• Don't want to see ads? Install an adblocker like uBlock Origin or use a Europe-based privacy-friendly browser like Vivaldi or Mullvad.

Admixtools My qpAdm Results Please Share Yours!

1772496192085.png

1772496217578.png

1772496132279.png


1772496472044.png

p=0.8
1772496408109.png

1772496565583.png


1772496541919.png



NUE001 is the Old Kingdom Egyptian from the Morez study

1772496671451.png
 

Attachments

  • 1772496439789.png
    1772496439789.png
    18.6 KB · Views: 85
I wanted to know how much ancestry I have considering the Bronze / Iron Age. Below is the best model I made, with p-value of 0.173 , low SE, all z > 2 (statistically supported).

Before I’ve tried to reconstruct the two Indo-European waves into Iberia separately, the Bell Beaker in the Bronze Age and the Celts in the Iron Age. I try to make some models with Portugal_BA and several Central European populations like Halberstadt (Urnfield) , Knoviz , Hallstatt or La Tene, but they sit on a shared cline and create instability, high SE and not good p-values.
When the sources sit on a shared cline that creates collinearity, qpAdm struggles because left and right are similar. That doesn’t mean the model is invalid, it means that the proportions become unstable, the SEs inflate and the interpretation becomes muddy.

By the Iron Age, Iberia already had a stabilized Bell Beaker-derived gene pool with continental interaction embedded in it. Here I use an already admixed historical Iberian population, this is statistically cleaner and historically more realistic. This model avoids the earlier collinearity problem because Spain_IA_Celt already includes Iberian Bronze Age substrate with some steppe ancestry (via Bell Beaker) and Iron Age continental input. Here I’m no longer trying to separate gene flows from the Bronze Age vs flows from the Iron Age, I collapsed them into a single historical population, that removes artificial dimensionality.

This model says that my ancestry is statistically consistent with being primarily derived from the genetic profile of Celtiberian-like Iron Age Iberia, with a modest eastern Mediterranean–related component and a small North African component.
It does not mean that I’m 80% Celtiberian, it means that my ancestry fits within that Iron Age Iberian genetic structure. That it is statistically compatible with that genetic profile. The Italy_Imperial_oAnatoliaCaucasus component is not equivalent to “Roman ancestry”, it is a proxy for Eastern Mediterranean / Anatolian-shifted drift. This could reflect Punic-era or Roman/post-Roman-era gene flow. This model is statistically stable, historically coherent, not overparameterized and avoids collinearity.
Spain_IA_Celt represents three individuals from La Hoya in northern Spain.
lwbD7Uf.png

The model works well also for some samples of Spain_IA, that occupy the same position on the PCA than Spain_IA_Celt.
LmDy9VZ.png


lc47ZAz.png

​
 
The Italy_Imperial_oAnatoliaCaucasus component is the responsible for my slight shift toward Anatolia Chalcolithic in the PCA, that may represent a slightly more eastern Mediterranean farmer ancestry. So, I revised my basal ancestry model to include Greece Nea Nikomedeia EN, instead of Barcin. Nea Nikomedeia has a predominantly Barcin Anatolia Neolithic ancestry, but also additional Central Anatolian Neolithic, Tepecik-Çiftlik related, and is slightly more CHG shifted.

gKh4Vwj.png
 
In this two-dimension PCA the sample most close to me is Hungary MBA Vatya RISE480. Not because any direct ancestral relationship, we have independent historical admixture events, but because in the PC1 / PC2 axis she should have similar proportions of "basal" ancestry (Barcin, Yamnaya,….).​

mKBm2cs.png
 
Last edited:
Tautalus said:
Not because any direct ancestral relationship, we have independent historical admixture events, but because in the PC1 / PC2 axis she should have similar proportions of basal ancestry (Barcin, Yamnaya,….).​
Click to expand...
That is the same phenomen occurring between some modern northern Italians and some modern Iberians isn't it?
 
HereToLearn said:
That is the same phenomen occurring between some modern northern Italians and some modern Iberians isn't it?
Click to expand...
The people of the Vatya culture did not migrate to Iberia, nor did they participate in the ethnogenesis of the Iberian peoples, but they share “basal” ancestry, ancestral components or deep ancestry components related to ANF + WHG + Steppe.
Between Northern Italy and Iberia, the same proportional effect is observed in the PCA (similar deep ancestry ratios), but they also share some historical migrations, like the Cardial Neolithic expansion, Bronze/Iron Age migrations (Ligurians ?), Roman-era mobility.​
 
1773159560184.png


My FST genetic distances to various populations across the world. This is my merged DNA file (Ancestry, 23&Me v4, 23&Me v5, MyHeritage)
 
I can be modeled as a single-source left with Early Medieval Tuscany and Lazio.

1773167420730.png

1773167656947.png

1773167937052.png


1773167989439.png
 
My F2 stats

1773170723801.png

1773170779893.png

1773170830209.png
 
In the spirit of St. Patrick's day I modeled myself as 100% Medieval Irish:

1773697853004.png

References:
1773697894178.png

1773697919320.png
 
What do you guys think about these outgroups. I used to use Estonia_BA.SG to as a Slavic outgroup, but that shared too much drift with Slavs. So I decided to use BHG and EHG instead
MDP91oc.png
 
I pooled these 2 samples together

1774307216216.png


And modeled myself as 100% Wales IA

1774307279284.png

pat wt dof chisq p f4rank Wales_IA.AG feasible best dofdiff chisqdiff p_nested
0 0 9 5.875532596691188 0.7523080292208336 0 1 TRUE NA NA NA NA

References:

1774307392277.png
 
I separated Kartal_A from Kartal_B as it is done in the study, Penske et al. 2023.

This demonstrates the high-steppe admixed population that mixed with the local farmers in the Balkens during the Copper age.

This serves as a good proxy for the Trans-Adriatic Steppe-enriched Balkan populations that were formative in Apulia.

1774447848727.png
1774447873203.png
 
I can be modeled completely as an Upper Paleolithic Caucasian. This is a core West Eurasian ancestry, and was seminal to the majority of WHG.

This sample comes from Allentoft et al. 2024

1775155765107.png

1775156021168.png
 
1775349098462.png

1775349336527.png
 
Wonder how you can model han chinese and Australian aboriginals. Those are two groups I’d like to look up origins on.
 
Jovialis said:
I can be modeled completely as an Upper Paleolithic Caucasian. This is a core West Eurasian ancestry, and was seminal to the majority of WHG.

This sample comes from Allentoft et al. 2024

View attachment 19467
View attachment 19468
Click to expand...
I wonder why the model passed because you definitely need some ANE to account for steppe ancestry. It should be UP Caucasian+ANE for most West Eurasians.
 
Anfänger said:
I wonder why the model passed because you definitely need some ANE to account for steppe ancestry. It should be UP Caucasian+ANE for most West Eurasians.
Click to expand...
It is probably just a fluke. Frankly, I wouldn't trust the results of this model... TBH, I am not sure what kind of outgroups would be optimal for such a distal Left pop.
 
You must log in or register to reply here.
Back
Top