LOL, I loaded my Ancestry data to MyHeritage and got absolutely no eastern European or Balkan, according to MyHeritage that is.Originally Posted by Fungene
LOL, I loaded my Ancestry data to MyHeritage and got absolutely no eastern European or Balkan, according to MyHeritage that is.
I've never tried MyHeritage. Do you recommend it?
hell no. LOL its now giving me 1.3% Nigerian. LOL
Fungene (01-31-2018)
On comparisons between medieval elongated skull individuals and modern populations:
Several mentions of modern Romanians and Bulgarians in Veeramah, 2018, “Population genomic analysis of elongated skulls reveals extensive female-biased immigration in Early Medieval Bavaria” caught my attention. I thought the article would be relevant to a thread on Romanian genetics. Turns out, it is only marginally related to Romania or Bulgaria (although, judging by the spin in phys.org, one would think Romania and Bulgaria are of central interest.) But I’ll post this information anyway.
Thanks to alert DNA aficionados who share their efforts online:
User Matt (March 13, 2018), drawing on Global 25 suggests, “The Artificial Cranial Deformation (ACD) samples do indeed look very Southeast European, or more precisely at the modern day intersect of Central Europe-North Balkans.” http://eurogenes.blogspot.com.au/201...-in-early.html. March 12, 2018. (Comments)
Anthrogenica’s LukacszM’s post offers more, using Eurogenes’ K36.
https://anthrogenica.com/showthread....es-and-moderns
The information below is derived from the PCA in that post:
Comparison between individuals with intermediately elongated skulls and modern populations:
STR_355: between Wales and Cumbria
STR_220: between Holland and Mecklenburg
STR_360: closest to North Norway
STR_310: between central Europe and northern Italy
Comparison between individuals with elongated skulls and modern populations:
AED_125: between Austria and Central Romania
STR_228: between Hesse and Brabant
AED_513: Austria
NW_54: close to Italy Piedmont
BIM_ 33: close to Thuringia
STR_535: close to Italy Friuli
AEH_1: between Albania and Italy Abruzzo
AED_1108: between Macedonia and Albania
Last edited by Fungene; 03-14-2018 at 05:10 PM.
Here is information from a preprint useful for interpreting Moldovan, and of course, Romanian, calculator results.
All information is from Zheong et al. 2018. "Characterizing the genetic history of admixture across inner Eurasia"
https://www.biorxiv.org/content/bior...27122.full.pdf
Zheong et al., Table S1. lists ten new samples of Moldovans, all from Căplani village in Ștefan Vodă District, Moldova: MOL-005, MOL-008, MOL-015, MOL-024, MOL-058, MOL-064, MOL-065, MOL-066, MOL-067, MOL-069.
The first illustration is from Zheong et al., and zooms in on a portion of Fig. S1. “PC1 separates western and eastern Eurasian populations…. PC2 separates eastern Eurasians along the north-south cline and also separates Europeans from West Asians.”
All the southeastern European samples are coded in red (in red: Rom for Romanian and Mld for Moldavian). Below the first illustration are the full PCA and the key for interpreting population names.
The next illustration zooms in on Fig. S2. PC3 is intended to separate the West Asians (top) from the Europeans (bottom).
Below, the full PCA.
The Căplani Moldovans are looking very southeastern European.
No surprise: Romanians and Moldovans cluster with Europeans, more specifically with southeastern Europeans.
This has been known for some time. The news for us here is the ten new Moldovan samples.
Looks like FTDNA, still hanging on to its all-encompassing East Europe category, didn’t get the memo yet.
Last edited by Fungene; 06-08-2018 at 07:22 PM.
Just a reminder, the material in this thread is being posted in the international section so that people can respond in Romanian if they choose to.
Dorkymon has been busy modeling himself using Eurogenes Global 25.
https://anthrogenica.com/showthread....l=1#post414781
Dorkymon: it seems that there is a running commentary with an interlocutor as you are going through the steps of modeling yourself using ancient samples. Could you give the general reader just a little background?
