G25 global ancient PCA of european individuals

[SouthDutch7991]

Here I've prepared a PCA with modern European individuals plotted against relevant ancient sources. I added in Papuan as well.

https://imgur.com/OH842pu.png


https://imgur.com/j3kyZMU.png

You can see the Anatolian-Steppe cline very well, you can also see that the vast majority of Europeans form a cluster between Basques/Italians and Russians respectively with some outliers. Note that the long tail leading from Eastern Europe to East Eurasians is all Finns, Northeast Russians, and Karelians.

[Ibericus]

There is something wrong about this plot. The WHG should not be plotting among modern europeans.

[Lemminkäinen]

Were Yamnayas mixed with East Asians? I didn't know this.

[Russki]

The gradient is from Finns to Italians and the native European (WHG) is essentially in the middle.


https://sun9-67.userapi.com/impg/JW5GLt7VISwHP-G4MjjC_s5U2xRCMOcF4bal8g/Zet1DD6l2RY.jpg?size=617x763&quality=95&sign=5bdd4f4aa9e3255880119b143b415480&type=album

[gixajo]

There is something wrong about this plot. The WHG should not be plotting among modern europeans.

It´s just a 2 dimension projection of 25 (if he used G25), actually WHG doesn´t plot among all those, althought as Russky soon would point that Baltics are the modern European peoples closest to them.:)

So yes, it´s not a realistic thing or at least is not what we are used to see, but we can present the reality also in this way. And if we want, playing with features and assigning different components to every axis, we can show very different PCAs, some of them surprising.

[Petalpusher]

This way it basically shows everything going up is more Eastern Eurasian in the broader sense and everthing going left is more African but besides that fact it's all a projection bias. Mesolithic are nowhere near that position on a proper multidimensional PCA:


https://chart-studio.plotly.com/~Urfaust/28.embed


I can even recreate the same projection bias taking a 2d capture of this 3d plot. When you navigate in it, you realize that's not where things are

https://i.postimg.cc/NF56rbQN/Capturepb1.png

(use CTRL + click to navigate with zoom/pan/orbital mode)

This is a more regular view we are used to see on 2d pca which suffers the least amount of projection distortion (for Euros at least, but still some)

https://i.postimg.cc/HLWC5vh4/Capture-RV1.png

[SouthDutch7991]

There is something wrong about this plot. The WHG should not be plotting among modern europeans.

Typically it shouldn't, but in 2D they will overlap. Weirdly, adding modern melanesians will collapse the 3rd dimension of WHG completely, to the point it plots between Yamnaya and Anatolian samples.

Try seeing what happens when you add/remove Papun from this in the 3-dimensional PCA. Does anyone know what causes this?

https://imgur.com/Ay1hWPE.png


TUR_Barcin_N,0.1175998,0.180118,0.0035312,-0.101158,0.0510443,-0.0483875,-0.0043582,-0.0069334,0.0362287,0.0807473,0.0079718,0.0118803,-0.0234545,0.0004691,-0.0419807,-0.0101913,0.0233091,0.0019866,0.0136954,-0.0097489,-0.0142249,0.0057723,-0.0041232,-0.0031658,-0.0043437
TUR_Tepecik_Ciftlik_N,0.108701,0.1731478,-0.0143308,-0.100534,0.0270818,-0.0402298,-0.0045828,-0.005019,0.02623,0.05745,0.008444,0.0080555,-0.0101832,0.0052298,-0.0352192,-0.0104748,0.0088662,0.0004432,0.005342,-0.0032828,0.0026827,0.0044515,-0.0071795,-0.0038258,-0.0045805
BRA_LapaDoSanto_9600BP,0.050765,-0.3095332,0.1206032,0.098515,-0.1156522,-0.017626,-0.284645,-0.3400012,-0.0136212,-0.0184422,0.002241,-0.003267,0.0004162,0.017478,-0.0065688,0.0064438,0.006493,-0.001267,0.0026898,0.003727,-0.0010482,0.007963,-0.002662,-0.0043378,0.001413
Dinka,-0.577083,0.0507765,-0.0003773,-0.0075098,-0.0053855,-0.0016735,-0.0176848,0.0204222,0.081145,-0.0969495,-0.02107,0.022742,-0.0383172,-0.0011698,0.0101452,-0.021347,0.0186125,-0.0094382,0.0241968,-0.0241678,0.002402,0.003308,0.001479,0.0009038,0.0 096995
ETH_4500BP,-0.511066,0.043668,0.000754,0.000969,-0.00277,-0.011435,0.050997,-0.045229,0.089172,-0.087838,-0.012991,-0.002997,-0.031219,0.000688,0.02158,-0.029965,0.027772,0.039273,0.00176,-0.009004,0.000374,0.006183,-0.003451,-0.00241,-0.000838
Gambian,-0.6064875,0.0671942,0.0197362,0.0114665,0.0021028, 0.0055777,-0.0358392,0.0387677,-0.0317353,0.0241463,0.0054132,-0.0036467,0.0162287,-0.0009865,0.011016,-0.0153362,0.009692,-0.0011823,-0.002116,-0.0032307,-0.001643,-0.0018137,0.0043958,-0.0001005,0.001377
Han,0.0206952,-0.4507105,-0.0082967,-0.0651579,0.0780846,0.0374475,0.0029055,-0.004804,-0.0150232,-0.0031146,-0.0495875,-0.0070573,0.0072707,-0.0081072,-0.0043305,0.0012535,0.0017542,-0.0008638,-0.0020455,-0.0102206,0.0122397,0.0069582,0.0141063,-0.0012925,-0.0022315
IRN_Ganj_Dareh_N,0.0430252,0.0664158,-0.1550722,0.0047158,-0.122669,0.0235384,0.017109,-0.0011998,-0.082546,-0.0544158,-0.0028258,-0.0016186,0.0044896,-0.0062756,0.0316498,0.0561384,-0.0054242,0.0068664,0.0136508,-0.0334162,0.00856,-0.028836,-0.0110678,-0.039331,0.0222254
IRN_Shahr_I_Sokhta_BA2_I8728,0.033009,-0.081242,-0.184789,0.125325,-0.09848,0.06777,0.002585,0.007615,0.03661,0.02606,-0.002273,0.010041,-0.005798,0.006744,0.004479,0.001326,0.008605,0.001 267,0.00352,0.001376,0.003743,-0.004946,-0.00037,-0.00253,-0.007305
Jarawa,-0.0219821,-0.2430289,-0.1324164,0.0986968,0.0323713,-0.0045841,-0.01012,0.0065044,0.0546974,0.0238044,0.0228054,0. 0027349,-0.0041532,0.0103562,-0.0108067,-0.0114441,0.0094529,-0.0019082,-0.0062221,0.0286778,-0.0046401,0.0111366,-0.0123402,0.0022668,0.0049844
Kura-Araxes_ARM_Kaps,0.1013025,0.133034,-0.07071,-0.043605,-0.0498555,-0.0092035,0.008108,-0.011653,-0.066368,-0.01631,0.0030855,0.011015,-0.019846,0.005849,0.0042075,-0.0102095,-0.0001305,-0.003927,-0.001697,0.0032515,0.003494,-0.000247,-0.0036975,-0.0090975,-0.0030535
Levant_PPNB,0.072847,0.1639064,-0.0316026,-0.1361132,0.0332986,-0.0645352,-0.0134426,-0.0147684,0.0741604,0.03601,0.0188046,-0.0150764,0.035738,0.0025596,-0.0217696,0.006099,0.0098048,-0.0013176,-0.0047264,0.0188088,-0.001797,0.0071472,0.0008872,-0.0056874,-0.0037602
MAR_Taforalt,-0.189857,0.0814452,-0.0242866,-0.085595,0.027636,-0.0552202,-0.0705968,0.0184146,0.155397,0.003499,0.0209156,-0.0318316,0.0747168,-0.0513334,0.0711988,-0.0363032,0.0052676,-0.066106,-0.1424162,0.0389938,-0.0376836,-0.1255322,0.0730118,-0.0137606,0.0164534
Nganassan,0.0476917,-0.4066181,0.1557885,0.0023902,-0.1594452,-0.0882129,0.0285066,0.0433367,0.0310876,0.0128477, 0.1028569,0.0094115,-0.0040734,-0.0261619,-0.0219731,-0.0123307,-0.0010952,0.0134165,0.0268365,-0.0008505,0.0431363,-0.0118954,0.0336096,0.0003977,0.0135556
WHG,0.1246365,0.116278,0.184789,0.189279,0.1546445 ,0.0464355,0.0131605,0.0372675,0.0890705,0.017768,-0.0153455,-0.015811,0.0159065,-0.0030275,0.053338,0.0582065,0.00502,0.016343,-0.0093015,0.055589,0.0944585,0.0111905,-0.049607,-0.160866,0.0170045
Yamnaya_RUS_Samara,0.1255849,0.089028,0.0426986,0. 1153479,-0.0287232,0.0450564,0.0036033,-0.0025642,-0.0559032,-0.0728943,0.0018222,3.32e-05,-0.0026924,-0.0233041,0.0366141,0.0157633,-0.0012316,-0.0017879,-0.0038408,0.0137704,-0.0031749,0.0007557,0.0110649,0.0186102,-0.004537
Yoruba,-0.6300625,0.0625011,0.022113,0.0167079,0.0005035,0 .0124741,-0.044417,0.0477673,-0.0488813,0.0327694,0.0046205,0.0007904,0.0230561, 0.0009509,0.0125232,-0.0096067,0.0070763,0.0004491,0.006022,-0.00299,0.0015542,0.0023156,-0.0017592,-0.0004711,-0.0004246

