Mega QAPDM Models thread

[Birchy]

Have you got your G25 cords yet? If so, would you mind sharing your raw scaled modern pop results ?

Just requested them from Davidski, should get them very soon.

[Gannicus]

AndreiDNA did some f4 statistics for me also:

result: Gannicus Denmark_IA.SG Scotland_EIA.AG Dinka.DG 0.000099 0.000626 0.158 8974 8961 134189

what this means is: Gannicus is symmetrically related to Denmark Iron Age and Scotland Iron Age

Positive f4 means you are slightly more related to Denmark_IA

however the Z is insignificant (0.158)

result: Gannicus Scotland_EIA.AG Greek.HO Estonian.HO 0.000399 0.000240 1.658 8457 8403 134234

This suggests you are slightly Greek shifted relative to scotland_EIA, but the Z (1.658) is in the grey zone. Some would consider this result statistically significant and some wouldnt.

result: Gannicus Denmark_IA.SG Greek.HO Estonian.HO 0.000786 0.000211 3.720 11490 11347 182109

This suggests, without a shadow of a doubt, that you are Greek shifted relative to Denmark_IA. The F4 is positive (0.000786) and the Z is above 3 (3.720)

result: Gannicus Denmark_IA.SG Spain_IA_Celt.AG Estonian.HO 0.001931 0.000508 3.802 9923 9622 155551
Statistically significant (Z=3.8) and positive f4. You are Celtiberian shifted relative to Denmark IA

result: Gannicus Scotland_EIA.AG Spain_IA_Celt.AG Estonian.HO -0.000107 0.000562 -0.189 7914 7928 126089
Statistically insignificant (Z=-0.189) and negative F4. You are not celtiberian shifted relative to Scotland EIA

result: Gannicus Scotland_EIA.AG Bulgaria_LIA.AG Estonian.HO 0.000741 0.000880 0.842 3690 3647 58724
Statistically insignificant (Z=0.842) and positive f4. You are somewhat Thracian shifted relative to Scotland EIA.

result: Gannicus Denmark_IA.SG Bulgaria_LIA.AG Estonian.HO 0.001646 0.000749 2.196 4591 4472 72424
Statistically significant (Z=2.196) and positive F4. You are thracian shifted relative to Denmark IA.



OpenAI Breakdown:

1. Denmark IA vs Scotland EIA (North Sea baseline)
f4(Gannicus, Denmark_IA; Scotland_EIA, Dinka)
Z = 0.158


Interpretation:

Essentially zero.

You are symmetrically related to Iron Age Denmark and Iron Age Scotland.

This confirms a North Sea–balanced NW European profile, not a skew toward one side.

This matches qpAdm behavior where Denmark and Britain are hard to separate.

2. Eastern Mediterranean signal: Greece_HO tests
Scotland comparison
f4(Gannicus, Scotland_EIA; Greek_HO, Estonian_HO)
Z = 1.658


Slight Greek/Eastern Mediterranean pull relative to Scotland.

Not formally significant, but directionally consistent.

Denmark comparison
f4(Gannicus, Denmark_IA; Greek_HO, Estonian_HO)
Z = 3.720


Statistically significant

You are clearly more Greek/Eastern Mediterranean–shifted relative to Denmark IA

Key point:
The Eastern Mediterranean signal is stronger relative to Denmark than to Britain, which is exactly what your qpAdm + G25 models show.

3. Iberian (Celtiberian) signal
Denmark comparison
f4(Gannicus, Denmark_IA; Spain_IA_Celt, Estonian_HO)
Z = 3.802


Statistically significant

You are Celtiberian-shifted relative to Denmark IA

Scotland comparison
f4(Gannicus, Scotland_EIA; Spain_IA_Celt, Estonian_HO)
Z = -0.189


Not significant

No detectable Celtiberian shift relative to Scotland

Interpretation:
Your Iberian/Celtic affinity is shared with Britain, but exceeds Denmark, again reinforcing:

Britain ≈ Denmark for you overall, but Denmark lacks some southern/continental pull that you have.

4. Southeast European (Thracian / Balkan) signal
Scotland comparison
f4(Gannicus, Scotland_EIA; Bulgaria_LIA, Estonian_HO)
Z = 0.842


Weak, non-significant SE European pull relative to Scotland

Denmark comparison
f4(Gannicus, Denmark_IA; Bulgaria_LIA, Estonian_HO)
Z = 2.196


Borderline significant

You are more Balkan/SE-European-shifted relative to Denmark

[Xavier727]

Hey Xavier, would you mind trying the model I did in post #354 with the same references? I’m curious to see how it turns out for you.

It’s a kind of odd that you end up scoring more Germanic than me on this one, would you mind posting your G25 NW PCA, I’m just curious to see where you plot.

[drb234]

using England_MIA_LIA for Insular gets me roughly 1-2% higher germanic and insular but w/ slightly higher standard error values. I also tried using Denmark_Jutland_IA however it bloated my Germanic % at the cost of Insular, so i stuck with Denmark_IA instead

[Gannicus]

It’s a kind of odd that you end up scoring more Germanic than me on this one, would you mind posting your G25 NW PCA, I’m just curious to see where you plot.

Gannicus_MergedFile_officialDavidski_scaled,0.1297 58,0.139128,0.057322,0.03553,0.040315,0.011156,-0.00611,0.003923,0.005113,0.008565,-0.009094,0.009891,-0.008622,-0.020231,0.023208,0.010475,-0.004433,-0.000507,0.004148,0.003001,0.011729,0.004699,-0.00419,0.015544,0.001437

https://vahaduo.github.io/g25views/#NorthwestEurope

[drb234]

This one looks good
Attachment 146076
Wales_IA.AG + France_Aisne_IA.SG + Norway_IA.SG

use scotland_ia and create a custom pop with nor2b6 and r2055 for iron age gaul

[Gannicus]

With drb's rights plus additions:

[Gannicus]

Same rights just Vk522 taken out of the list because it is now a source.

[Xavier727]

@Gannicus

Thanks for sharing, yeah, you plot quite clearly on the Irish / Brittany side on G25. The higher Germanic in qpAdm might be due to elevated Steppe ancestry. I think it could very well represent actual Germanic ancestry; do you have any in your paper trail?

[Gannicus]

@Gannicus

Thanks for sharing, yeah, you plot quite clearly on the Irish / Brittany side on G25. The higher Germanic in qpAdm might be due to elevated Steppe ancestry. I think it could very well represent actual Germanic ancestry; do you have any your paper trail?

Yeah, I've got some German ancestry. From the Rhineland Palatinate, Northwest Germany (Lippe), and Hesse. This is mostly following the trail on Ancestry.com. The Germanic can come from multiple sources when it comes to my overall ancestry.

On 23&Me historical matches I do match VK522 as well as other ancient samples that cluster with IA Scandinavia:




If any of you want to check, here are the other matches: