search this blog

Thursday, February 2, 2017

First look at Polish Early Bronze Age genome PL_N17


I still don't know the archaeological context of this new sample from Gustorzyn, Northern Poland, but I've now managed to get my hands on his genome-wide data. The files I have look sound enough for a preliminary analysis, so this is how he compares to a variety of ancient and modern-day individuals and populations from around the world. Look for Poland_EBA PL_N17 in the various datasheets linked to below. Let's kick off with a few qpAdm models.

Outgroups
AG3-MA1
Chukchi
Dusun
Igorot
Karitiana
Kosipe
Kostenki14
Lebbo
Levant_Neolithic
Mbuti
Satsurblia
Ust_Ishim
Villabruna

PL_N17
Yamnaya_Samara 0.687±0.040
Lengyel_LN 0.249±0.037
Western_HG 0.064±0.028

chisq 3.434 tail_prob 0.969281

Corded_Ware_Germany
Yamnaya_Samara 0.688±0.021
Lengyel_LN 0.246±0.020
Western_HG 0.066±0.015

chisq 2.848 tail_prob 0.984786

Corded_Ware_Germany (2)
Yamnaya_Samara 0.701±0.045
Lengyel_LN 0.200±0.043
Western_HG 0.099±0.027

chisq 10.065 tail_prob 0.434827

Corded_Ware_Estonia
Yamnaya_Samara 0.649±0.047
Lengyel_LN 0.268±0.043
Western_HG 0.083±0.030

chisq 18.477 tail_prob 0.0474288

It's great to see that I get very similar results for PL_N17 by plugging his Global 10 coordinates into nMonte. The Global 10 plot with PL_N17 can be seen here; look for the black cross.

PL_N17
Yamnaya_Samara 61.9
Lengyel_LN:I1495 25.6
Western_HG 12.6

distance%=0.7639 / distance=0.007639

Obviously, PL_N17 packs a lot of Bronze Age steppe or Yamnaya-related ancestry. His overall genetic structure is similar to that of the Estonian and German Corded Ware individuals, suggesting that he's a direct offshoot of the steppe-derived circum-Baltic Corded Ware population.

Intriguingly, he shows inflated affinity to the Sintashta samples from the Trans-Ural steppe in the f3 outgroup shared drift stats. However, these Sintashta sequences appear to be affected by relatively strong post-mortem damage, so it's hard to say at this stage whether the result is meaningful. We really need more Sintashta genomes to be sure.


PL_N17 scores an impressive 26.36% Ancient North Eurasian ancestry in my Basal-rich K7 test. Hence, on the K7 pyramid plot he is sitting on the edge of the khaki colored Steppe_MLBA cluster, in fact right next to Sintashta and a few Srubnaya individuals.


And here's a plot based on raw SNP data instead of ancestry proportions. It shows results almost identical to the K7 pyramid plot.


Note that modern-day Poles are shifted west-northwest of PL_N17 on both of the West Eurasia-specific plots above. This perhaps suggests that at some point after 2000 BC there was an influx of Early European farmer (EEF) and Western Hunter-Gatherer (WHG) related gene flow into what is now Poland. But it's also possible that Corded Ware-derived groups lived alongside EEF-like farmers with inflated WHG ancestry in the South Baltic without mixing with them to any significant degree until the later stages of the Bronze Age, or even until the Iron Age.

However, these speculations are based on the assumption that modern-day Poles are direct descendants of PL_N17, which is probably true, but not guaranteed. So my next goal is to pinpoint the most closely and directly related modern-day ethnic groups to PL_N17. This is probably best done with some type of haplotype or rare alleles test (rather than, say, overall shared drift or ADMIXTURE output). Need to think about this a bit.

See also...

R1a-Z280 from Early Bronze Age Northern Poland

47 comments:

Nirjhar007 said...

Which language he likely spoke? :) . Nice corresponding analysis btw.


EastPole said...

@ Nirjhar007
I think some kind of early Slavic or Balto-Slavic.
From linguistic and genetic analysis most probable is that R1a-Z645 was an Indo-Balto-Slavic lineage. R1a-Z93 was an Indo-Iranian lineage and it’s brother R1a-Z283 was a Balto-Slavic lineage. R1a-Z280 then was also a Balto-Slavic lineage.

