The Uralic cline with kra001 - no projection this time

A whole lot of nonsense was posted online, often by people who should've known better, after I claimed that kra001 was a solid proxy for a proto-Uralic genome (see here).

For those of you who still don't get it, below are three Principal Component Analysis (PCA) plots featuring Uralic speakers and other present-day Eurasians. Kra001 is also there. These graphs are based on genotype data not reprocessed Global25 data. The relevant datasheet is available here.

Compared to my previous PCA with kra001, here I included a bigger range of East Eurasian populations to help mitigate the effects of extreme genetic drift in some of the Siberian groups, at least on the first few Principal Components (PCs). Moreover, kra001 wasn't projected onto PCs computed with modern-day samples, so he was free to influence the outcome of the PCA.

Note the east to west clines made up largely of Uralic speaking groups on the first two plots. These plots are based on PCs 1/2 and 1 /3, respectively. The third plot, based on PCs 1/4, is more complex and thus more difficult to interpret, but it also manages to isolate many of the Uralic populations from the others.

The Uralic-specific clines do intersect with the clines and clusters formed by the other linguistic groups. However, based on the three plots, the Yeniseian-speaking Kets are the only Asian group that can plausibly be confused for Uralic speakers.

Importantly, apart from the Kets, kra001 is the only Asian individual who shifts his position on all three plots as if he were a Uralic speaker. This might well be a coincidence, and we'll never know what language was spoken by kra001, but it does suggest to me that his genome is a solid proxy for a proto-Uralic genome.

See also...

First taste of Early Medieval DNA from the Ural region

The BOO people: earliest Uralic speakers in the ancient DNA record?

Fresh off the sledge

Finally, a proto-Uralic genome

Obviously, genes don't speak languages, people do. But sometimes it's possible to associate a linguistic group with a very specific genetic signature.

A while ago many of us in the blogosphere spotted an uncanny connection between the Uralic language family, Y-haplogroup N-L1026 and Nganasan-like genome-wide genetic ancestry.

As a result, we expected a Nganasan-like population rich in N-L1026 to eventually appear in the ancient DNA record, probably somewhere in Siberia and in burials from a likely proto-Uralic archeological culture. This hasn't happened yet, but we now have direct evidence that such a population must have existed somewhere deep in Siberia as early as the Bronze Age.

Kra001, whose genome was published recently along with Kilinc et al., belongs to a pre-N-L1026 lineage and, at least in terms of genome-wide genetic structure, could well be from a population directly ancestral to present-day Nganasans. Of course, the Nganasan language is part of the Samoyedic branch of Uralic.

Below is a series of Principal Component Analyses (PCA) featuring kra001. He's labeled RUS_Krasnoyarsk_BA, after the location and age of his burial. Note the obvious Uralic cline running across the plots. That is, from west to east. Kra001 is positioned at the end of this cline very close to a small cluster of Nganasans. To see interactive versions of the plots, paste the Global25 coordinates here into the relevant field here.

Admittedly, there's no way of knowing whether this individual spoke proto-Uralic or not. Indeed, he may have spoken something totally unrelated. The important point is that the very specific genetic signature shared by almost all present-day Uralic speakers, except perhaps Hungarians, is now finally represented in the ancient DNA record. And I can reveal to you that we'll soon be seeing many more ancients very similar to kra001 in upcoming papers.

See also...

The Uralic cline with kra001 - no projection this time

Fresh off the sledge

Fresh off the sledge

As things stand, the closest individual to a Proto-Uralic speaker in the ancient DNA record is arguably 0LS10 from an Iron Age tarand grave in what is now Estonia. I say that because:

- isotopic data suggest that 0LS10 wasn't born where he died, and considering his elevated Siberian ancestry relative to earlier and most contemporaneous Baltic ancients, he was very likely a migrant to the Baltic region from the east

- the tarand grave tradition appears to be specifically a Finnic (west Uralic) phenomenon that probably spread from the Volga-Oka region, which is just west of where most people place the Proto-Uralic homeland

- 0LS10 belongs to Y-chromosome haplogroup N-L1026, a paternal marker that is especially closely associated with Uralic-speaking populations and probably only appeared in the East Baltic region during the transition from the Bronze Age to the Iron Age

You can find more background info about 0LS10 and other relevant samples in Saag et al. 2019 (see here). This is where he sits in my Principal Component Analyses (PCA) focusing on fine scale Northern European genetic diversity. The relevant datasheets are available here and here, respectively.