I tried to produce some fits with Global 25 based on my admixture. The models might make sense for this region overall.
----------------------------------------------
1) Neolithic and Yamnaya migration
Input references
https://docs.google.com/document/d/1...it?usp=sharing
If you want to play around with this model, feel free to drop the Japanese and Eskimo references. Han eats them up anyway when the degree of penalisation is forced to 0 to account for intra-regional overlap (pen=0).
[1] "distance%=4.0061"Code:[1] "1. CLOSEST SINGLE ITEM DISTANCE%" Yamnaya_Samara:I0357 Yamnaya_Samara:I0370 Yamnaya_Samara:I0231 15.16141 16.43888 16.62792 Yamnaya_Samara:I0443 Yamnaya_Samara:I0444 Balkans_N:I2529 16.64321 16.76355 16.78688 Yamnaya_Samara:I0429 Balkans_N:I0634 17.63493 17.72987
Dorkymon
Yamnaya_Samara 48.2
Balkans_N 44.2
Iron_Gates_HG 4.6
Japanese 1.8
Eskimo_Naukan 0.8
Han 0.4
----------------------------------------------
2) Bronze Age (IN: Trypillia, Globular Amphora; OUT: Balkans Neolithic)
Since Trypillia:I926
is not really representative of the average and that's the only one I have, the following have been considered as Trypillian in order to get closer to the actual average as seen in Mathieson et al. 2018.
Our extra Trypillians are these 2 samples from Balkans_BA:
Input references
https://docs.google.com/document/d/1...it?usp=sharing
[1] "distance%=3.6477"Code:[1] "1. CLOSEST SINGLE ITEM DISTANCE%" Trypillia:adapted_I2175 Trypillia:adapted_Bul6 11.31006 11.41893 Trypillia:I1926 Yamnaya_Samara:I0357 12.96256 15.16141 Globular_Amphora:I2441 Trypillia:adapted_I1109 15.50558 16.04360 Yamnaya_Samara:I0370 Yamnaya_Samara:I0231 16.43888 16.62792
Dorkymon
Trypillia 58.4
Yamnaya_Samara 36.6
Han 2.6
Iron_Gates_HG 2
Onge 0.2
Pima 0.2
We can do better.
----------------------------------------------
3) Bronze Age (IN: Balkans_BA (minus I2163, who's like 3/4th Yamnaya), Poland_BA (minus I6537, which skews towards Balkans_BA); OUT: Trypillia, Globular Amphora)
We will run the script only with pen=0 in order to prevent the overlap between closely related groups (here Balkans_BA and Poland_BA).
ANI163 from Varna has been adapted into the Balkans_BA group in order to compensate for the clearly outlying I2163. ANI163 looks more logical, when compared with the rest of Balkans_BA.
Input references
https://docs.google.com/document/d/1...it?usp=sharing
pen=0Code:[1] "1. CLOSEST SINGLE ITEM DISTANCE%" Balkans_BA:adapted_ANI163 Balkans_BA:I2165 5.639317 5.890353 Balkans_BA:I4332 Balkans_BA:I4331 7.405923 7.531860 Balkans_BA:I3313 Balkans_BA:I2520 8.371998 9.196043 Poland_BA:I6531 Poland_BA:I6579 9.435305 9.956505
[1] "distance%=2.4347"
Dorkymon
Balkans_BA 80
Yamnaya_Samara 14.6
Han 4.8
Poland_BA 0.6
There's too much East Asian at pen=0, so the model can still be improved upon. Moving on.
