mycenaean. I used to be very impressed by the fact that nMonte run on Global 10 data suggested a Minoan-like component in my own Italian ancestry. But that's over, with Global 25 data I don't get anything Minoan anymore.

Simon_W said...


I noticed that in the updated Global 25 average spreadsheets the Beaker averages are now split into a Beaker average and a Beaker average without steppe. What's the former average exactly? The average of the total or just the average of the Beakers with steppe admixture?

Simon_W said...

@Anthro Survey

Re: The Gauls, I absolutely agree that they were more EEF shifted than the average central European Beaker, after all the modern French are more EEF shifted, and that's hardly the legacy of Roman input or constant gene-flow from Iberia over the last centuries...

I didn't quite keep in mind how far north the Occitan language originally extended. So Occitan in any case doesn't equal French_South, which makes your suggestion a bit less outlandish than I initially had believed. Yet I wouldn't equate the original area of the langue d'oil with the Germanic admixed part either. There were no Frankish settlers in the Poitou or in Berry for example. In contrast, the Burgundians settled mostly in the Franco-Provencal area and the Visigoths in the far southwest. And then all these Germanic settlers didn't manage to impose their language on the locals, simply because they were not numerous enough.

Indeed, I think we cannot exactly calculate the percentage of Germanic admixture in northern or eastern France, as long as we lack genomes from LBA, IA or Roman Age France. I customarily use the modern French sample as a proxy for the Gauls because there is nothing better at the moment, but I keep in mind that it's only a proxy.

French_East as a mix of Germany_Roman and Germany_Medieval may yield a close fit, but 54% Germanic admixture is outlandish, given the Romance language in eastern France. And so it's probably as wrong to assume the Gallo-Roman locals to have been like Maximus, the Germany_Roman.

Davidski said...


What's the former average exactly? The average of the total or just the average of the Beakers with steppe admixture?

You can see the values that I averaged by looking at the non-averaged datasheets.

Simon_W said...

Well of course they're averages of the non-averaged datasheets, that goes without saying. But I was wondering if the general Beaker averages include those without steppe or if you made some kind of apartheid there, with an average of the steppe-admixed ones, besides the average of those without steppe.

Alogo said...


Yeah, shame, your former scenario was more interesting but who knows - reality was likely a bit more complex than the temporally distant samples make it look like. :-)


Thanks for taking yet another close look.

Ideally, we of course want (as Chad mentioned too) to have the relevant, later general northern Balkan-Pannonian-Pontic area more sampled as well.

As for the Sarmatians, while quite similar to Scythian_Samara overall they seem to work somewhat better in some models for at least some of the samples (like STR_310) that don't have much if any extra East Asian. The other Scythians matter more obviously due to their eastern input and are potentially a stand-in for various other later steppic groups with various other languages (say, Huns) as you wrote.

On average, the deformed/intermediate group members predictably and certainly seem like complex, err 'non-standard' (by our current sampling standards, that is) mixes at any rate that need more sampling of the general area to be better clarified. This basic sort of model (something Scandinavian-like, something steppe-like, something Balkan or more geographically northern-like) does seem plausible though.


Here's one with England_Roman as a Celtic proxy. The initial model that you used:

[1] "distance%=2.5356"



with addition:

[1] "distance%=1.2355"



Alogo said...

Oops, correction on the first:

[1] "distance%=3.093"



André de Vasconcelos said...

England_Roman might not be an ideal proxy because he could have carried more steppe related ancestry than the hallstatt groups who made it into Iberia. Lusitanian appears to be have some Italic-ish features, so unless it's a proto-celtic language dating from the Beaker period, it shouldn't be surprising if it came from somewhere close to a contact area between Italic and Celtic - maybe the Alps - and I'd guess they'd have a higher EEF and lower steppe than Roman era Briton did.
Speculative, I admit, but that's my gut feeling

Anyway how would Spanish Basques fare with England_Roman?

Simon_W said...

@ Davidski

"You can see the values that I averaged by looking at the non-averaged datasheets."

Seriously, no shit? Lol. Alright, I'm going to calculate the average of the total myself and then compare it with your cryptic average. I can do that in no time.

Simon_W said...

Right, here's what I found: Beaker_Iberia average only includes steppe admixed Iberian beakers. I don't find that very useful, because steppe admixed Beakers and those without steppe were not distinct populations who didn't mix. But at least now I know it and can calculate my own averages.

Davidski said...


I don't find that very useful, because steppe admixed Beakers and those without steppe were not distinct populations who didn't mix.

Obviously, they were distinct populations until they started mixing, that's why some of them lack steppe admixture.

So the "no steppe" averages are very important, because they're informative about the types of long-standing local populations that the steppe-like Beakers mixed with in each part of Europe.

In fact, they're more important than the lumped averages, which are each just an arbitrary snapshot of a population in flux that will change as more samples come in.

Simon_W said...

OK, that's a valid point indeed. Still, to me the lumped averages also have some importance, because they're the best proxy available (though not perfect) for the post-admix Beaker population of a given region.

