Gedmatch calculators on Vahaduo

**Zanzibar** · 10-22-2021, 05:36 PM

Originally Posted by Komintasavalta

I don't know if Udmurts have that much steppe in absolute terms, but for example in the qpGraph model in Tambets et al. 2018, Udmurts had 30% Nganasan, 63% Corded Ware, 1% LBK, and 6% EHG, so the proportion of Corded Ware in their non-Nganasan ancestry was around 90%. The proportion of Corded Ware was 62% in Central Russians and 61% in Mordovians, but both obviously had lower Nganasan.

Also in the G25 model below, out of the ancestry of Udmurts that is not from Krasnoyarsk_BA, about 82% is from Yamnaya (but the model doesn't have an EHG source, and maybe Udmurts have extra EHG that is not from Yamnaya):

Target: Udmurt
Sources: GEO_CHG IRN_Ganj_Dareh_N RUS_Krasnoyarsk_BA TUR_Barcin_N Yamnaya_RUS_Samara
Distance: 0.045
62% Yamnaya_RUS_Samara
24% RUS_Krasnoyarsk_BA
14% TUR_Barcin_N

Other models for Udmurts in G25:

Target: Udmurt
Distance: 3.9071% / 0.03907110
57.6 Yamnaya_RUS_Samara
25.6 RUS_Krasnoyarsk_BA
11.8 TUR_Barcin_N
5.0 WHG

Target: Udmurt
Distance: 3.3185% / 0.03318480
45.6 Yamnaya_RUS_Samara
26.0 RUS_Krasnoyarsk_BA
19.4 Baltic_EST_BA
9.0 TUR_Barcin_N

Target: Udmurt
Distance: 3.3774% / 0.03377407
37.2 Yamnaya_RUS_Samara
23.6 RUS_Krasnoyarsk_BA
21.6 RUS_Volga-Kama_N
9.2 TUR_Barcin_N
7.6 TUR_Tepecik_Ciftlik_N
0.8 Han

**Leto** · 10-22-2021, 05:44 PM

Originally Posted by Joqool

Other models for Udmurt:

Target: Udmurt
Distance: 3.9071% / 0.03907110
57.6 Yamnaya_RUS_Samara
25.6 RUS_Krasnoyarsk_BA
11.8 TUR_Barcin_N
5.0 WHG

Target: Udmurt
Distance: 3.3185% / 0.03318480
45.6 Yamnaya_RUS_Samara
26.0 RUS_Krasnoyarsk_BA
19.4 Baltic_EST_BA
9.0 TUR_Barcin_N

How much Mongoloid is that Krasnoyarsk sample? It sure isn't Andoronvo since those were pretty much whites. Why not use Ymiyakhtakh Yakutia instead?

**~~Komintasavalta~~** · 10-22-2021, 05:48 PM

Originally Posted by Joqool

Other models for Udmurts in G25:

Target: Udmurt
Distance: 3.9071% / 0.03907110
57.6 Yamnaya_RUS_Samara
25.6 RUS_Krasnoyarsk_BA
11.8 TUR_Barcin_N
5.0 WHG

Target: Udmurt
Distance: 3.3185% / 0.03318480
45.6 Yamnaya_RUS_Samara
26.0 RUS_Krasnoyarsk_BA
19.4 Baltic_EST_BA
9.0 TUR_Barcin_N

Target: Udmurt
Distance: 3.3774% / 0.03377407
37.2 Yamnaya_RUS_Samara
23.6 RUS_Krasnoyarsk_BA
21.6 RUS_Volga-Kama_N
9.2 TUR_Barcin_N
7.6 TUR_Tepecik_Ciftlik_N
0.8 Han

Oh sorry, I had WHG in my sources at first, but I forgot to add it back. Here's the models I ran:

(I use a modified version of Michał's CVXPY script, and I have shell functions for quickly selecting lines from G25 datasheets.)

Originally Posted by Leto

Why not use Ymiyakhtakh Yakutia instead?

