PDA

View Full Version : How much of Bulgarian ancestry is Aegean/Sicilian-like versus NE European/Slavic?



Sikeliot
08-15-2019, 12:17 AM
How much of Bulgarian ancestry is East Mediterranean vs from Slavic settlers? And is there any additional Near Eastern that cannot be explained by the former?

How do they compare in this regard to Serbs, Romanians, Bosnians, and Croats?

Can someone model them? Thanks :)

Aspar
08-15-2019, 08:51 AM
I've made a PCA using Past3 a while ago based on a simple 3 population model: Greek_Crete, Italian_Tuscan and Ukrainian!
Using these three populations as references pretty much you can model very accurately all of the Balkans.

The PCA:

https://i.postimg.cc/nh3s5Xgg/fafa.jpg (https://postimages.org/)

Admixture stacked chart:

https://i.postimg.cc/TwMWjZZ3/fafa1.jpg (https://postimages.org/)

I found that using Greek_Crete as a reference instead of Sicilian_East gives much better fit and visual results when it comes to PCA for the Balkan populations!

Aspar
08-15-2019, 08:59 AM
I forgot to say that I am using unscaled coordinates for the modeling!

Serbs and Montenegrins:

https://i.postimg.cc/XJLdVszN/fafa2.jpg (https://postimages.org/)

Bosnians:

https://i.postimg.cc/L4T5bqcS/fafa6.jpg (https://postimages.org/)

Romanians:

https://i.postimg.cc/fbj0KHZ7/fafa3.jpg (https://postimages.org/)

Greeks:

https://i.postimg.cc/fLRkvD8Z/fafa4.jpg (https://postimages.org/)

Albanians:

https://i.postimg.cc/3RNJv9wx/fafa5.jpg (https://postimages.org/)

Sikeliot
08-15-2019, 10:04 PM
I found that using Greek_Crete as a reference instead of Sicilian_East gives much better fit and visual results when it comes to PCA for the Balkan populations!

Because Cretans have minor Slavic themselves, and lack the North African of Sicilians. But we can clearly see Bulgarians are, combined, much more Cretan + Tuscan like (Southcentral/southeast European) than they are like Ukrainians.

Bosnians seem the most genuinely 'Slavic' while Serbs and Romanians the most Tuscan-like, and Albanians have the least Slavic of all the Balkanites, less than Greeks themselves. Some Greeks are coming up as being as "Ukrainian" like as some of the South Slavs.

Kale
08-16-2019, 05:20 AM
Somewhat related to the topic at hand. A while back I ran an interesting, though by no means literal or particularly informative experiment.
Step 1) Run F3 outgroup stats to show shared drift modern populations (in HumanOrigins2068 dataset) have with key ancient references, these being...
Barcin_N Levant_N Iran_N Kotias Taforalt MA1 Anzick_Kennewick Iron_Gates_HG EHG Yamnaya_Samara Ami Australian Ust_Ishim
Step 2) Discover which modern populations could NOT be modeled as a mix of other moderns.
Step 3) Model other moderns with only the populations that could not be modeled themselves.

Bulgarians came out as...
33.85% Greek
33.15% Lithuanian
22.15% Sardinian
5.65% Saami
5.15% Georgian
0.05% Chukchi
With a quite good fit.
Most of North/West/Central Europe came out as a mix of Greek, Lithuanian, and Sardinian. The Saami & Georgian combo is a bit unique.

Compare to Romanians...
45.35% Lithuanian
27.5% Greek
17.1% Sardinian
6.2% Balkar (there were 3 groups from the Caucasus which couldn't be modeled)
2.7% BedouinB
1.15% Mansi

And Croatians...
49.25% Lithuanian
36.1% Greek
13.45% Sardinian
1.2% Somali (noise)

Albanian was pretty unexpected...
51.0% Sardinian
29.5% Chechen (the 3rd Caucasus pop)
13.1% Lithuanian
3.15% Georgian
2.4% Greek
And some noise...
0.7% Saharawi
0.15% Makrani
Interestingly, Italian_North would come out almost exactly halfway between Albanian and Basque.