----------------------------------------------
4) Bronze Age (IN: Yamnaya Ukraine (the one that's modelled as 100% steppe in Mathieson et. al 2018); OUT: Yamnaya Samara, Poland_BA, Han (I ran them separately and those that had even the smallest trace of non-East Asian/Siberian where removed)
Input references
https://docs.google.com/document/d/1...it?usp=sharing
[1] "distance%=3.4018"Code:[1] "1. CLOSEST SINGLE ITEM DISTANCE%" Balkans_BA:adapted_ANI163 Balkans_BA:I2165 5.639317 5.890353 Balkans_BA:I4332 Balkans_BA:I4331 7.405923 7.531860 Balkans_BA:I3313 Balkans_BA:I2520 8.371998 9.196043 Balkans_BA:I2163 Balkans_BA:I2176 10.163270 10.969873
Dorkymon
Balkans_BA 92
Yamnaya_Ukraine 4.8
Han 1.4
Pima 1.4
Onge 0.4
pen=0
[1] "distance%=2.5128"
Dorkymon
Balkans_BA 80.2
Yamnaya_Ukraine 15.4
Han 4.4
Hmm, it looks better now that it's ~3% East Asian/Siberian. The run without regularisation (pen=0) still overfits the Han with at least 1.5% more. Let's add Hungary_BA to see if anything changes.
----------------------------------------------
5) Bronze Age (IN: Hungary_BA)
We will run the script only with pen=0 in order to prevent the overlap between closely related groups (here Balkans_BA and Hungary_BA).
Input references
https://docs.google.com/document/d/1...it?usp=sharing
pen=0
[1] "distance%=2.4796"Code:Balkans_BA:adapted_ANI163 Balkans_BA:I2165 5.639317 5.890353 Hungary_BA:I1504 Hungary_BA:I7043 6.671139 6.864227 Balkans_BA:I4332 Balkans_BA:I4331 7.405923 7.531860 Balkans_BA:I3313 Hungary_BA:I7041 8.371998 8.479207
Dorkymon
Balkans_BA 70.4
Yamnaya_Ukraine 15.4
Hungary_BA 9.6
Han 4.6
Still overfitted for Han at pen=0, so the following model is the winner for me:
[1] "distance%=3.4018"Code:[1] "1. CLOSEST SINGLE ITEM DISTANCE%" Balkans_BA:adapted_ANI163 Balkans_BA:I2165 5.639317 5.890353 Balkans_BA:I4332 Balkans_BA:I4331 7.405923 7.531860 Balkans_BA:I3313 Balkans_BA:I2520 8.371998 9.196043 Balkans_BA:I2163 Balkans_BA:I2176 10.163270 10.969873
Dorkymon
Balkans_BA 92
Yamnaya_Ukraine 4.8
Han 1.4
Pima 1.4
Onge 0.4
PS: Look at MyHeritage from my signature. Coincidence?
----------------------------------------------
Bonus: 23andme vs nMonte
Just for the sake of curiosity, I'll run an overfitted model with all the moderns from Global 25 PCA_scaled at pen=0. Then I will combine the individual results into regions, following 23andme's nomenclature. After that, I will take my Broadly European bit from 23andme and try to break it down into regions, by following the proportions from nMonte. Let's see what we get.