Krasnoyarsk_BA is closer to Nganasans. This is a PLINK PCA based on the Reich dataset, where some of the samples from KolymaRiver_LN and CentralYakutia_LN are called RUS_Yakutia_Ymyiakhtakh_LN in G25:

**Zanzibar** · 10-22-2021, 05:52 PM

Originally Posted by Leto

How much Mongoloid is that Krasnoyarsk sample? It sure isn't Andoronvo since those were pretty much whites. Why not use Ymiyakhtakh Yakutia instead?

Like 100% Mongoloid it seems:

Target: RUS_Krasnoyarsk_BA
Distance: 6.5864% / 0.06586387
90.6 Nganassan
8.0 Han
1.2 BRA_LapaDoSanto_9600BP
0.2 Yoruba

Target: RUS_Krasnoyarsk_BA
Distance: 6.2975% / 0.06297547
81.2 Nganassan
18.4 MNG_North_N
0.4 BRA_LapaDoSanto_9600BP

Here are the models with Ymiyakhtakh, fit worsens it seems.

Target: Udmurt
Distance: 4.8925% / 0.04892541
57.4 Yamnaya_RUS_Samara
25.6 RUS_Yakutia_Ymyiakhtakh_LN
11.8 TUR_Barcin_N
5.2 WHG

Target: Udmurt
Distance: 4.3376% / 0.04337573
44.6 Yamnaya_RUS_Samara
26.0 RUS_Yakutia_Ymyiakhtakh_LN
20.6 Baltic_EST_BA
8.8 TUR_Barcin_N

**Leto** · 10-22-2021, 06:00 PM

Originally Posted by Joqool

Like 100% Mongoloid it seems:

Target: RUS_Krasnoyarsk_BA
Distance: 6.5864% / 0.06586387
90.6 Nganassan
8.0 Han
1.2 BRA_LapaDoSanto_9600BP
0.2 Yoruba

Target: RUS_Krasnoyarsk_BA
Distance: 6.2975% / 0.06297547
81.2 Nganassan
18.4 MNG_North_N
0.4 BRA_LapaDoSanto_9600BP

Here are the models with Ymiyakhtakh, fit worsens it seems.

Target: Udmurt
Distance: 4.8925% / 0.04892541
57.4 Yamnaya_RUS_Samara
25.6 RUS_Yakutia_Ymyiakhtakh_LN
11.8 TUR_Barcin_N
5.2 WHG

Target: Udmurt
Distance: 4.3376% / 0.04337573
44.6 Yamnaya_RUS_Samara
26.0 RUS_Yakutia_Ymyiakhtakh_LN
20.6 Baltic_EST_BA
8.8 TUR_Barcin_N

Interesting. Looks like it was some kind of indigenous inhabitant of Siberia, unlike the Andronovans. I wonder if modern Siberian Turkics/Altaians can be understood as a mix of the two. They are heavily R1a-Z93 but autosomally 60-75% East Eurasian depending on the group.

- - -

By the way, why can't anyone convert the Yunusbayev data to G25 format? I emailed David weeks ago - no response. On Anthrogenica no one has replied to my thread either. Is it that difficult? I need the Tajiks and Kumyks from that old-ass study.

**Zanzibar** · 10-22-2021, 06:40 PM

Originally Posted by Komintasavalta

I don't know if Udmurts have that much steppe in absolute terms, but for example in the qpGraph model in Tambets et al. 2018, Udmurts had 30% Nganasan, 63% Corded Ware, 1% LBK, and 6% EHG, so the proportion of Corded Ware in their non-Nganasan ancestry was around 90%. The proportion of Corded Ware was 62% in Central Russians and 61% in Mordovians, but both obviously had lower Nganasan.

Also in the G25 model below, out of the ancestry of Udmurts that is not from Krasnoyarsk_BA, about 82% is from Yamnaya (but the model doesn't have an EHG source, and maybe Udmurts have extra EHG that is not from Yamnaya):

Target: Udmurt
Sources: GEO_CHG IRN_Ganj_Dareh_N RUS_Krasnoyarsk_BA TUR_Barcin_N Yamnaya_RUS_Samara
Distance: 0.045
62% Yamnaya_RUS_Samara
24% RUS_Krasnoyarsk_BA
14% TUR_Barcin_N

Do you know why Udmurts on Tambets using qpAdm are 30% Mongoloid whereas Udmurts on G25 are like 25-26% Mong on average?