I'm thinking about trying this experiment again with the new 1240k DG samples from Reichlab 37.2 dataset, but then again, there are less populations :\

Dorkymon
08-16-2019, 05:54 AM
I don't think that using moderns is the right approach for OP's question.

pegasus
08-16-2019, 06:16 AM
Somewhat related to the topic at hand. A while back I ran an interesting, though by no means literal or particularly informative experiment.
Step 1) Run F3 outgroup stats to show shared drift modern populations (in HumanOrigins2068 dataset) have with key ancient references, these being...
Barcin_N Levant_N Iran_N Kotias Taforalt MA1 Anzick_Kennewick Iron_Gates_HG EHG Yamnaya_Samara Ami Australian Ust_Ishim
Step 2) Discover which modern populations could NOT be modeled as a mix of other moderns.
Step 3) Model other moderns with only the populations that could not be modeled themselves.

Bulgarians came out as...
33.85% Greek
33.15% Lithuanian
22.15% Sardinian
5.65% Saami
5.15% Georgian
0.05% Chukchi
With a quite good fit.
Most of North/West/Central Europe came out as a mix of Greek, Lithuanian, and Sardinian. The Saami & Georgian combo is a bit unique.

Compare to Romanians...
45.35% Lithuanian
27.5% Greek
17.1% Sardinian
6.2% Balkar (there were 3 groups from the Caucasus which couldn't be modeled)
2.7% BedouinB
1.15% Mansi

And Croatians...
49.25% Lithuanian
36.1% Greek
13.45% Sardinian
1.2% Somali (noise)

Albanian was pretty unexpected...
51.0% Sardinian
29.5% Chechen (the 3rd Caucasus pop)
13.1% Lithuanian
3.15% Georgian
2.4% Greek
And some noise...
0.7% Saharawi
0.15% Makrani
Interestingly, Italian_North would come out almost exactly halfway between Albanian and Basque.

I'm thinking about trying this experiment again with the new 1240k DG samples from Reichlab 37.2 dataset, but then again, there are less populations :\

Wow Albanians are pretty different.

Coldmountains
08-16-2019, 07:27 AM
Bulgarians have around 40% early Slavic dna.

"sample": "Test1:Bulgarian_-_Bulgaria1",
"fit": 2.48,
"HUN_Avar_Szolad": 45,
"BGR_IA": 30,
"GRC_Mycenaean": 25,

"sample": "Test2:Bulgarian_-_BulgarianD6",
"fit": 2.265,
"BGR_IA": 47.5,
"HUN_Avar_Szolad": 40.83,
"GRC_Mycenaean": 11.67,

Aspar
08-16-2019, 08:37 AM
Bulgarians have around 40% early Slavic dna.

"sample": "Test1:Bulgarian_-_Bulgaria1",
"fit": 2.48,
"HUN_Avar_Szolad": 45,
"BGR_IA": 30,
"GRC_Mycenaean": 25,

"sample": "Test2:Bulgarian_-_BulgarianD6",
"fit": 2.265,
"BGR_IA": 47.5,
"HUN_Avar_Szolad": 40.83,
"GRC_Mycenaean": 11.67,

I am not sure if GRC_Mycenaean is the right sample to model Bulgarians...
The Bulgarian ethnogenesis happened in the Northern Balkans and the Pontic steppe, but even in the borders of modern Bulgaria up to date, there were no any Mycenaean-like samples found.

Coldmountains
08-16-2019, 08:52 AM
I am not sure if GRC_Mycenaean is the right sample to model Bulgarians...
The Bulgarian ethnogenesis happened in the Northern Balkans and the Pontic steppe, but even in the borders of modern Bulgaria up to date, there were no any Mycenaean-like samples found.

well Bulgaria_IA can be modelled as some kind of mix between GRC_Mycenaean and Hungary_BA. Also late roman Bulgaria would probably be more southern/Greek shifted because of migrations from the mediterranean

"sample": "Test1:BGR_IA",
"fit": 2.7384,
"GRC_Mycenaean": 59.17,
"HUN_BA": 40.83,

Adding some more reference populations i get this.

"sample": "Test1:Bulgarian",
"fit": 1.6074,
"HUN_Avar_Szolad": 38.33,
"BGR_IA": 31.67,
"Anatolia_IA": 10,
"GRC_Mycenaean": 10,
"HUN_BA": 8.33,
"Hun_Tian_Shan": 1.67,

So there is definetly some extra Greek-like affinity which can not only be explained by BGR_IA.

Dorkymon
08-16-2019, 09:05 AM
Bulgarians have around 40% early Slavic dna.