Papuan,-0.0459844,-0.2410866,-0.2637584,0.3092414,0.2069308,-0.509589,-0.001974,0.0070152,-0.041273,-0.0119546,-0.018902,0.000869,0.0013974,-0.00311,0.0001628,-0.0003186,-0.0021906,-0.0018498,-0.001081,-0.0015256,0.0023458,0.0018052,0.0012324,0.0018314, 0.0016048

[SouthDutch7991]

How it looks in 3D with individuals. Really weird effect. Usually you see changes when adding new samples but never something this drastic.

https://imgur.com/xZow5Q5.png

https://imgur.com/1jLQaY0.png

[SouthDutch7991]

Were Yamnayas mixed with East Asians? I didn't know this.

This PCA may exaggerate things, but as they have North Eurasian ancestry, they are mixed with ancient East Eurasians to some extent.


Target: Yamnaya_RUS_Samara
Distance: 0.0577% / 0.05770055
56.7 RUS_Karelia_HG
24.6 GEO_CHG
18.7 Kura-Araxes_ARM_Kaps


Target: RUS_Karelia_HG
Distance: 0.0642% / 0.06416915
68.5 RUS_AfontovaGora3
30.0 WHG
0.9 GEO_CHG
0.6 TUR_Tepecik_Ciftlik_N


It's hard to accurately model Ancient North Eurasian because it's just too far back on the human dispersion tree for modern human populations to make sense. Also BRA_LapaDoSanto_9600BP presumably has Western Non-African ancestry himself. NOTE THE HUGE DISTANCE IN THE MODEL. But most people put it somewhere between 1/4 and 1/3 Eastern Non-African.


Target: RUS_AfontovaGora3:AfontovaGora3
Distance: 0.2340% / 0.23395072
31.7 BRA_LapaDoSanto_9600BP
29.5 WHG
22.4 IRN_Ganj_Dareh_N
14.3 GEO_CHG
2.1 Nganassan



So all told, Yamnaya were probably between 10% and 15% Eastern non-African.


Target: Yamnaya_RUS_Samara
Distance: 0.0817% / 0.08172921
38.5 RUS_AfontovaGora3
25.4 GEO_CHG
19.5 Kura-Araxes_ARM_Kaps
16.1 WHG
0.5 TUR_Barcin_N

[Lemminkäinen]

This PCA may exaggerate things, but as they have North Eurasian ancestry, they are mixed with ancient East Eurasians to some extent.


Target: Yamnaya_RUS_Samara
Distance: 0.0577% / 0.05770055
56.7 RUS_Karelia_HG
24.6 GEO_CHG
18.7 Kura-Araxes_ARM_Kaps


Target: RUS_Karelia_HG
Distance: 0.0642% / 0.06416915
68.5 RUS_AfontovaGora3
30.0 WHG
0.9 GEO_CHG
0.6 TUR_Tepecik_Ciftlik_N


It's hard to accurately model Ancient North Eurasian because it's just too far back on the human dispersion tree for modern human populations to make sense. Also BRA_LapaDoSanto_9600BP presumably has Western Non-African ancestry himself. NOTE THE HUGE DISTANCE IN THE MODEL. But most people put it somewhere between 1/4 and 1/3 Eastern Non-African.


Target: RUS_AfontovaGora3:AfontovaGora3
Distance: 0.2340% / 0.23395072
31.7 BRA_LapaDoSanto_9600BP
29.5 WHG
22.4 IRN_Ganj_Dareh_N
14.3 GEO_CHG
2.1 Nganassan



So all told, Yamnaya were probably between 10% and 15% Eastern non-African.


Target: Yamnaya_RUS_Samara
Distance: 0.0817% / 0.08172921
38.5 RUS_AfontovaGora3
25.4 GEO_CHG
19.5 Kura-Araxes_ARM_Kaps
16.1 WHG
0.5 TUR_Barcin_N

So below 1%, which in common in modern Europeans using qpAdm, even more.

[SouthDutch7991]

So below 1%, which in common in modern Europeans using qpAdm, even more.

For modern East Asian-related yes. But for modern amerindian-related East Eurasian it should be as much as 10%. See how much of American-related ancestry ANE will score. Keep in mind modern amerindian is around 70-80% Eastern Non-African in ancient terms.

This is not a good model, the percents could be off by a good amount either way. But it demonstrates my point.

For a group that's more closely related to farmers:

Target: Yamnaya_RUS_Samara
Distance: 0.2034% / 0.20341464
87.9 Basque_Spanish
12.1 Karitiana

For modern Europeans with more Yamnaya related ancestry:

Target: Yamnaya_RUS_Samara
Distance: 0.1455% / 0.14545821
91.8 English
8.2 Karitiana

Target: Yamnaya_RUS_Samara
Distance: 0.1367% / 0.13674789 | R2P
92.2 Lithuanian_VZ
7.8 Karitiana

With Italians for control to see if Basque drift exaggerated things:

Target: Yamnaya_RUS_Samara
Distance: 0.1961% / 0.19612885
88.6 Italian_Tuscany
11.4 Karitiana

So yeah, you can see there is a good amount. Not great models, again, but it shows that a relatedness does exist and it's more than 1%

[Lemminkäinen]

For modern East Asian-related yes. But for modern amerindian-related East Eurasian it should be as much as 10%. See how much of American-related ancestry ANE will score. Keep in mind modern amerindian is around 70-80% Eastern Non-African in ancient terms.

Again, all North Europeans have ANE ancestry. PCA plots never reveal the whole picture. I don't know what happens if two eigenvectors figure earlier calculated 25 eigenvectors.

[SouthDutch7991]

Again, all North Europeans have ANE ancestry. PCA plots never reveal the whole picture. I don't know what happens if two eigenvectors figure earlier calculated 25 eigenvectors.

Yeah with yamnaya and Karelia_HG removed my family gets about 17-19% ANE. Which works out to 4-7% Eastern Non-African. That's pretty similar to what I see if I only model myself using Barcin_N, WHG, and BRA_LapaDoSanto_9600BP. Although it's a problem that we don't have a more amerindian-like ancient East-Eurasian source without the ANE component. I dont even know what that would look like.

[SouthDutch7991]

I ran some models to see how well you can estimate East-Eurasian from ancient sources, and if drift has made the models too inaccurate.

Target: RUS_AfontovaGora3:AfontovaGora3
Distance: 0.2340% / 0.23395072
31.7 BRA_LapaDoSanto_9600BP
29.5 WHG
22.4 IRN_Ganj_Dareh_N
14.3 GEO_CHG
2.1 Nganassan

Target: RUS_Karelia_HG
Distance: 0.0642% / 0.06416915
68.5 RUS_AfontovaGora3
30.0 WHG
0.9 GEO_CHG
0.6 TUR_Tepecik_Ciftlik_N

Target: Yamnaya_RUS_Samara
Distance: 0.0577% / 0.05770055
56.7 RUS_Karelia_HG
24.6 GEO_CHG
18.7 Kura-Araxes_ARM_Kaps

Target: Welsh
Distance: 0.0449% / 0.04486597
46.8 Yamnaya_RUS_Samara
38.8 TUR_Barcin_N
14.4 WHG

Welsh with no Yamnaya or Karelia_HG to see what their relatedness to AG3 is.

Target: Welsh
Distance: 0.0758% / 0.07582045
42.0 TUR_Barcin_N
22.0 WHG
19.0 RUS_AfontovaGora3
11.5 Kura-Araxes_ARM_Kaps
5.5 GEO_CHG

From this we expect to see the Welsh come up as about 6.1-6.3% Eastern Non-African from ANE alone. Finally some models using only base samples:

Target: Welsh
Distance: 0.1264% / 0.12636423
60.9 TUR_Barcin_N
31.6 WHG
6.3 BRA_LapaDoSanto_9600BP
1.0 Han
0.2 Yoruba

Simplified

Target: Welsh
Distance: 0.1264% / 0.12636423
92.5 West_Eurasian
7.3 East_Eurasian
0.2 Sub_Saharan_African

So it's about what we expect to see. Some extra Eastern non-African from CHG as well. Fairly accurate but still, there could be something I'm missing.