Alogo said...


That's most likely true, probably could get even better results with some other samples but just out of convenience. Re-running the Extremadura one along with the Basques as well since I finally started using David's most recent spreadsheets (one more Iberia_BA sample, with more steppe). Added some runs with Iberomaurusians too:

André de Vasconcelos said...


Thanks, the Spanish has higher Briton-related ancestry than Sp_Basques do, which is not unexpected (~22% vs 33%)

Arza said...

Vadim Verenich said...
I would assign his [KER_1] Y-haplogroup into Y J-L27 J-L26 J2a1- category

which is consistent with his F1-statistics:

Ker1.full-hg19 Ker1.full-hg19 1 0.6462 -3.882 French S_French-2
Ker1.full-hg19 Ker1.full-hg19 2 0.6456 -4.069 Armenian S_Armenian-1
Ker1.full-hg19 Ker1.full-hg19 3 0.6443 -4.27 Greek S_Greek-2
Ker1.full-hg19 Ker1.full-hg19 4 0.6442 -4.403 Sardinian B_Sardinian-3
Ker1.full-hg19 Ker1.full-hg19 5 0.6441 -4.441 Czech S_Czech-2
Ker1.full-hg19 Ker1.full-hg19 6 0.6441 -4.521 Jordanian S_Jordanian-3
Ker1.full-hg19 Ker1.full-hg19 7 0.6438 -4.552 Adygei S_Adygei-1
Ker1.full-hg19 Ker1.full-hg19 8 0.6436 -4.609 Tuscan S_Tuscan-2
Ker1.full-hg19 Ker1.full-hg19 9 0.6436 -4.653 Lezgin S_Lezgin-1
Ker1.full-hg19 Ker1.full-hg19 10 0.6432 -4.729 Armenian S_Armenian-2
Ker1.full-hg19 Ker1.full-hg19 11 0.6429 -4.772 Iraqi_Jew S_Iraqi_Jew-2
Ker1.full-hg19 Ker1.full-hg19 12 0.6429 -4.797 BedouinB S_BedouinB-1
Ker1.full-hg19 Ker1.full-hg19 13 0.6427 -4.826 Georgian S_Georgian-1
Ker1.full-hg19 Ker1.full-hg19 14 0.6427 -4.838 Samaritan S_Samaritan-1
Ker1.full-hg19 Ker1.full-hg19 15 0.6426 -4.861 Sardinian S_Sardinian-2
Ker1.full-hg19 Ker1.full-hg19 16 0.6425 -4.888 Hungarian S_Hungarian-2
Ker1.full-hg19 Ker1.full-hg19 17 0.6422 -4.91 English S_English-2
Ker1.full-hg19 Ker1.full-hg19 18 0.6422 -4.927 French S_French-1
Ker1.full-hg19 Ker1.full-hg19 19 0.6422 -4.955 Bergamo S_Bergamo-1
Ker1.full-hg19 Ker1.full-hg19 20 0.6416 -5.016 Basque S_Basque-1
Ker1.full-hg19 Ker1.full-hg19 21 0.6415 -5.446 Druze S_Druze-2
Ker1.full-hg19 Ker1.full-hg19 22 0.6415 -5.477 Greek S_Greek-1
Ker1.full-hg19 Ker1.full-hg19 23 0.6414 -5.569 Abkhasian S_Abkhasian-2
Ker1.full-hg19 Ker1.full-hg19 24 0.6413 -5.614 North_Ossetian S_North_Ossetian-2
Ker1.full-hg19 Ker1.full-hg19 25 0.6412 -5.662 Tuscan S_Tuscan-1
Ker1.full-hg19 Ker1.full-hg19 26 0.6412 -5.523 Orcadian S_Orcadian-2
Ker1.full-hg19 Ker1.full-hg19 27 0.6412 -5.514 Georgian S_Georgian-2
Ker1.full-hg19 Ker1.full-hg19 28 0.6411 -5.521 Orcadian S_Orcadian-1
Ker1.full-hg19 Ker1.full-hg19 29 0.6409 -5.534 Sardinian S_Sardinian-1
Ker1.full-hg19 Ker1.full-hg19 30 0.6409 -5.545 Albanian S_Albanian-1

This French at the top is especially interesting not only in the context of Gothic migration, but also because it's fairly easily detectable in G25.

Unknown said...
This comment has been removed by the author.
Unknown said...

Hmm tried to post a long list, but only a little bit of the comment showed up? Anyway long story short many users including myself, Alex Williamson, and Wayne K at anthrogenica have found 5 of the Lombards of Northern E. ancestry and 5 males from the Bavarian paper of Northern E. ancestry to be positive for the R1b-U106 SNP and some sub groups under that. Main difference is the Longobards were L48 or Z18+ while the Baiuvarii appear Z156 and Z305-307 (and even ALH 1 is DF96) according to sub group data!

«Oldest ‹Older   201 – 222 of 222   Newer› Newest»