Nirjhar007 said...

So East pole , in your view, L-664 will be representing which group ? :) .

EastPole said...

@ Nirjhar007
I really don’t know. Probably some PIE dialect related to pre-Indo-Balto-Slavic. But no ethnicity which can be linked to it exists anymore. L-664 is a very rare lineage.

Tesmos said...

Is Lengyel_LN:I1495 the best Neolithic fit for that sample up to date?

Davidski said...

I don't know, I haven't tested that.

I used Lengyel_LN because it worked well for most Europeans when paired with Yamnaya, and it also works very well for PL_N17. But there might be better Neolithic/Copper Age samples for PL_N17?

Nirjhar007 said...

We need more research focusing on Slavs and that is certain , they are somewhat discriminated in European history , in most of the occasions intentionally IMO.
http://dispatchesfromturtleisland.blogspot.in/2017/02/slavic-origins.html

Eastpole,

While I appreciate your reasoning, which I consider has some basis, but I strictly forbid correlating SNPs with language , especially speaking of a time of which don't have any direct source . Another thing is that Balto-Slavs were attested quite late and perhaps if we stretch , we may take back their proto- material culture to 2nd millennium bc with some confidence yes.

Davidski said...

Yeah, we need early Slavic and all sorts of Iron Age genomes form Central and Eastern Europe to work this out.

But already I'm giggling at the suggestions I've seen that Slavs came from somewhere around Kiev, displaced people like PL_N17, and made Poland and the rest of Eastern Europe more western genetically than PL_N17.

Up is down, down is up.

Nirjhar007 said...

What we call 'Slavs' have their oldest identifiable materials from those cultures, like Kiev. It is suggestive, that the notion of Slav was a late development , with including other not so related tribes into the fold . The time before that, perhaps nothing like 'Slav' existed at all ! :) .

Simon said...

Apparently the sample is dated around 1900 BC which would point toward early Proto-Balto-Slavic or a contemporary parallel branch, perhaps also related to Germanic or the extinct NW Block language.

Simon_W said...

@Davidski
"But already I'm giggling at the suggestions I've seen that Slavs came from somewhere around Kiev, displaced people like PL_N17, and made Poland and the rest of Eastern Europe more western genetically than PL_N17"

Of course that's weird logic, but 2000 BC is still very early. If the Wielbark culture will turn out the way I've heard it will, then we'd still need Slavs coming in from the east to fresh up the Steppe ancestry. This sample doesn't tell us anything about the Lusatian culture either, which came later, and which covered most of modern Poland up to the Baltic Sea.

It's interesting how Poles are keen to see population continuity in their country, but this is a very relative term. Some measure of continuity can almost always be expected, but from which percentage on can we speak of a real continuity? 30%? 50%? 80%? And genetic continuity doesn't equal ethnic/linguistic continuity.

Shaikorth said...

@Simon

Like with Rathlin1 and BR2, we'll be able to connect it to specific modern European groups once we get a high coverage sample.


What we know: if Bronze Age Hungarians left direct ancestry, it's apparently more prevalent in Poland than anywhere else:
http://www.pnas.org/content/113/2/368/F3.large.jpg

BR2 has higher WHG than CW etc. and haplotype donations from BR2 correlate with Loschbour donations quite well in the Polish sample of Cassidy et al.

http://oi68.tinypic.com/33axzdu.jpg

On the SNP-based PCA Poles are between BR2 and Srubnaya/CW/PL_N17, and one of the latter should show similar elevated affinity to Poles, or Belarusians or Ukrainians if they are this kind of Bronze Age 2-way mix.

According to David someone suggested Slavs could have come from Kiev and been more western than the preceding Polish population. This could have happened if BR2-like populations extended deep into Ukraine and if CW-types survived in Poland until the Roman era (big "if's"). Then if high BR2 affinity in Poland comes from a pre-Slavic substrate we get a different story.

Lots of Late Bronze and Iron Age DNA needed to resolve this.

Slumbery said...

Davidski:
I can't see why it is so impossible or even upside down that there was a higher EEF + higher WHG population South-East of them at some point. I do not have a strong opinion on the question of Slavic source area, but there is definitely no reason to giggle.

epoch2013 said...