Interesting, I heard a rather expert user on Anthrogenica mention that Udmurts also have some Iran_N or extra CHG so I'm surprised they don't score any in G25 or that Tambets qpAdm model.

**Leto** · 10-22-2021, 06:44 PM

Originally Posted by Joqool

Do you know why Udmurts on Tambets using qpAdm are 30% Mongoloid whereas Udmurts on G25 are like 25-26% Mong on average?

Prolly their Siberian reference is not 100% East Eurasian.

Originally Posted by Joqool

Interesting, I heard a rather expert user on Anthrogenica mention that Udmurts also have some Iran_N or extra CHG so I'm surprised they don't score any in G25 or that Tambets qpAdm model.

They score that even on Gedmatch. Their West Asian average is double that of ethnic Russians.

**~~Komintasavalta~~** · 10-22-2021, 08:29 PM

Originally Posted by Joqool

Do you know why Udmurts on Tambets using qpAdm are 30% Mongoloid whereas Udmurts on G25 are like 25-26% Mong on average?

Interesting, I heard a rather expert user on Anthrogenica mention that Udmurts also have some Iran_N or extra CHG so I'm surprised they don't score any in G25 or that Tambets qpAdm model.

It's not qpAdm but qpGraph. And there was no Iran_N or CHG source in their qpGraph model.When I recreated the qpGraph model of Tambets 2018 in ADMIXTOOLS 2, Udmurts only got 26% Nganasan, but it could be because I used different samples:

Code:

library(admixtools)

t=read.table(text="R Yoruba.DG
R A
A O
A AA
O E
O CH
E EU
E L1
EU WH
EU EH
AA EH
EH EH2
EH EH1
EH AAA
AA AAA
WH CCE
WH Italy_North_Villabruna_HG
CH YM
CH CH1
CH AAN
EH1 YM
EH1 EH0
AAA AAN
CH1 EH0
CH1 Georgia_Kotias.SG
L1 Turkey_N_published
L1 LB
L1 YM1
L1 CW
EH2 LB
EH2 CWC
YM YM1
YM CW
EH0 Russia_HG_Karelia
AAN Nganasan
AAN CCA
LB Germany_EN_LBK
LB CCW
YM1 Russia_Samara_EBA_Yamnaya
CW CCW
CW Estonia_CordedWare
CCW CWC
CWC CCE
CCE CCA
CCA Udmurt")

pop=c("Udmurt","Germany_CordedWare","Estonia_CordedWare","Germany_EN_LBK","Italy_North_Villabruna_HG","Nganasan","Russia_HG_Karelia","Russia_Samara_EBA_Yamnaya","Turkey_N_published","Yoruba.DG","Georgia_Kotias.SG","Chimp.REF")

f2=f2_from_geno("v50.0_HO_public",pops=pop)
gr=qpgraph(f2,t)

plot_graph(gr$edges)
ggsave("1.png",width=9,height=9)
system("mogrify -trim -border 64 -bordercolor white t/1.png")

# plotly_graph(gr$edges)

write_dot(gr$edges) # generate black-and-white graph at graphviz.it

Edit: But when I used `maxmiss=1`, Udmurts got 32% Nganasan. (By default, ADMIXTOOLS 2 discards each SNP which is missing from every sample in a population, but `maxmiss=1` discards no SNPs.)

**Lucas** · 10-22-2021, 08:31 PM

Originally Posted by Komintasavalta

It's not qpAdm but qpGraph. The results of qpGraph depend a lot on what outgroups you include, just like with qpAdm, but it also depends on the order of the nodes and what intermediate nodes you include. And there was no Iran_N or CHG source in their qpGraph model.

When I recreated the qpGraph model of Tambets 2018 in ADMIXTOOLS 2, Udmurts only got 26% