Modeling Polish ancestry

[ambron]

Poles are basically the Trzciniec population (80-90%), slightly shifted towards the Balkans, whit a small Sacandinavian admixture. So, the most realistic models of the Polish population can be obtained by juxtaposing:
Trzciniec samples
Wielbark samples
indigenous Balkan samples from the Iron Age or early Middle Ages.

[Riverman]

Riverman

U106 is the best example of the autochthonous character of most of the Polish R1b. Many Polish haplotypes belong to the oldest layer U106, dating to 2000-4000 ybp:

R-Z345 - 3900 ybp
R-Y264455 - 3800 ybp
R-Y291823 - 3600 ybp
R-Y165307 - 3500 ybp
R-Y45730 - 3400 ybp
R-FT147099 - 3400 ybp
R-Y413603 - 3300 ybp
R-Y245050 - 3300 ybp
R-FGC57405 - 3100 ybp
R-ZP159 - 3100 ybp
R-BY18855 - 3100 ybp
R-Y349706 - 3100 ybp
R-Z8185 - 3000 ybp
R-FGC11784 - 2900 ybp
R-BY41723 - 2800 ybp
R-Y413876 - 2700 ybp
R-FT46604 - 2700 ybp
R-A14205 - 2700 ybp
R-FGC52137 - 2700 ybp
R-Y15791 - 2600 ybp
R-S15823 - 2500 ybp
R-S12035 - 2500 ybp
R-BY105611 - 2400 ybp
R-TY22801 - 2300 ybp
R-Y55050 - 2200 ybp
R-Y61294 - 2100 ybp
R-Y132003 - 2000 ybp

You wrote about Eneolithic-EBA cultures like Corded Ware and Bell Beakers. We can discuss whether these branches of R-U106 ended up in Poland in the Germanic Antiquity or with later Germans, but you can't declare them autochthonous from the Bronze Age, that's not possible, just look at the branch members. They split from Western, Germanic branches usually between about 100 AD-1500 AD.

The question is therefore did the early Slavs soak up so many Germanics in Poland, much more than elsewhere, and how much of this Germanic contribution came from later German settlers. That's the real question, Germanic substrate vs. Medieval German settlers.

If you use YFull for comparisons, it won't suffice, because most of the Germans and German Americans tested on FTDNA. On YFull is just a fraction of German testers for most branches.

[Mythbuster General]

From what modern Polish results I've seen throughout the years, what became clear to me is how much ancient Germanic ancestry they lack compared to other East-Central Europeans, looking more similar to East Slavs.

[leonardo]

Riverman

Many Polish haplotypes belong to the oldest layer U106, dating to 2000-4000 ybp:

Remnants of the Celts who once worked the Amber Road?

[Tomenable]

From what modern Polish results I've seen throughout the years, what became clear to me is how much ancient Germanic ancestry they lack compared to other East-Central Europeans, looking more similar to East Slavs.

I guess you are talking about 23andMe results, where Poles are among references for the "Eastern European" component, which is why they score almost 100% of it. In other companies it looks differently. For example this is how DNA Tribes company modelled my ancestry:



For comparison my DNA Land results:



If you use Belarusians as reference for Slavic, then Poles often score substantial amounts of Germanic ancestry.

But this doesn't apply to North-Eastern Poles from the borderland with Belarus who are often very Belarusian-like.

[Riverman]

Obviously Czechs have way more Germanic and Celtic ancestry, but even among Czechs there can be huge differences, depending on the region in question. Typically the regions with known stronger German settlement, where many German communities became Czech over time, are way more German shifted than others. In the end, there is a strong overlap of Czechs and Poles, because the least German influenced Czechs are less German shifted than the more German shifted Poles. That's something which can be shown no every PCA too.

About one third of the Polish G25 sample is in the clearly German influenced Czech range. Pretty obvious case. Everything "right of the Slovaks" is less German influenced, the clearest borderline being marked by Kashubians and Sorbs - it is in itself quite remarkable, that Sorbs are on average less German shifted than about 1/3 of the Polish sample:



That's not to say that Kashubians and Sorbs have no German admixture at all, even on the contrary if looking at the next table, but clearly less than Czechs, Slovaks and one third of the Poles.