Input references
https://drive.google.com/file/d/1IZ2...ew?usp=sharing
pen=0
[1] "distance%=0.5865"Code:[1] "1. CLOSEST SINGLE ITEM DISTANCE%" Moldovan:747_R01C02 Montenegrin:Montenegro7 2.723537 2.832282 Bosnian:Bosnian_14 Moldovan:747_R02C02 2.917600 2.958859 Hungarian:NA15207 Bosnian:Bosnian_13 2.965516 3.357705 Romanian:A362 Croatian:Croatia_Cro26 3.453939 3.454474
Dorkymon
Montenegrin,40.4
Russian_Kursk,15.8
Ukrainian,9.8
Macedonian,9.2
Belgian,7.4
Mordovian,5.8
Albanian,4.8
Greek,3.2
Han_NChina,2
Mongolian,1.4
Hungarian,0.2
Country to region
[1] "distance%=0.5865"
Dorkymon
Southeast European 57.8
Northeast European 31.4
Northwest European 7.4
Northeast Asian 3.4
Compared to:
23andme
South Euro 53.5% (41.8% Balkan)
East Euro 22.4%
Northwest Euro 7.3%
Ashkenazi 0.3%
Broadly Euro 14.7%
East Asian 1.8%
23andme (Broadly European broken down according to nMonte vs nMonte)
Code:South Euro 63.0% 57.8% East Euro 26.4% 31.4% Northwest Euro 8.5% 7.4% Ashkenazi 0.3% N/A East Asian 1.8% 3.4%
Known ancestry: 1/2 Romanian Northeast + 1/4 Romanian Southeast + 1/4 Romanian Bukovina Ukraine
23andme: 53.5% South Euro (41.8% Balkan), 22.4% East Euro, 7.3% Northwest Euro, 0.3% Ashkenazi, 14.7% Broadly Euro, 1.8% East Asian
MyHeritage: 92.2% Balkan, 5.8% English, 1.1% Eskimo, 0.9% Japanese
FTDNA: 45% Southeast Euro, 45% East Euro, 4% British Isles, 3% West Middle East, 2% NE Asia
Global 25: Anatolia_N 50.2%, EHG 32.6%, CHG 11.2%, WHG 3.8%, East Asian 2.2%
This is great fun, and clearly you enjoy doing this too. Now, since I don't do modeling myself I'm not going to comment on what I don't know, so what you’ll see here are just a few comments from an interested amateur.
Looking through past posts in other threads, you had started out maybe a few weeks ago with LBK_EN as the EEF component.
You’re a veteran of Kurd’s calculators, so you know that you have more eastern EEF than western (correct me if I’m wrong.)
In this series, your first model uses Gomolava (Vinca) and Yabalbovo for the Balkans Neolithic and looks really not bad at all.
For the next model, which you call Bronze Age, you included Smyadovo, which is Copper Age, Beli Breyag, which is Bronze Age, and a Verteba Cave Trypillian, which is Copper Age again. To this you added a Malak Preslavets sample, which is described as being from Criș Culture, so Neolithic. Plus you have the Kierzkowo Globular Amphora (I think.) When you get Pima and Onge, as you do with this model, the developer of Global 25 himself suggests you add Siberian and perhaps South Asian reference populations.
That might be something to be pursued, because otherwise, this model gives you the eastern EEF you know you have.
Dorkymon (06-14-2018)
Yes, the model can be improved. In terms of the EEF ratio, it's usually an almost even split, with EEF coming ahead in some instances, but not the other way around.
I liked the Balkans Neolithic run since it almost perfectly agreed with the LBK_EN one to which you have referred.
https://anthrogenica.com/showthread....l=1#post404366
I should look into qpAdm moving forward for an even greater precision, but glancing through Davidski's outputs, my final result shouldn't deviate significantly.
Consider that I usually cluster visibly Northeast of the Romanian pool, falling somewhere between Bosniaks and Croats.
And in one of his qpAdm models, both Romanian and Croatian references are made available, so we can attempt an educated guess.
Barcin_Neolithic
Romanian average: 52.3%
Croatian average: 49.4%
My average LBK_EN: 46.8%
My average Balkans_N: 44.2%
So, I should expect something in the region of 51%.
https://docs.google.com/spreadsheets...it?usp=sharing
Last edited by Dorkymon; 06-14-2018 at 11:04 PM.
Known ancestry: 1/2 Romanian Northeast + 1/4 Romanian Southeast + 1/4 Romanian Bukovina Ukraine
23andme: 53.5% South Euro (41.8% Balkan), 22.4% East Euro, 7.3% Northwest Euro, 0.3% Ashkenazi, 14.7% Broadly Euro, 1.8% East Asian
MyHeritage: 92.2% Balkan, 5.8% English, 1.1% Eskimo, 0.9% Japanese
FTDNA: 45% Southeast Euro, 45% East Euro, 4% British Isles, 3% West Middle East, 2% NE Asia
Global 25: Anatolia_N 50.2%, EHG 32.6%, CHG 11.2%, WHG 3.8%, East Asian 2.2%
Posting Permissions
Reply With Quote