@Simon_W

"If the Wielbark culture will turn out the way I've heard it will, then we'd still need Slavs coming in from the east to fresh up the Steppe ancestry."

Can you elaborate on that? Are there Wielbark samples available or on the way? How, according to what you've heard, will they turn out?

rozenfag said...

Baltic and Ukrainian Mesolithic published: http://www.cell.com/current-biology/fulltext/S0960-9822(16)31542-1

Alberto said...

So according to qpAdm this sample is identical to Corded_Ware_Germany? While from the PCA data it does look more western:

Corded_Ware_Germany:I0103
Yamnaya_Samara:I0231 71.8 %
Hungary_N:I1495 21 %
Loschbour:Loschbour 7.2 %

Distance 0.001553

PL_N17
Yamnaya_Samara:I0231 62.2 %
Hungary_N:I1495 26.4 %
Loschbour:Loschbour 11.4 %

Distance 0.001925

Differences become more clear with Kotias-Karelia_HG instead of Yamnaya (I'm using weighted data, since without weighting it looks more dramatic, so probably less likely to be correct):

Corded_Ware_Germany:I0103
Kotias:KK1 33.8 %
Karelia_HG:I0061 29.4 %
Loschbour:Loschbour 18.4 %
Hungary_N:I1495 18.4 %

Distance 0.003166

PL_N17
Kotias:KK1 34 %
Loschbour:Loschbour 27 %
Hungary_N:I1495 19.8 %
Karelia_HG:I0061 19.2 %

Distance 0.003105

So basically CHG and Hungary_N are the same, but there is a shift from EHG to WHG. I wonder if this could be tested with a D-stat like:

Corded_Ware_Germany PL_N17 EHG WHG

To check what model is better supported.

Chad Rohlfsen said...

Latvian HGs are R1b... geez. Looking for Ukraine.

Chad Rohlfsen said...

R1b1b is all they have though.

Shaikorth said...

In fig S3 what's up with Ukraine_HG having higher affinity to Norwegians, Orcadians and Scots than to Belorussians, Russians, Ukrainians and Mordovians? Neolithic doesn't show this.

Arza said...

@David et al.

Can you take a look and say what you think about this:

Slavic homeland, origin of Finns, no extra WHG in Balts.
http://cordedware.blogspot.com/2017/02/slavic-homeland-origin-of-finns-no.html

I wonder if my simulated BalticHG will be close to actual Latvian HG. What a timing.

Azarov Dmitry said...

@Chad Rohlfsen
R1b1b is all they have though

Unfortunately they did not elaborate about methods they used to determine haplos. If they tested SNPs then phrase “the maximum-likelihood sub-haplogroup is R1b1b” sounds weird. They should know it for sure if they tested SNPs.

Matt said...

@ Davidski, is it possible to get this sample on the "Days of High Adventure" PCA (http://eurogenes.blogspot.co.uk/2016/11/days-of-high-adventure.html)?

That has some interesting extra dimensions that peak between Iberia_Chalcolithic and other samples and where recent Europeans and contribute to recent European differences. Calculating euclidean between population average distance on that PCA gives nice logical NW-East European splits, where the Globe10 just gives a single cline.

(Though I'm not totally sure about Bronze Age European and Steppe samples position on those dimensions).

Matt said...

Wow new paper. So Latvia HG: "In keeping with their geographical origins, they are in an intermediate position between Western European hunter-gatherer samples (WHG; from Luxembourg, Hungary, Italy, France, and Switzerland) and Eastern European hunter-gatherer samples (EHG; from Russia). They are composed of the same (blue) major component as these other hunter-gatherer groups in an ancestry coefficient decomposition analysis performed using ADMIXTURE (Figure 2B), suggesting a close relationship between these groups."

Sounds like SHG to me? That always seemed like the logical thing rather than SHG being somehow much more EHG like than Baltic HG (not sure why that was thought to be the case, IIRC and some thought that).

"We found that although the Latvian Mesolithic samples share closer affinity to WHG than to EHG, the Latvian Mesolithic samples do not belong entirely to either hunter-gatherer group (tested using D statistics [27], which offer a formal test of admixture; Table 2)."