Note that 0LS10 doesn't cluster strongly with any ancient or modern populations. To investigate this in more detail I ran a series of two-way qpAdm analyses, testing tens of ancient individuals and populations as potential admixture sources. These two models stood out above the rest in terms of their statistical fits, chronology and overall plausibility.

Baltic_EST_IA_0LS10
Baltic_EST_BA 0.826±0.045
RUS_Sintashta_MLBA_o1 0.174±0.045
chisq 12.527
tail prob 0.564048
Full output

Baltic_EST_IA_0LS10
Baltic_EST_BA 0.683±0.102
RUS_Mezhovskaya 0.317±0.102
chisq 13.811
tail prob 0.463864
Full output

Please note that RUS_Sintashta_MLBA_o1 isn't representative of the Sintashta culture population as a whole. It's a group of the most extreme genetic outliers among the Sintashta samples, and they may or may not have been Uralic speakers (see here). Interestingly, the Mezhovskaya culture population is generally associated with the Ugric branch of the Uralic language family.

I was also able to closely replicate these results with the Global25/nMonte method; down to almost one per cent. However, the statistical fits (distances) are poor, probably because the reference populations aren't the real mixture sources. This is in line with the fact that their Y-haplogroups are Q1a, R1a and R1b, rather than any type of N.

Baltic_EST_IA:0LS10
Baltic_EST_BA,83.8
RUS_Sintashta_MLBA_o1,16.2
distance%=4.7955

Baltic_EST_IA:0LS10
Baltic_EST_BA,69.8
RUS_Mezhovskaya,30.2
distance%=3.5783

I do realize that two Bronze Age samples from Bolshoy Oleni Ostrov, Kola Peninsula, belong to N-L1026, but adding them to my mixture models doesn't help. Little wonder, because the Kola Peninsula lies within the Arctic Circle, and I'm pretty sure that 0LS10 and his N-L1026 came from somewhere just north of the mixture cline marked on the map below. Unfortunately, I can't test this directly yet due to the scarcity of ancient samples from this region.

See also...

It was always going to be this way

On the association between Uralic expansions and Y-haplogroup N

Uralic-specific genome-wide ancestry did make a signifcant impact in the East Baltic

Uralic-specific genome-wide ancestry did make a signifcant impact in the East Baltic

I've started analyzing the ancient genotype data from the recent Saag et al. paper on the expansion of Uralic languages and associated spread of Siberian ancestry into the East Baltic region. The paper is freely available here and the data are here.

I really like the paper, but I don't agree with the authors' claim that the appearance of Y-chromosome haplogroup N in what is now Estonia and surrounds during the Iron Age is "not matched by a clear shift in autosomal profiles". In my opinion it certainly is, and, as one would expect, it's a shift towards a genetic profile typical of western Uralic speakers.

I'd say that the easiest way to find this signal is with a Principal Component Analysis (PCA) focusing on fine scale genetic substructures within Northern Europe, like the one below. The relevant datasheet is available here.

Note that the East Baltic Iron Age samples, all from burial sites in what is now Estonia, appear to be peeling away from their Bronze Age predecessors and overlapping strongly with present-day Estonians, who are Uralic speakers. Indeed, the PCA suggests to me that the formation of the greater part of the present-day Estonian gene pool took place in the East Baltic during the transition from the Bronze Age to the Iron Age. That is, when Uralic languages are generally accepted to have arrived in the region from near the Ural Mountains in the east.

I was also able to closely replicate these outcomes with my Global25 data using the method described here. However, in this effort, present-day Estonians are clearly more western than the Estonian Iron Age samples (EST_IA), which might be due to the presence of low level Germanic ancestry in Estonia dating to the medieval period. The relevant datasheet is available here.

Interestingly, the Estonian Bronze Age samples (EST_BA) come from stone-cist graves which are widely hypothesized to have been introduced to the East Baltic from the Nordic Bronze Age civilization. I even recall reading a paper on the topic which claimed that the remains buried in such graves were those of Proto-Germanic-speaking Scandinavian migrants. Well, I haven't had a chance to study these samples in any great detail yet, but considering that in both of the PCA above they're overlapping strongly with Latvian Bronze Age samples (LVA_BA) and sitting far away from the nearest Scandinavians, I'd say they're probably of local stock from way back.

See also...

It was always going to be this way

On the association between Uralic expansions and Y-haplogroup N

Inferring the linguistic affinity of long dead and non-literate peoples: a multidisciplinary approach

On the trail of the Proto-Uralic speakers (work in progress)

Historical linguists have long posited that Fennoscandia was a busy contact zone between early Germanic and Uralic languages. The first ancient DNA samples from what is now Finland have corroborated their inferences, by showing that during the Iron Age the western part of the country was inhabited by a genetically heterogeneous population closely related to both the Uralic-speaking Saami and Germanic-speaking southern Scandinavians.

The samples were sequenced and analyzed by two different teams of researches, and their findings published recently in Lamnidis et al. and Sikora et al. (see here and here, respectively).

This is how most of these ancients, whose remains were excavated from the Levanluhta burial site dated to 300–800 CE, behave in a Principal Component Analysis (PCA) based on my Global25 data. Levanluhta_IA are the Saami-related samples, while Levanluhta_IA_o is an Scandinavian-like outlier. Baltic_IA is an Iron Age individual from what is now Lithuania from the recent Damgaard et al. paper (see here). Note the accuracy of the Global25 data in pinpointing their genetic affinities and also the trajectory of the Levanluhta_IA cluster, which seems to be "pulling" towards Levanluhta_IA_o.

The Saami and Levanluhta_IA are clear outliers from the main Northern European cluster. There are two reasons for this: excess East Asian/Siberian-related ancestry and Saami-specific genetic drift. However, this eastern admixture and genetic drift are shared in varying degrees by other North European populations, especially those that also speak Uralic languages, and this is why they appear to be "pulling" towards the Saami/Levanluhta_IA clusters in my PCA. Thus, what this suggests is that the expansion of Uralic languages across Northeastern Europe was intimately linked with the spread of Siberian-related ancestry into the region.

This idea has been around for a long time and is now becoming even more widely accepted (see here). However, Lamnidis et al. also featured samples from a likely pre-Uralic (1523±87 calBCE) burial site at Bolshoy Oleni Ostrov in the Kola Peninsula, present-day northern Russia, and, perhaps surprisingly, found that they showed even more Siberian-related ancestry than Levanluhta_IA. So what's going on?

I'm confident that this discrepancy can be explained by multiple waves of migrations from the east into Northeastern Europe, possibly before, during and after the time of the people buried at Bolshoy Oleni Ostrov, by pre-Uralic, para-Uralic and/or Proto-Uralic-speaking populations.

Consider the following qpAdm output, in which Levanluhta_IA is just barely modeled successfully as a two-way mixture between Levanluhta_IA_o and Bolshoy_Oleni_Ostrov. The statistical fit improves significantly with the addition of Glazkovo_EBA as a third mixture source. This is an ancient population from near Lake Baikal dated to 4597-3726 BC from the aforementioned Damgaard et al. paper.

Levanluhta_IA
Bolshoy_Oleni_Ostrov 0.468±0.036
Levanluhta_IA_o 0.532±0.036
chisq 19.129
tail prob 0.0854706
Full output

Levanluhta_IA
Bolshoy_Oleni_Ostrov 0.241±0.092
Glazkovo_EBA 0.162±0.059
Levanluhta_IA_o 0.597±0.046
chisq 7.756
tail prob 0.734966
Full output

For the sake of being complete, I also tested whether Levanluhta_IA_o could be substituted by other similar ancient samples from the neighborhood, including those associated with the Battle-Axe and Corded Ware cultures. There's not much to report; qpAdm returned poor statistical fits and/or implausible ancestry proportions (for the full output from my runs, see here). Baltic_IA did produce a statistically sound model, but with excess Glazkovo_EBA-related ancestry. I also had to drop Bolshoy_Oleni_Ostrov from the analysis to make things work, which suggests to me that the result shouldn't be taken too literally.