Many Poles are as Slavic as Eastern Ukrainians, there is really no generalisation possible:

 


A lot of Poles score more than 90 % Dnipro Ukrainian. But many others score much less. And there are clear geographical patterns, like the two Silesians with high German vs. the Kashubians with medium to low German. I'm sure if doing the same with various regional samples, there would be clear trends in German admixture.

I used a very simplified model with just Ukrainian Dnipro, German, Albanian and Ashkenazi. Of course there are better ones around, but all will show a significant German-like input, be it from ancient Celts and Germanics, or Medieval and modern German settlers, or most likely both.

[ambron]

Riverman

I don't want to check all these lines, but for example the Polish R-FT9791 separates from the Germanic R-Z2 in 1900 BC.

The oldest sample U106 belongs to the CWC and comes from the Czech-Polish border from around 2900 BC.

The first Slavs (Proto-Slavs) were in Poland in the Trzciniec culture, so they could not absorb the Germanics, because they did not exist at that time.

[Riverman]

Riverman

I don't want to check all these lines, but for example the Polish R-FT9791 separates from the Germanic R-Z2 in 1900 BC.

The oldest sample U106 belongs to the CWC and comes from the Czech-Polish border from around 2900 BC.

The first Slavs (Proto-Slavs) were in Poland in the Trzciniec culture, so they could not absorb the Germanics, because they did not exist at that time.

Since old TMRCA don't change the overall pattern. It might be just testing bias. I can't exclude single lineages being older in Baltoslavs or outside of the Germanic ethnos, especially in Celts, but most of it seems to be Germanic and has a more recent (post first century) TMRCA with Germanic people.

[ambron]

Riverman

The east-west gradient already existed in the Polish Trzciniec culture, so it is not the result of medieval German settlement, but of a different local demographic background, assimilated by the CWC.

[Riverman]

Riverman

The east-west gradient already existed in the Polish Trzciniec culture, so it is not the result of medieval German settlement, but of a different local demographic background, assimilated by the CWC.

Most of the local admixture is present in all Slavs, including those Ukrainians from Dnipro. The German shift in Poles is very clearly more Germanic/German specific overall. But of course, its hard to distinguish more ancient Germanic, Celtic and Lusatian admixture from more recent German, since those are closer related to each other than they are to Slavs.

[Chad]

Riverman

The east-west gradient already existed in the Polish Trzciniec culture, so it is not the result of medieval German settlement, but of a different local demographic background, assimilated by the CWC.

Most of the local admixture is present in all Slavs, including those Ukrainians from Dnipro. The German shift in Poles is very clearly more Germanic/German specific overall. But of course, its hard to distinguish more ancient Germanic, Celtic and Lusatian admixture from more recent German, since those are closer related to each other than they are to Slavs.

Try a model with a northern source as well. If memory serves, Poles are between 4-15% YDNA N, depending on the region.. If this is all Uralic related, Ingria might make an okay source and help clarify things.

[Riverman]

Riverman

The east-west gradient already existed in the Polish Trzciniec culture, so it is not the result of medieval German settlement, but of a different local demographic background, assimilated by the CWC.

Most of the local admixture is present in all Slavs, including those Ukrainians from Dnipro. The German shift in Poles is very clearly more Germanic/German specific overall. But of course, its hard to distinguish more ancient Germanic, Celtic and Lusatian admixture from more recent German, since those are closer related to each other than they are to Slavs.

Try a model with a northern source as well. If memory serves, Poles are between 4-15% YDNA N, depending on the region.. If this is all Uralic related, Ingria might make an okay source and help clarify things.

I tried it with Balts, which clearly play a role in Poles, but Baltic eats up too much of the East Slavic/old Slavic ancestry. Saami is no good fit and Ingrian, while affecting just a minority of the samples at all, cannibalised old Slavic as well.
Another issue is, that with the Baltoslavic component being too extreme, Poles get split too much and German too high. I wasn't able to put up a better model and I know this one is not ideal, but it suffices to prove that significant German ancestry must be present, either of old Antiquity or newer Medieval and modern source. Ashkenazi on the other hand, just as a side note, plays a significant role - only possibly - in just two samples, both approaching an equivalent of about 1/16 Ashkenazi if its nothing else from the South East.

[Tomenable]

Ashkenazi on the other hand, just as a side note, plays a significant role - only possibly - in just two samples, both approaching an equivalent of about 1/16 Ashkenazi if its nothing else from the South East.

Out of curiosity, check if you detect Ashkenazi in this person (my maternal grandfather's brother):

WMeller_scaled,0.124067,0.125926,0.069767,0.059755,0.040007,0.023427,0.00893,0.009,0.007158,-0.024602,-0.003735,-0.006444,0.013082,0.017203,-0.012622,0.006099,-0.001434,-0.000633,0.004148,0.013006,-0.002496,-0.005564,0.008134,0.004699,-0.005149

He scores 2% Levant on FTDNA MyOrigins and also 3% Levant in Davidski's standard model:

Target: WMeller_scaled
Distance: 6.3992% / 0.06399218
Aggregated
52.0 Yamnaya_RUS_Samara
29.0 Anatolia_Barcin_N
16.0 WHG
3.0 Levant_PPNB

His K36 results also show 4% of MENA admix:

Kit number - WL5801302

Arabian 1.85 Pct
Near_Eastern 2.00 Pct
Omotic 0.07 Pct

And here is Davidski's custom model for him:

73.2% Polish
20.6% Lithuanian
3.8% Ashkenazi_Jew
2.4% German

[Tomenable]

Bolek is right about the German surnames of Poles. The best example is the surname Miller, given by the Germans to Polish millers.

This seems to be the case with my maternal grandfather. His surname is Meller, but he has a Slavic R1a haplogroup.

Also the distribution of Mellers in 1890 in eastern Germany mostly matches Polish settlement area, except for Sambia:

https://nvk.genealogy.net/map/1890:Meller

[Riverman]

Ashkenazi on the other hand, just as a side note, plays a significant role - only possibly - in just two samples, both approaching an equivalent of about 1/16 Ashkenazi if its nothing else from the South East.

Out of curiosity, check if you detect Ashkenazi in this person (my maternal grandfather's brother):

WM_scaled,0.124067,0.125926,0.069767,0.059755,0.040007,0.023427,0.00893,0.009,0.007158,-0.024602,-0.003735,-0.006444,0.013082,0.017203,-0.012622,0.006099,-0.001434,-0.000633,0.004148,0.013006,-0.002496,-0.005564,0.008134,0.004699,-0.005149

He scores 2% Levant on FTDNA MyOrigins and also 3% Levant in Davidski's standard model:

Target: WM_scaled
Distance: 6.3992% / 0.06399218
Aggregated
52.0 Yamnaya_RUS_Samara
29.0 Anatolia_Barcin_N
16.0 WHG
3.0 Levant_PPNB

His K36 results also show 4% of MENA admix:

Kit number - WL5801302

Arabian 1.85 Pct
Near_Eastern 2.00 Pct
Omotic 0.07 Pct

And here is Davidski's custom model for him:

73.2% Polish
20.6% Lithuanian
3.8% Ashkenazi_Jew
2.4% German

Well, Davidski surely got it better than myself, but in this simplified model, he doesn't pick up Ashkenazi, Sephardic or Roma, just an average-low amount of German (16). I used all Ashkenazi samples (standard was just Polish AJ), Sephardic, Roma and Armenian. He didn't pick up any, which is quite remarkable, since Dnipro is quite Slavic and he just picks up German.
But this could be a case of high Baltic masking a more Southern admixture. Since David gave him quite a bit of Baltic. However, with Ingrian, like suggested by Chad, he gets this:

Target: WM_scaled
Distance: 2.3339% / 0.02333852 | ADC: 0.25x RC
84.0
Ukrainian_Dnipro
16.0
German

---

Target: WM_scaled
Distance: 2.3301% / 0.02330093
80.6
Ukrainian_Dnipro
16.4
German
3.0
Ingrian

He just doesn't pick up any Ashkenazi, Armenian or Roma, which he might do even by chance otherwise. I don't really see it. He should test at 23andMe, they are best with picking up Ashkenazi ancestry, FTDNA can't be trusted as much on that matter.