Shouldn't they be testing D(Outgroup,LatvianHG,EHG,SHG) and D(Outgroup,LatvianHG,WHG,SHG) as well? Unless they have and I just haven't read that bit of their paper yet.

Plus, Latvia_HG3 has R1b1b then, and is no different in their ADMIXTURE to any of the other Latvia HGs.

Alberto said...

On the PCA, Latvia_HGs look in between WHG and SHG. And then Latvia_MN2 looks like an EHG (arrived around 4000 BC). The Ukrainian HGs look in between SHG and EHG, so they're probably the ones referred to in that Mathieson talk who mixed with Eastern Balkans farmers. I wonder how they will help to explain the genesis of Yamnaya. If they're a better fit that EHG, then you'd need a more eastern source from the south then the models we can test now.

Great to see new stuff finally!

Romulus said...

Told ya so

Matt said...

Alberto, right, looks like Baltic are more WHG than SHG on average, but not clearly sitting with WHG, while Ukraine are more where geography would've been expected to place them. (The ADMIXTURE in the supplement also matches with this).

Btw, re: selection, SLC24A5 and SLC45A2 mirror the pattern of more EHG like samples - the Ukraine_HG1 and Ukraine N1 - having both at least one copy of derived variants, while WHG like - Latvia_HG - tend to have ancestral variants, but more derived than WHG and also like WHG, to tend to have light eye variants (Ukraine both look dark eyed).

Hard to understand this long term pattern - it looks like the SLC24A5 and SLC45A2 had reasonably ancient antiquity among Euro HG, and there was opportunity to cross over into WHG, as there was some ancestry continuum (and crossed over between Anatolian HG and Euro HG in some direction). So poss. for whatever reason these would appear to have just made no selective headway among WHGs (Bichon, La Brana, Loschbour)...

Also if I'm reading Table S1 properly indicates a continuation of the pattern of WHG samples (Latvian HG) tending to have more of the derived variant in IRF4 (the variant associated with freckles, pale skin and early greying hair that reaches a world peak in present day Irish), being heterozygous for the derived, while the Ukraine HG is homozygous for the ancestral.

(Question: Given there's more of an apparent ancestry continuum, should we continue talking about WHG and EHG or talk about Western Euro HG and Eastern Euro HG, or some other language?)

Romulus said...

I don't think WHG SHG EHG are meaningful terms they just represent points on a spectrum between Pre-Ice age Indigenous European DNA and ANE. WHG is already obviously a mixture between the two, probably falling more towards ANE than Indigenous European.

rozenfag said...

Important note: The study of Mesolithic Baltic DNA that was talked about last autumn is not this one. This paper is done by group in Dublin, while "Mittnik et al., A genetic perspective on population dynamics of the pre-historic Eastern Baltic region" should be done by Johannes Krause's group.

Rob said...

We were right to keep bringing up odd SHG looking phantom populations
They'll probably be better fits for BA Hungary, and Balkans Greece too imo

Romulus said...

Latvian Neolithic is already Perfectly Yamnaya-like with 0 farmer ancestry prior to existence of Yamnaya. No need for Yamnaya anymore we can just throw that name in the trash. CWC = Funnel beaker + Latvian Neolithic. It's already well known HGs and farmers lived side by side for the duration of the Neolithic in the East LBK/FB zone. CWC = Farmers + Neighbouring HGs. Nobody came from the Steppe in the BA,no need for it, and Z2103 proves it.

Matt said...

LN1 does looks like an interesting sample in PCA and ADMIXTURE. ADMIXTURE looks like Steppe Early/Middle BA without any Anatolian ancestry, but its position looks to overlap on West-East with the Steppe Middle/Late BA who do have Anatolian ancestry and to be somewhat North of them. I can't quite tell from their PCA though. More samples will flesh out if this is an outlier.

It intersects with where a line between CHG to Latvia_HG/Ukraine_HG would be drawn, where some Yamnaya look instead to intersect CHG to EHG.

Though D( Mbuti Kotias (CHG) Latvia_HG Latvia_LN1) = 0.0182, Z: 2.138 seems like a relatively weak stat (and is the strongest suggestive of admixture in LN1 and the broader CHG-Iran N grouping).

Based on PCA, Latvian MN1 and MN2 are quite different. MN1 looks to be a light skinned individual (SLC24A5 derived, unknown genotype for SLC45A2), light eye (HERC2 derived) and predicted light hair and is pretty close to the Latvian HGs. MN2 essentially looks on PCA like an EHG, and it has very light skin (homozygous derived in both SLC24A5 and SLC45A2) with dark hair and dark eyes. What does the mtdna say about this pair?

Rob said...

Rom
By this LN sample is from 2000BC

Chad Rohlfsen said...

Lol, just wait until you guys see the Y-DNA from Ukraine Mesolithic. The Baltic hunters aren't very relevant.

Matt said...

I do think some kind of Corded Ware / Yamnaya type expansion still remains necessary to spread R1b across Europe.

Personally, while I'm not sure of the direct relevance of Baltic HGs to ancestry today, I do think this changes my priors on whether for R1b a sequence "ANE"->EHG->Corded/Yamnaya->Europe LNBA is necessarily more likely than Villabruna->Ukraine/Latvian HG->Corded/Yamnaya->Europe LNBA (if R1b is present in rather WHG like Latvian HGs). Time may tell.

Romulus said...

@Chad

Yeah R1b in Villabruna 12,000 B.C. isn't very relevant, all the R1b in Neolithic Spain isn't very relevant, Mesolithic R1b from the Baltic isn't relevant. If we're waiting for some confirmation of Steppe theory to turn up before it's relevant we might die waiting because classic interpretation gets more wrong by the day.

Rob said...

Let's take talk to new thread !

Arza said...

@Romulus
At least in this picture Yamna looks completely... unimportant?

https://1.bp.blogspot.com/-85ZEKR030kI/WJOi0exeWPI/AAAAAAAAAFw/I2QUCd32YJMMoCr4pYENt8vEPKE3_BYYwCLcB/s1600/latvian.png

PC1,PC2,PC6 3d plot from a spreadsheet that Matt posted here once (that one with "specific Balto-Slavic drift").

Davidski said...

@Alberto

Yeah, straight D-stats suggest that Corded Ware is somewhat more eastern and PL_N17 more western.

Mbuti Eastern_HG Poland_EBA Corded_Ware_Germany 0.013 3.088 813774
Corded_Ware_Germany Poland_EBA Eastern_HG Western_HG 0.0162 3.436 813591

I think some of the more western-shifted Corded Ware individuals might be screwing with the qpAdm model.

Seems like using all of the available Corded Ware samples depresses the Yamnaya-related ratio in the qpAdm model. After taking out the RISE individuals I get this.

Corded_Ware_Germany
Yamnaya_Samara 0.701±0.045
Lengyel_LN 0.200±0.043
Western_HG 0.099±0.027

chisq 10.065 tail_prob 0.434827

Arza said...

@Davidski

Can you look at this?

https://4.bp.blogspot.com/-mx0F2mHFzV4/WJPBK0ISmQI/AAAAAAAAAGA/kPT27GXVmF4POzTgTis2sLI0oa13TURkwCLcB/s1600/ukrainian.png

And then:

http://cordedware.blogspot.com/2017/02/slavic-homeland-origin-of-finns-no.html

There is a Balto-Slavic cline from Unetice to this new Ukrainian samples.

Davidski said...

@Arza

Not really sure what I'm looking at there. I need to run the samples myself. But I seriously doubt that Baltic HGs explain the ancestry of PL_N17, which is clearly Yamnaya-related.

@Matt

Days of high adventure coordinates.

Poland_EBA:PL_N17 -0.039 0.0156 0.0025 -0.009 0.0062 -0.0169 -0.0037 -0.0014 -0.0008 0.0003

Matt said...

@ Davidski, thanks. Sticking those into the spreadsheet I have with population averages on that PCA, euclidean distance for PCA Days of High Adventure calculates top 10 closest populations for Poland_EBA:PL_N17 as: Russian_Central, Ukrainian_North, Polish, Ukrainian_Poltava, Ukrainian_East, Slovakian, Russian, Ukrainian_West, Belarusian, Corded_Ware_Germany. Does seems regionally specific in connection.

(Compare to:

Bell Beaker Germany population average top 10: Irish, Dutch, English_Kent, Scottish, English_Cornwall, French, Orcadian, Alberstedt_LN, German, Unetice_EBA

Nordic_LBA population average top 10: Orcadian, Norwegian, Unetice_EBA, Alberstedt_LN, English_Cornwall, Irish, Scottish, Dutch, French, Bell_Beaker_Germany

Alberstedt_LN: Irish, Scottish, Orcadian, Bell_Beaker_Germany, Dutch, English_Cornwall, English_Kent, Unetice_EBA, German, Austria

Unetice_EBA: Nordic_LN, Norwegian, Irish, Bell_Beaker_Germany, Swedish, Alberstedt_LN, Orcadian, Scottish, German, Austria

Hungary_BA: Czech, Nordic_BA, German, Slovenian, Hungarian, Swedish, Dutch, Norwegian, Croatian, Slovakian

Though admittedly this is maybe a bit apples to oranges, as these are averages for heterogenous LNBA populations and PL_N17 is a single sample).

Matt said...

On the neighbour joining tree structure based on distances in the Days of High Adventure PCA, for population averages, sits on a branch between the North Slavic subbranch and the steppe / Euro HG subbranch: http://i.imgur.com/6RCbfu7.png

Arza said...

"BalticHG" here is a ghost population made by me, including the name.

I was trying to explain what is pulling Balto-Slavs out of Unetice. There is a clearly visible (in 3D) cline from Czechs/Unetice/Germanics to Latvians.

Apparently new Ukrainian samples are on the same cline, just further away from Latvians than my ghost, that was based on BB Czech 568 and different HG.

But I seriously doubt that Baltic HGs explain the ancestry of PL_N17, which is clearly Yamnaya-related.

That's the point! As you wrote - Poles are more western than Kuyavian guy, and they are more western because Proto-Balto-Slavs migrated from the West (Unetice-like) to the East (Ukrainian HG or N), not the other way (and they did it after the time when our Kuyaviak lived, but apparently before the arrival of Uralics - BTW when they appeared around Baltic?).

Even the Finns and Karelians are as western as Germans in terms of Yamna/Farmer ratio if you subtract Saami and this Balto-Ukraino-whathever_HG ancestry.

Here is a 2-way mix. Maybe fits are not so impressive because there are single points at the ends and a cloud of samples between them, but the behaviour is consistent.

CentralWestEurasia3
Balto,-0.0606,0.0224,0.0232,-0.0066,0.016,-0.0384,-0.0135,-0.0049,-0.0088

Population,Unetice_EBA:I0047,Balto,X,X,Dstatistic
Latvian:GS000035027 18 82 _ _ 0.012
Latvian:latvian54A2 19 81 _ _ 0.0066
Latvian:latvian22J5 23 77 _ _ 0.0089
Latvian:latvian54H7 24 76 _ _ 0.0105
Latvian:GS000016903 31 69 _ _ 0.0109
Lithuanian:lithuania7 32 68 _ _ 0.0127
Lithuanian:GS000016905 33 67 _ _ 0.013
Latvian:GS000035148 35 65 _ _ 0.0046
Lithuanian:lithuania5 35 65 _ _ 0.01
Lithuanian:GS000016904 36 64 _ _ 0.0111
Latvian:latvian54F2 38 62 _ _ 0.0068
Latvian:latvian58C6 38 62 _ _ 0.008
Latvian:latvian58C8 44 56 _ _ 0.0104
Lithuanian:lithuania4 45 55 _ _ 0.0148
Belarusian:GS000035241 50 50 _ _ 0.0114
Belarusian:belorus9 52 48 _ _ 0.0111
Lithuanian:lithuania3 53 47 _ _ 0.0111
Belarusian:GS000014351 53 47 _ _ 0.009
Polish:GS000016889 55 45 _ _ 0.0119
Belarusian:bela29zp 57 43 _ _ 0.0188
Belarusian:bela32zp 58 42 _ _ 0.01
Belarusian:bela54zp 58 42 _ _ 0.0099
Lithuanian:lithuania6 59 41 _ _ 0.0152
Polish:Polish9H 59 41 _ _ 0.0127
Polish:GS000016890 60 40 _ _ 0.0136
Belarusian:GS000014324 61 39 _ _ 0.0079
Belarusian:bela52vp 61 39 _ _ 0.0129
Belarusian:bela45vp 62 38 _ _ 0.0156
Polish:Polish14H 62 38 _ _ 0.0106
Polish:Polish15H 62 38 _ _ 0.016
Polish:Polish13H 63 37 _ _ 0.0152
Polish:Polish16H 63 37 _ _ 0.0139
Polish:Polish7H 63 37 _ _ 0.0138
Polish:Polish10H 65 35 _ _ 0.0094
Polish:Polish8H 66 34 _ _ 0.0095
Belarusian:GS000016104 67 33 _ _ 0.0116
Belarusian:belorus7 67 33 _ _ 0.0065
Lithuanian:GS000035040 68 32 _ _ 0.0126
Belarusian:bela50vp 70 30 _ _ 0.0135
Polish:Polish11H 72 28 _ _ 0.0124
Polish:GS000015869 73 27 _ _ 0.0095
Polish:Polish6H 77 23 _ _ 0.0107
Polish:Polish12H 88 12 _ _ 0.0104

"New Ukrainians" are about two times further from Unetice than this "Balto", so proportions will change to more Unetice-like.

Rob said...

@ Arza

How can we be sure that the western shift is directly due to something Unetice like. It could be a cumulative sum of other processes. Broadly, something Balkan Bronze Age impacting a still very steppic population, which then encountered remnant Germani in west Slavic areas (Poland and Czeczhia).

Alberto said...

@Davidski

Thanks for testing that. And I'm glad to see that the Global 10 PCA is working so good. (I'm rather inclined now to think that the weighted dimensions according to their eigenvalue tend to work better, but that's a complicated story and at least i can test both and check which one makes sense for each model).

With D-stats based sheet I see more noise and variability depending on the outgroups. In the tests about Onge-like admixture, I found that even including populations like Yamnaya, BB, Armenia_EBA, Iran Chalcolithic, CHG... I still get a good amount of Pathan/Kalash admixture throughout Europe (Kalash peaks in West Ukrainian at 17%!). I'm a bit sceptic about that (not seen in the PCA at all). Though I started testing this also because of the f3 admixture stats in the latest Laz. 2016 paper, where even with Yamnaya samples available, some European populations (Albanians, Greek, Sicilian,...) were getting the lowest score with something like Europe_EN + Pathan. So maybe there is something real to it. Time will tell.

G. Dekaen said...

@Arza

For what it's worth, I think you're correct. Funny, I posted something a few days ago postulating the exact same thing, a probable large contribution of Unetice/Tumulus/Urnfield to Balto-Slavs.

There's a remarkable correspondence of admix percentages between BB/Unetice and East/West Slavic populations, far closer than with Corded Ware who may have been successively pushed out by C. European cultures like BB/Unetice or Urnfield and back onto the Steppe.

It would also explain why all non-Southern Slavs (with the exception of Poles) consistently carry Y-DNA I2a1 at approximately 15-18% in core Slavic areas (Ukraine, S. Russia etc.) with frequencies dropping off only in lands historically inhabited by non-I2a1 rich peoples like Balts and Finns (NE Belarus, N. Russia).

Alternatively, if proto-Slavs were more like Poles with minor I2a1 <10%, then maybe the I2a1 so frequent among Ukrainians and S. Russians could be from Thracian admixture as Sorin Paliga speculates on linguistic grounds? That would be a HUGE contribution from a rather unknown population though!

All in all, my guess would be that this Polish sample does not represent a Balto-Slavic speaker.

Tesmos said...

@David

''I used Lengyel_LN because it worked well for most Europeans when paired with Yamnaya, and it also works very well for PL_N17. But there might be better Neolithic/Copper Age samples for PL_N17?''

Hmm... Nmonte2 seems to pick LBK over Lengyel_LN for PL_N17.

PL_N17

Yamnaya_Samara:I0429 62.70
LBK_EN:I0048 19.05
Loschbour:Loschbour 11.00
Hungary_N:I1495 7.25