Levanluhta_IA
Baltic_IA 0.677±0.034
Glazkovo_EBA 0.323±0.034
chisq 8.547
tail prob 0.741095
Full output

So as far as I can see, the western ancestry in Levanluhta_IA is likely to be mostly of Germanic origin, and thus Indo-European, meaning that it's logical to look east, perhaps far to the east, for the source of its Uralic ancestry. This might seem like a complicated and uncertain task, considering that Levanluhta_IA could well be at least a thousand years younger than the first entry of Uralic speakers into Fennoscandia. However, take a look what happens when I substitute Glazkovo_EBA with a variety of Uralic-speaking populations from around the Ural Mountains, which is where the Proto-Uralic homeland is generally considered to have been located.

Levanluhta_IA
Bolshoy_Oleni_Ostrov 0.210±0.091
Khanty 0.283±0.090
Levanluhta_IA_o 0.507±0.035
chisq 7.007
tail prob 0.798532
Full output

Levanluhta_IA
Bolshoy_Oleni_Ostrov 0.193±0.098
Levanluhta_IA_o 0.495±0.035
Mansi 0.312±0.100
chisq 7.884
tail prob 0.7237
Full output

Levanluhta_IA
Bolshoy_Oleni_Ostrov 0.300±0.065
Levanluhta_IA_o 0.337±0.072
Mari 0.363±0.121
chisq 8.393
tail prob 0.677705
Full output

Levanluhta_IA
Bolshoy_Oleni_Ostrov 0.238±0.084
Levanluhta_IA_o 0.553±0.036
Nenets 0.209±0.067
chisq 7.210
tail prob 0.78181
Full output

Levanluhta_IA
Bolshoy_Oleni_Ostrov 0.302±0.069
Levanluhta_IA_o 0.324±0.081
Udmurt 0.373±0.135
chisq 9.195
tail prob 0.60393
Full output

All of these models look great, and easily rival the best model with Glazkovo_EBA. Moreover, they make good sense in terms of linguistics. The only problem is that they're anachronistic, because the Uralic-speaking reference populations are younger than Levanluhta_IA. So I can't be certain that they reflect reality without corroboration from ancient DNA. It might turn out, for instance, that a Glazkovo_EBA-like population was already present somewhere deep in Europe before or during the time of Bolshoy_Oleni_Ostrov, while no such population existed around the Ural Mountains until the time of Levanluhta_IA.

By the way, it might be important to note that the present-day Finnish samples in my dataset can't be modeled as a mixture between Levanluhta_IA and Levanluhta_IA_o. But they can be modeled as a mixture between Baltic_IA and Levanluhta_IA. I don't know which part of Finland they're from exactly; probably all over the place, so it'd be useful to test regional Finnish populations to see how they behave in such models. Of course, Finns aren't Saamic speakers, they're Finnic speakers, and they're probably the result of a more recent Uralic expansion into Fennoscandia than the one that gave rise to the Saami.

Finnish
Baltic_IA 0.671±0.076
Levanluhta_IA 0.329±0.076
chisq 14.114
tail prob 0.293508
Full output

Damgaard et al. didn't report the Y-haplogroup for Baltic_IA, but the word round the campfire is that this individual belonged to N1c, which is today the most common Y-haplogroup among Uralic speakers. Obviously, we need a lot more ancient DNA to sort all of this out, but things are already looking pretty much as expected. Stay tuned for new posts in this series following the publication of more ancient DNA relevant to this fascinating topic.

See also...

How did Y-haplogroup N1c get to Bolshoy Oleni Ostrov?

The Uralic cline in the Global25

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...