my updated non overfit G25 model

[Kurdquistador]

vahaduo : https://vahaduo.github.io/vahaduo/

here the model fuckers : https://justpaste.it/7sdcv


some notes regarding this model :

- you will get a little worse fits with this model than you are used to with many other models . simply because i have used references that have as little overfit with each other as possible .

- i have used an Iran Ganj Dareh N individual that is the least similar to CHG . therefore you will score more CHG here . i think it is hidden CHG that wont show in many other models due to the fact that Iran N eats it up usually


my results

 

[idklolimo]

 

[Kurdquistador]

View attachment 13599

bro please tell me what was your arab ancestor like ?

the Kurdish reference scores 8.8 Arabian hunter gatherer while you score 14.4

you said you are 1/8 arab .

if all 7/8 of your Kurdish ancestors were 8.8 Arabian HG then your arab ancestor must have been gulf arab level with 55% or so arabian hg

 

[idklolimo]

bro please tell me what was your arab ancestor like ?

the Kurdish reference scores 8.8 Arabian hunter gatherer while you score 14.4

you said you are 1/8 arab .

if all 7/8 of your Kurdish ancestors were 8.8 Arabian HG then your arab ancestor must have been gulf arab level with 55% or so arabian hg

Maybe my father's side may be completely Mihallami/Islamized Assyrian, but this is not possible because there are no Arabic or Semitic speaking people on my father's side lol

 

[Kurdquistador]

Maybe my father's side may be completely Mihallami/Islamized Assyrian, but this is not possible because there are no Arabic or Semitic speaking people on my father's side lol

i mean your 1/8 arab ancestor . he/she must have been gulf arab like

 

[Kurdquistador]

you are still northern west asian . 14.4 is not as much as kicking you out of it . but i wonder about your arab ancestor tbh

 

[idklolimo]

you are still northern west asian . 14.4 is not as much as kicking you out of it . but i wonder about your arab ancestor tbh

I think g25 coordinates and results can be misleading. Gedmatch also gives conflicting results. I think this is more related to the quality of raw data.

 

[Kurdquistador]

I think g25 coordinates and results can be misleading. Gedmatch also gives conflicting results. I think this is more related to the quality of raw data.

learn from me . i personally take anything giving information from every tool

s-indian from harappaworld (also south asian from eurogenes k12b and dodecad k12b and south indian from mdlp k18 ancient roots) to see if someone is gypsoid

red sea from dodecad k10a to see if someone is sandniggeroid lol

west-central asian from Eu K12b to see if someone from Turkey is Kurdish

Atlantic Baltic from Dodecad K10a to see north euro stuff

and many more such things

 

[Kurdquistador]

i have ran many ethnicities and this is the best model i have made so far . it makes a lot of sense . check yourself . run people

 

[Kurdquistador]

i updated it again . have used an tepe abdul hossein individual for Iran N

try it

the model : https://justpaste.it/fzdfu

vahaduo :

Vahaduo: Admixture JS 22.05

Vahaduo: Admixture JS - Rapid estimation of admixture proportions and calculation of Euclidean distances.

vahaduo.github.io

my result :

 

[Ben Dover]

i updated it again . have used an tepe abdul hossein individual for Iran N

try it

the model : https://justpaste.it/fzdfu

vahaduo :

Vahaduo: Admixture JS 22.05

Vahaduo: Admixture JS - Rapid estimation of admixture proportions and calculation of Euclidean distances.

vahaduo.github.io

my result :

exaggerated sand for me but otherwise it is decent

 

[Ben Dover]

 

[Kurdquistador]

you can also substitute Kotias with Satsurblia :

Satsurblia_CHG,0.092197,0.101553,-0.093526,-0.000969,-0.092633,0.020917,0.030786,-0.001615,-0.139281,-0.085833,-0.002923,0.024278,-0.058424,0.009634,0.036373,-0.022938,0.044591,-0.008488,-0.027025,0.042896,0.04517,-0.009521,0.001849,-0.030245,-0.002515

 

[Kurdquistador]

allright this is updated again and i like it the most . the fits will be even "worse" but that is because as the title says as little overlapping / components eating up each other as possible

here you go : https://justpaste.it/628mu

vahaduo : https://vahaduo.github.io/vahaduo/


my results :

 

[KARABOĞA]