"sample": "Test1:Bulgarian_-_Bulgaria1",
"fit": 2.48,
"HUN_Avar_Szolad": 45,
"BGR_IA": 30,
"GRC_Mycenaean": 25,

"sample": "Test2:Bulgarian_-_BulgarianD6",
"fit": 2.265,
"BGR_IA": 47.5,
"HUN_Avar_Szolad": 40.83,
"GRC_Mycenaean": 11.67,

Yep, 40% is more or less what I also got with one of my models (Balkans K4) (http://185.144.156.77:3000/custom_calculators.html).
However, it needs be stressed that there are few samples for all populations who are being discussed here.
I'm not familiar with all, but for Romanians at least, I know that the samples letter coded with "A" come from Alba county, while those with "G" are from Gorj. The former is located in the West, while the latter in the Southwest. So that's not really representative for all Romanians and that can be verified by comparing against my results (75% Northeast, 25% Southeast).




Sample
Details
Fit
Mongols Et Al.
Pre-Slavic Greek World K4
Pre-Slavic Thracian World K4
Slavic Migration


1
Albanian_Averaged:Average

1.4233
0.83
40.83
20.83
37.5


2
Bosnian_Averaged:Average

2.0905
0
4.17
36.67
59.17


3
Bulgarian_Averaged:Average

1.3661
0.83
15
45.83
38.33


4
Croatian_Averaged:Average

1.8945
0
0
32.5
67.5


5
Custom:AGUser_Dorkymon

2.2293
4.17
7.5
30.83
57.5


6
Greek_Averaged:Average

1.1323
0
38.33
31.67
30


7
Greek_Crete_Averaged:Average

1.9342
0.83
65.83
15
18.33


8
Macedonian_Averaged:Average

1.9311
0
13.33
42.5
44.17


9
Moldavian_Averaged:Average

1.3282
0.83
24.17
26.67
48.33


10
Montenegrin_Averaged:Average

1.6245
0
9.17
40.83
50


11
Romanian_Averaged:Average

1.341
0.83
10.83
46.67
41.67


12
Serbian_Averaged:Average

1.7051
0
9.17
42.5
48.33


13
Slovenian_Averaged:Average

2.1959
0
0
29.17
70.83


14
Ukrainian_Averaged:Average

3.2418
0.83
0
0
99.17

Alain
08-16-2019, 09:06 AM
well Bulgaria_IA can be modelled as some kind of mix between GRC_Mycenaean and Hungary_BA. Also late roman Bulgaria would probably be more southern/Greek shifted because of migrations from the mediterranean

"sample": "Test1:BGR_IA",
"fit": 2.7384,
"GRC_Mycenaean": 59.17,
"HUN_BA": 40.83,

Adding some more reference populations i get this.

"sample": "Test1:Bulgarian",
"fit": 1.6074,
"HUN_Avar_Szolad": 38.33,
"BGR_IA": 31.67,
"Anatolia_IA": 10,
"GRC_Mycenaean": 10,
"HUN_BA": 8.33,
"Hun_Tian_Shan": 1.67,

So there is definetly some extra Greek-like affinity which can not only be explained by BGR_IA.

And the central Asian mixture from Proto-Bolgars is in Hun Tian Shan?

Aspar
08-16-2019, 09:39 AM
well Bulgaria_IA can be modelled as some kind of mix between GRC_Mycenaean and Hungary_BA. Also late roman Bulgaria would probably be more southern/Greek shifted because of migrations from the mediterranean

"sample": "Test1:BGR_IA",
"fit": 2.7384,
"GRC_Mycenaean": 59.17,
"HUN_BA": 40.83,

Adding some more reference populations i get this.

"sample": "Test1:Bulgarian",
"fit": 1.6074,
"HUN_Avar_Szolad": 38.33,
"BGR_IA": 31.67,
"Anatolia_IA": 10,
"GRC_Mycenaean": 10,
"HUN_BA": 8.33,
"Hun_Tian_Shan": 1.67,

So there is definetly some extra Greek-like affinity which can not only be explained by BGR_IA.

My understanding is that this extra Greek-like affinity comes from the 6th century Romans.
We still need more ancient samples from the Balkans but I think that with the few studies about the Romans that are due to arrive, a lot of light will be shed on the past of the Balkan nations as well!
There are already some speculations that the late Romans were "ITA_Collegno_Med_o1"-like and as such very Cretan-like as well!

We already know of early interactions between the Bulgarians and the Romans and mixing between each other by the account of the Bulgarian prince Kuber and the Sermesians(Roman citizens captured earlier by the Avars): https://archive.org/stream/wojnikov_mail_PRBG/PRBG_djvu.txt


A large group of local Thracian-Gettic-Dardanian
population, a so-called “Sermesians” Byzantine
population that was captured by the Avars and moved to
the surroundings of Sirmium, went together with Kuber

https://en.wikipedia.org/wiki/Kuber


Kuber (also Kouber or Kuver) was a Bulgar leader who according to the Miracles of Saint Demetrius led in the 670s, a mixed Bulgar and Byzantine Christian population, whose ancestors had been transferred from the Eastern Roman Empire to the Syrmia region in Pannonia by the Avars 60 years earlier. According to a scholarly theory, he was a son of Kubrat, brother of Khan Asparukh and member of the Dulo clan1

Greekscholar
08-16-2019, 09:13 PM
Fun thread. Does anyone know if the Greek samples are mainland, Aegean island, or a mix? Any guesses on why Ukrainians, in particular, are a good proxy for ancient Slavs? Thanks.

td120
08-16-2019, 09:21 PM
...

Albanian avg. slavic migration 37.5

Ain't' this kinda high?

Dorkymon
08-16-2019, 10:21 PM
Albanian avg. slavic migration 37.5

Ain't' this kinda high?

Yes, it should be closer to 15-20%, I'll post a more up-to-date model later.

Dorkymon
08-18-2019, 09:56 PM
Here it goes.
This is a test model that starts small with 5 proxies (AED 106 for NW Euro ancestry, Avar Szolad - Slavic-like, Han North - East Asian-like, Collegon CL25 - Supposedly Ancient Greek/Roman, Scythian Moldova - Thracian)



Sample
Fit
DEU MA--AED 106
HUN Avar Szolad
Han NChina
ITA Collegno MA O1--CL25
Scythian MDA


Albanian:Average
1.7724
0
21.67
0
26.67
51.67


Bulgarian:Average
1.3243
0
40.83
0.83
22.5
35.83


Croatian:Average
1.0388
10
63.33
0
9.17
17.5


Custom:AGUser_Dorkymon
2.3405
0
47.5
2.5
4.17
45.83


Greek:Average
1.3233
0
22.5
0
38.33
39.17


Greek_Crete:Average
2.255
0
4.17
0.83
78.33
16.67


Moldavian:Average
1.1382
0
40
0
30.83
29.17


Romanian:Average
1.2389
1.67
43.33
0
12.5
42.5


Serbian:Average
1.4962
0
50
0
15.83
34.17


Ukrainian:Average
1.6975
5
92.5
0.83
1.67
0




In each iteration, I will add one more population and look at how it influences the fit. If the improvement is insignificant then that population is not really needed for someone's model.
For example, STR 300 below seems to be crucial for Albanians and somewhat needed for Greeks. Perhaps an Illyrian?




Sample
Fit
DEU MA--AED 106
DEU MA O--STR 300
HUN Avar Szolad
Han NChina
ITA Collegno MA O1--CL25
Scythian MDA
Fit improvement


Albanian:Average
1.2809
0
33.33
22.5
0.83
15
28.33
0.4915


Bulgarian:Average
1.244
1.67
12.5
38.33
0.83
16.67
30
0.0803


Croatian:Average
1.0299
10.83
4.17
63.33
0
8.33
13.33
0.0089


Custom:AGUser_Dorkymon
2.2147
5
20
45
3.33
0
26.67
0.1258


Greek:Average
1.0418
0
21.67
21.67
0
29.17
27.5
0.2815


Greek_Crete:Average
2.1493
0
16.67
4.17
0.83
70.83
7.5
0.1057


Moldavian:Average
1.0529
4.17
12.5
38.33
0.83
30
14.17
0.0853


Romanian:Average
1.0627
7.5
18.33
39.17
0.83
9.17
25
0.1762


Serbian:Average
1.3706
2.5
15.83
47.5
0
10
24.17
0.1256


Ukrainian:Average
1.6975
5
0
92.5
0.83
1.67
0
0




The Chernyakhiv Goth; seems to affect significanly only me. For the rest it's optional.



Sample
Fit
DEU MA--AED 106
DEU MA O--STR 300
HUN Avar Szolad
Han NChina
ITA Collegno MA O1--CL25
Scythian MDA
UKR Chernyakhiv Shyshaky
Fit improvement


Albanian:Average
1.259
0
34.17
19.17
0.83
13.33
26.67
5.83
0.0219


Bulgarian:Average
1.1604
0
15
34.17
0.83
14.17
24.17
11.67
0.0836


Croatian:Average
1.0294
10.83
3.33
63.33
0
8.33
14.17
0
0.0005


Custom:AGUser_Dorkymon
1.4704
0
16.67
28.33
4.17
0
2.5
48.33
0.7443


Greek:Average
1.0064
0
22.5
17.5
0
26.67
25.83
7.5
0.0354


Greek_Crete:Average
2.0861
0
15
0
0.83
70.83
3.33
10
0.0632


Moldavian:Average
1.0011
0.83
14.17
35
0.83
26.67
12.5
10
0.0518


Romanian:Average
0.97
3.33
19.17
35.83
0.83
5
22.5
13.33
0.0927


Serbian:Average
1.3705
2.5
15.83
46.67
0
9.17
25
0.83
0.0001


Ukrainian:Average
1.6991
5
0
93.33
0.83
0.83
0
0
-0.0016




Baltics, seem to be important for Ukrainians and to a lesser extent for Moldovans, Croatians, Serbians and myself.



Sample
Fit
Baltic EST MA
DEU MA--AED 106
DEU MA O--STR 300
HUN Avar Szolad
Han NChina
ITA Collegno MA O1--CL25
Scythian MDA
UKR Chernyakhiv Shyshaky
Fit improvement


Albanian:Average
1.2654
7.5
0
31.67
11.67
0
13.33
34.17
1.67
-0.0064


Bulgarian:Average
1.0951
12.5
0
18.33
19.17
0.83
14.17
24.17
10.83
0.0653


Croatian:Average
0.8887
27.5
5
10
34.17
0
9.17
14.17
0
0.1407


Custom:AGUser_Dorkymon
1.2959
23.33
0
23.33
6.67
3.33
0
6.67
36.67
0.1745


Greek:Average
0.9499
10
0
23.33
8.33
0
29.17
26.67
2.5
0.0565


Greek_Crete:Average
2.0861
0
0
15
0
0.83
70.83
3.33
10
0


Moldavian:Average
0.7601
25
0
15.83
9.17
0
29.17
20.83
0
0.241


Romanian:Average
0.9075
15
0.83
21.67
21.67
0.83
7.5
24.17
8.33
0.0625


Serbian:Average
1.2484
22.5
0
21.67
21.67
0
10.83
23.33
0
0.1221


Ukrainian:Average
1.2713
47.5
0
4.17
36.67
0
3.33
8.33
0
0.4278




The best model for Bulgarians from what I have looks to be



Sample
Fit
DEU MA--AED 106
DEU MA O--STR 300
HUN Avar Szolad
Han NChina
ITA Collegno MA O1--CL25
Scythian MDA


Bulgarian:Average
1.244
1.67
12.5
38.33
0.83
16.67
30



That's roughly
38% Slavic
30% East Balkan/Thracian
17% South Balkan/Early Greek (Aegean)
13% West Balkan/Illyrian
2% Germanic
1% East Asian

eastara
08-19-2019, 01:23 PM
Thank you, Dorkymon.
Is it possible to include also the Alan and Saltovo-Mayaki samples, do they have any connection to the modern Bulgarians? This is in view of this study as it tries to prove a major steppe-Caucasus influence.

https://www.biorxiv.org/content/10.1101/687384v3?fbclid=IwAR0iwGJtOTJZi9cqE1gtvV53qp5CZU9y KwqTqm2gMPE7S5xADRbuwy0jw6c

Dorkymon
08-19-2019, 02:52 PM
Thank you, Dorkymon.
Is it possible to include also the Alan and Saltovo-Mayaki samples, do they have any connection to the modern Bulgarians? This is in view of this study as it tries to prove a major steppe-Caucasus influence.

https://www.biorxiv.org/content/10.1101/687384v3?fbclid=IwAR0iwGJtOTJZi9cqE1gtvV53qp5CZU9y KwqTqm2gMPE7S5xADRbuwy0jw6c

The change is marginal, so those Alans can easily be left out.

"PCA results suggest genetic connection between contemporary Bulgarians andthe ancient individuals AlanDA243, AlanDA164 and Alan DA146 belonging toSM culture.

Our PCA (Fig.2) also revealed indirect connection between contemporaryBulgarians and central Asian Bronze Age nomads of East Iranic origin known asKangju group. This relation however is dependent on the presence of sample Alan DA146 from Saltovo-Mayaki (Saltovo, SM for short) culture on the PCAPlot and disappears if we remove this sample from the plot."


Here's the model with the addition of Saltovo-Mayaki's DA146.



Sample
Fit
DEU MA--AED 106
DEU MA O--STR 300
HUN Avar Szolad
Han NChina
ITA Collegno MA O1--CL25
RUS Alan MA--DA146
Scythian MDA


Bulgarian:Average
1.2352
0.83
12.5
39.17
0.83
15
1.67
30




And all 6 of those Alans



Sample
Fit
DEU MA--AED 106
DEU MA O--STR 300
HUN Avar Szolad
Han NChina
ITA Collegno MA O1--CL25
RUS Alan MA
Scythian MDA


Bulgarian:Average
1.223
1.67
14.17
38.33
0.83
13.33
3.33
28.33

Sorcelow
08-20-2019, 03:20 AM
[1] "distance%=2.3961"

Bulgarian

Ukrainian,51
BGR_IA,31.2
GRC_Mycenaean,14.6
Lebanese_Christian,3.2

Aspar
08-20-2019, 11:37 AM
IMO, you can not estimate the Slavic contribution in the autosomal DNA of the Bulgarians without observing their uniparental haplogroups.

The Bulgarian uniparentals are diverse and don't show recent founder effect, apart from the highly identifiable Slavic marker I-Y3120 with founder effect approximately 2200 ybp according to the scientist Ken Nordtvedt and also according the YFULL tree!

In that regard I will use the study "Y-Chromosome Diversity in Modern Bulgarians: New Clues about Their Ancestry" (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3590186/) to estimate the Slavic input in the modern Bulgarians and to find a correlation with a certain model using unscaled G25!

The Bulgarian ydna, 808 testes:

https://i.postimg.cc/qRBcLTLL/Bulgarians-ydna.png (https://postimages.org/)

As I already said, the marker I-Y3120 is the most identifiable Slavic marker, followed by R1a. Bear in mind, not all of R1a is of the Slavic variety but there is some R-Z93 as well in the range of 1-2%.However, there is probably some small percentage of other haplogroups that came with the Slavs. Using this parameters, the Slavic input is between 36-38%.

The best model for the Bulgarians it seems is this one:

[1] "distance%=0.7886"

Bulgarian

HUN_Avar_Szolad,37.6
Scythian_MDA,31.2
ITA_Collegno_Med_o1,16.8
BGR_IA,14.4

Although in the model were included other Balkan samples like HRV_IA, this samples is not preferred!
The main contributors of the Bulgarian ethnogenesis seem to be the people living in the Eastern Balkans, Thracians, Dacians, Getae, Scythians, along with some Greco-Roman(Byzantine) input and the Slavs! We can also observe some small percentage of EA lineages but when I include such populations in the model, the fit gets worse and that is because such kind of input is already represented by 'Scythian_MDA', which had some small EA input itself!

Dorkymon
08-20-2019, 12:30 PM
IMO, you can not estimate the Slavic contribution in the autosomal DNA of the Bulgarians without observing their uniparental haplogroups.



Bulgarian

HUN_Avar_Szolad,37.6
Scythian_MDA,31.2
ITA_Collegno_Med_o1,16.8
BGR_IA,14.4


How is this not identical to what I already shared? The only difference is that those Bulgarian samples prefer the Middle Ages sample STR 300, instead of Iron Ages BGR IA, but they both serve as proxies for a pre-Slavic population between the Danube and the Greeks. I also don't believe that small but still existent East Asian influence stuck around since the Iron Age. It's likely from the Turkics/Mongols who came to these parts later.

"The best model for Bulgarians from what I have looks to be



Sample
Fit
DEU MA--AED 106
DEU MA O--STR 300
HUN Avar Szolad
Han NChina
ITA Collegno MA O1--CL25
Scythian MDA


Bulgarian:Average
1.244
1.67
12.5
38.33
0.83
16.67
30


"

Aspar
08-20-2019, 01:10 PM
How is this not identical to what I already shared? The only difference is that those Bulgarian samples prefer the Middle Ages sample STR 300, instead of Iron Ages BGR IA, but they both serve as proxies for a pre-Slavic population between the Danube and the Greeks.

"The best model for Bulgarians from what I have looks to be



Sample
Fit
DEU MA--AED 106
DEU MA O--STR 300
HUN Avar Szolad
Han NChina
ITA Collegno MA O1--CL25
Scythian MDA


Bulgarian:Average
1.244
1.67
12.5
38.33
0.83
16.67
30


"

I never said that your model is not good or anything of that sort.
I am just giving mine insight on the matter.

Dorkymon
08-20-2019, 02:03 PM
I never said that your model is not good or anything of that sort.
I am just giving mine insight on the matter.

