there is a Iran N individual that is more like CHG than all the other Iran N are . and this Iran N works best for me
when i use this Iran N individual and Barcin N or Balkan N it is most of my genome .
and when i use only these 2 and add EHG then i am almost done . not much more is needed . and do not bitch about the distance of 3.5 . i never get much better fits ....and this fit is good enough . many europeans for example can not be modelled with a fit better than 4-5 . try to model latvians , lithuanians etc. with a fit better than 6 ...you wont be able to do it . when i make models i try to take components that eat up each other as little as possible . what many of you call "gOoD FiTs " are actually often overfits because of components eating up each other in the model etc.
i am using every relevant population . this works also well with ANF like Barcin N . but Greece N gives me better fits
let me show you :
first of all here is the model for you if you want to run yourself : https://justpaste.it/k4oqd
vahaduo link : https://vahaduo.github.io/vahaduo/
this is how i score with the full model
but i can easily be modelled with only 3 populations with almost the same fit .
look at the distances i am to Iran N and Greece N . distant to both . so the fit of 3-4 is very good regarding that
when i use this Iran N individual and Barcin N or Balkan N it is most of my genome .
and when i use only these 2 and add EHG then i am almost done . not much more is needed . and do not bitch about the distance of 3.5 . i never get much better fits ....and this fit is good enough . many europeans for example can not be modelled with a fit better than 4-5 . try to model latvians , lithuanians etc. with a fit better than 6 ...you wont be able to do it . when i make models i try to take components that eat up each other as little as possible . what many of you call "gOoD FiTs " are actually often overfits because of components eating up each other in the model etc.
i am using every relevant population . this works also well with ANF like Barcin N . but Greece N gives me better fits
let me show you :
first of all here is the model for you if you want to run yourself : https://justpaste.it/k4oqd
vahaduo link : https://vahaduo.github.io/vahaduo/
this is how i score with the full model
but i can easily be modelled with only 3 populations with almost the same fit .
look at the distances i am to Iran N and Greece N . distant to both . so the fit of 3-4 is very good regarding that
Last edited:
