Europeans have more Iran N related than CHG related ancestry (qpAdm models)

[IjustlikeHistory]

Europeans seem to have more Iran N related ancestry than Caucasus Hunter Gatherer related. I do not know if they actually have Iran N or some ancestral population to Iran N (Iran Hunter Gatherers?). In this model the Polish even score almost only Iran N related.

Corded Ware and Norwegians
unknown.png
Polish
unknown.png

Right pops:

right = c('Mbuti.DG', 'Ethiopia_4500BP_published.SG', 'Russia_Ust_Ishim.DG', 'Czech_Vestonice16', 'Belgium_UP_GoyetQ116_1_published', 'Russia_Kostenki14.SG', 'Russia_MA1_HG.SG', 'Italy_North_Villabruna_HG', 'Han.DG', 'Papuan.DG', 'ONG.SG',  'Karitiana.DG', 'Georgia_Satsurblia.SG', 'Turkey_Epipaleolithic', 'Morocco_Iberomaurusian', 'Israel_Natufian_published')

To make sure i was not wrong i've tried multiple right pops from papers, tried different combinations and even tried right pops by Davidski and i still got the same results.
Models made using Rstudio AdmixTools 2

People that know how to use qpAdm send models using the same left populations for Europeans.

Post models of modern Europeans with CHG/Iran N neolithic and not just models of Eneolithic people which do not represent fully modern Europeans. If you cannot model the Polish with just CHG or if the Polish can model better with Iran N than CHG then it means a migration happened AFTER the Eneolithic. If you post Eneolithic models then at least also post modern Europeans with similar proxies and experiment with Iran N too.

[Kale]

On a deep scale, CHG behave like a bottlenecked mix of Iran, Anatolian, and EHG populations.

CHG.SG
Iran_TepeAbdulHosein_N.SG 0.550973 0.0223940 24.6036
Anatolia_Epipaleolithic 0.338898 0.0284548 11.9100
RUS_Karelia_HG 0.110129 0.0187272 5.88069
Tail: 0.42
right = c('Congo_Mbuti.DG', 'Taforalt', 'Iran_GanjDareh_N', 'Onge.DG', 'Japan_Jomon', 'RUS_Primorsky_Boisman_MN', 'Peru_Laramate_900BP', 'Tarim_EMBA1', 'MA1.SG', 'Yana_UP.SG', 'Gravettian', 'Sunghir.SG', 'BachoKiro_IUP', 'Ust_Ishim.DG')
snps: 431182

The CHG ancestry present in Europeans branched off before Kotias and Satsurblia split from each other.
So with the bottleneck partially absent, if you have Kotias in the left and Satsurblia in the right, European CHG largely breaks apart into it's Iran, Anatolian, and EHG-like constituents.

CHG will work fine with Iran_N in the right pops.

RUS_Stavropol_Progress_Eneolithic
RUS_Arkhangelsk_HG.SG 0.336807 0.0446366
AG3 0.162081 0.0326854
CHG.SG 0.501112 0.0256997
Tail: 0.18
right = c('Congo_Mbuti.DG', 'Anatolia_Boncuklu_N.SG', 'Iran_Wezmeh_N.SG', 'Taforalt', 'Italy_GrottaContinenza_HG.SG', 'Botai.SG', 'Yana_UP.SG', 'Andaman_100BP.SG', 'RUS_Primorsky_DevilsCave_N.SG', 'Peru_RioUncallane_1800BP.SG')
Snps: 240234

RUS_Stavropol_Vonjucka_Eneolithic
RUS_Arkhangelsk_HG.SG 0.293275 0.0549543
AG3 0.148050 0.0396730
CHG.SG 0.558675 0.0324585
Tail: 0.40
right = c('Congo_Mbuti.DG', 'Anatolia_Boncuklu_N.SG', 'Iran_Wezmeh_N.SG', 'Taforalt', 'Italy_GrottaContinenza_HG.SG', 'Botai.SG', 'Yana_UP.SG', 'Andaman_100BP.SG', 'RUS_Primorsky_DevilsCave_N.SG', 'Peru_RioUncallane_1800BP.SG')
Snps: 219114

But it's a hard fail with Iran in the left and CHG in the right.
RUS_Stavropol_Vonjucka_Eneolithic
RUS_Arkhangelsk_HG.SG 0.407134 0.0556593
AG3 0.0663339 0.0433884
Iran_GanjDareh_N 0.526532 0.0297400
Tail: 0.0000002
right = c('Congo_Mbuti.DG', 'Anatolia_Boncuklu_N.SG', 'CHG.SG', 'Taforalt', 'Italy_GrottaContinenza_HG.SG', 'Botai.SG', 'Yana_UP.SG', 'Andaman_100BP.SG', 'RUS_Primorsky_DevilsCave_N.SG', 'Peru_RioUncallane_1800BP.SG')
Snps: 218316

[IjustlikeHistory]

On a deep scale, CHG behave like a bottlenecked mix of Iran, Anatolian, and EHG populations.

But your models did not include modern Europeans or Corded Ware. Try modeling modern Europeans with (Iran N), (CHG), (Iran N + CHG). There are a lot of ghost populations that contributed to modern Europeans and showing models of Eneolithic people when many other migrations took place after that won't show anything. The Iran N like shift that happened was probably after the Eneolithic so Eneolithic is a bit irrelevant.

I'm not saying that Iran N was ancestral to modern Europeans but that they have something more related to Iran N than CHG. That's my point.

[bce]

Europeans seem to have more Iran N related ancestry than Caucasus Hunter Gatherer related. I do not know if they actually have Iran N or some ancestral population to Iran N (Iran Hunter Gatherers?). In this model the Polish even score almost only Iran N related.

Corded Ware and Norwegians
unknown.png
Polish
unknown.png

Right pops:

right = c('Mbuti.DG', 'Ethiopia_4500BP_published.SG', 'Russia_Ust_Ishim.DG', 'Czech_Vestonice16', 'Belgium_UP_GoyetQ116_1_published', 'Russia_Kostenki14.SG', 'Russia_MA1_HG.SG', 'Italy_North_Villabruna_HG', 'Han.DG', 'Papuan.DG', 'ONG.SG',  'Karitiana.DG', 'Georgia_Satsurblia.SG', 'Turkey_Epipaleolithic', 'Morocco_Iberomaurusian', 'Israel_Natufian_published')

To make sure i was not wrong i've tried multiple right pops from papers, tried different combinations and even tried right pops by Davidski and i still got the same results.
Models made using Rstudio AdmixTools 2

People that know how to use qpAdm send models using the same left populations for Europeans.

are you using the "maxmiss=1" option?

[IjustlikeHistory]

are you using the "maxmiss=1" option?

No, i'm not. Why?

[bce]

No, i'm not. Why?

maxmiss=1 means no SNPs get filtered. results can change a lot when all 1240k SNPs are used.

[IjustlikeHistory]

maxmiss=1 means no SNPs get filtered. results can change a lot when all 1240k SNPs are used.

Alright, thanks. I will try that out.

[Ariel90]

I don't see the usefulness of these "let's re-write scientific papers written by the leading scientist of their field" posts.

[IjustlikeHistory]

maxmiss=1 means no SNPs get filtered. results can change a lot when all 1240k SNPs are used.

I've tried maxmiss = 1 and the results are almost identical.

Polish cannot be modeled with just CHG but they can be modeled with just Iran N.
unknown.png

[IjustlikeHistory]

I don't see the usefulness of these "let's re-write scientific papers written by the leading scientist of their field" posts.

Have any papers tried to model modern Europeans with Iran N and it failed? I would like to see the papers trying models of modern Europeans with CHG, Iran N and CHG + Iran N. I don't see how this is a re-writting if there are no such models attempted. Even if papers have shown the opposite then i don't see how it's bad to question them if we think they might've got something wrong. If i missed papers that attempted such qpAdm models then send them.