Ah, I thought you didn't see it, since you arrived at the same conclusion in the end.

ph2ter
08-20-2019, 02:15 PM
[1] "distance%=0.6029"


Bulgarian


SLAVIC,38.4
ILLYRIAN-THRACIAN-DACIAN,36.4
GRECO-ROMAN,21.2
BALTIC,2.6
BALTIC-INGRIA,1.4

or

[1] "distance%=0.6111"


Bulgarian


SLAVIC,36.4
ILLYRIAN-THRACIAN,29.8
GRECO-ROMAN,15
ANATOLIAN,6
LEVANTINE,4.8
BALTIC,2.4
IBERIAN,1.2
INGRIA,0.8
TARTESSIAN,0.8
ALANIC,0.4
ANDAMAN,0.4
BALTIC-INGRIA,0.4
GOTHS,0.4
LUSITANIAN,0.4
HIMALAYAN,0.2
JAPANESE,0.2
PACIFIC,0.2
SOUTH_AFRICAN,0.2

artemv
08-22-2019, 01:31 PM
IMO, you can not estimate the Slavic contribution in the autosomal DNA of the Bulgarians without observing their uniparental haplogroups.

The Bulgarian uniparentals are diverse and don't show recent founder effect, apart from the highly identifiable Slavic marker I-Y3120 with founder effect approximately 2200 ybp according to the scientist Ken Nordtvedt and also according the YFULL tree!

In that regard I will use the study "Y-Chromosome Diversity in Modern Bulgarians: New Clues about Their Ancestry" (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3590186/) to estimate the Slavic input in the modern Bulgarians and to find a correlation with a certain model using unscaled G25!

The Bulgarian ydna, 808 testes:

https://i.postimg.cc/qRBcLTLL/Bulgarians-ydna.png (https://postimages.org/)

As I already said, the marker I-Y3120 is the most identifiable Slavic marker, followed by R1a. Bear in mind, not all of R1a is of the Slavic variety but there is some R-Z93 as well in the range of 1-2%.However, there is probably some small percentage of other haplogroups that came with the Slavs. Using this parameters, the Slavic input is between 36-38%.
We should not forget that share of Y-chomosome markers can differ greatly from share in overall ancestry.
We all remember how greatly differs share of Anatolian farmers Y-chr branches and share of their ancestry in autosomes.

And yes, again - Y3120 is the only one young clade in the list. Most of other happlogroups include both Slavic, non-Slavic and "not possible to classify" branches.
R-Z93 also has some Slavic sub-branches, and R1a not Z93 has many non-Slavic branches. R1b, E, I-M223, G, even J2 - all have likely Slavic branches.
By the way, the olny one Bulgarian I-M223 sample currently on y-full is from a likely Slavic branch.



The best model for the Bulgarians it seems is this one:

[1] "distance%=0.7886"

Bulgarian

HUN_Avar_Szolad,37.6
Scythian_MDA,31.2
ITA_Collegno_Med_o1,16.8
BGR_IA,14.4

Although in the model were included other Balkan samples like HRV_IA, this samples is not preferred!
The main contributors of the Bulgarian ethnogenesis seem to be the people living in the Eastern Balkans, Thracians, Dacians, Getae, Scythians, along with some Greco-Roman(Byzantine) input and the Slavs! We can also observe some small percentage of EA lineages but when I include such populations in the model, the fit gets worse and that is because such kind of input is already represented by 'Scythian_MDA', which had some small EA input itself!

Pre-Slavic Roman era population is understimated here, just because BGR_IA sample is too old. If it is possible, one should try to use in model only contemporary populations.

Michał
08-23-2019, 01:30 AM
@artemv

How would you define a "Slavic branch"?

As for me, I would define it as a branch that is present in at least two out of the three major Slavic groupings (which are the West, East and South Slavs) and shows much higher frequency among those Slavs than in any remaining populations (making possible exception for some non-Slavic populations that are known to have been very strongly influenced by the Slavs, like the Hungarians, Romanians, Austrians, East Germans, etc.). This makes it almost certain that such a branch was originally associated with the expanding Slavs, and was not just a local non-Slavic branch that was assimilated by the incoming Slavs, or a foreign branch that was introduced to some Slavic groupings in more recent times.




R-Z93 also has some Slavic sub-branches

Which ones do you have in mind?



and R1a not Z93 has many non-Slavic branches.

This is correct, but these are extremely rare in the Balkans (or among the Slavs from other parts of Europe), so their contribution can be neglected in all similar calculations.



R1b, E, I-M223, G, even J2 - all have likely Slavic branches.
I am aware of only one Slavic subclade of R1b, and this is R1b-PH2302 (also known as R1b-Y14300 or L23EE). As for the remaining haplogroups (E, G, J2), could you please let us know which "Slavic" sub-branches you have in mind?



By the way, the olny one Bulgarian I-M223 sample currently on y-full is from a likely Slavic branch.

Again, which particular branch do you have in mind? And why do you think it should be called "Slavic"?

artemv
08-23-2019, 01:25 PM
@artemv

How would you define a "Slavic branch"?

Slavic branch is a branch, that was present among Early Slavs, who expanded in the 6th century.
As far as we do not have Early Slav DNA (and will not have, as far as Early Slavs used cremation as a burial ritual), we can only suppose looking at modern Ychr distribution.
This definition gives us one important formal criteria: only branches that were formed 1500 ybp or earlier can be Slavic.



As for me, I would define it as a branch that is present in at least two out of the three major Slavic groupings (which are the West, East and South Slavs) and shows much higher frequency among those Slavs than in any remaining populations (making possible exception for some non-Slavic populations that are known to have been very strongly influenced by the Slavs, like the Hungarians, Romanians, Austrians, East Germans, etc.). This makes it almost certain that such a branch was originally associated with the expanding Slavs, and was not just a local non-Slavic branch that was assimilated by the incoming Slavs, or a foreign branch that was introduced to some Slavic groupings in more recent times.

What you write here is not a definition. It is criteria. If we are talkig about historical migrations and ancestry (of Bulgarian Ychr branches, for example) the definition I gave earlier is the only possible definition. However, that definition doesn't give us a useful criteria of a Slavic happlogroup (except for happlogroup formation time, which is of course not enough to distinct Slavic happlogroups from not-Slavic).

I agree that happlogroups that follow these your criteria are likely Slavic, but:
- in case if we find some happlogroup, that follows all these criteria, but we know that it was not present among Early Slavs (got into several Slavic groups from elsewhere and became absent in the place where it came from), it should not be considered Slavic;
- if we have some happlogroup, that is present only in one Slav group, it can be well derived from Early Slavs.

I understand that it would be difficult for me to argue about each of the small groups, but let me ask you two questions:
1. What do you think was the fate of small Ychr branches of Early Slavs? (My answer is: some of them probably didn't survive, others likely survived only in a single major group of Slavs. This means, that even if Ychr branch is found only in a single major Slavic group it should not be labled automatically as not-Slavic).
2. All the current European populations have some percent of happlogroups E, G, J2. For example in Ukraine together they make about 15% of population. Do you think all of them are relatively recent immigrants from Balkans?

You formulated your criteria to find happlogroups, that can be called Slavic without serious doubt. This means, you automatically add to non-Slavic all the cases with some doubt, and, thus get much lower estimation of Slavic-Ychr.
Please note also, that Slavic Y-chromosome can come from a group that contributed little or close to nothing to Slavic autosomal DNA.



This is correct, but these are extremely rare in the Balkans (or among the Slavs from other parts of Europe), so their contribution can be neglected in all similar calculations.

Alltogether small branches will give significant percent. It would be a good idea to separate branches let's say for those definitely Slavic, those that were definitely present in pre-Slavic Romanized population of the region, those that definitely arrived later and branches under doubt. You will see that percent of those under doubt will be surprisingly high.



I am aware of only one Slavic subclade of R1b, and this is R1b-(also known as R1b-Y14300 or L23EE). As for the remaining haplogroups (E, G, J2), could you please let us know which "Slavic" sub-branches you have in mind?

I do not see really big, 100% sure Slavic branches, but there are number of old enough and potencially Slavic happlogroups. Many of them will turn out to be Slavic. Here are some examples:
https://www.yfull.com/tree/J-Y29718/
https://www.yfull.com/tree/R-YP5271/
https://www.yfull.com/tree/G-Z39501/
https://www.yfull.com/tree/E-Y41959/

https://www.yfull.com/tree/R-Y14088/
Let's discuss this group of close relatives, who desided to make a test alltogether.
You will probably say, their happlogroup came to region with some celtic group, or even stayed somewhere in the region since BB times. But this doesn't contradict the idea, that they could have been part of some Early Slav group in the 5th-6th centuries.
There are some Ychromosome branches, that stayed for many millenia on the British Isles, without emigrating outside. But it looks like this is the case only for some isolated ares: people always moved between the regions, so did Ychr happlogroups. Common expectation of one's male-line ancestors to live in the same place for several millenia in most cases would be wrong.

P.S. Ok, it looks that I've likely mistaken about Slavic R1a-Z93.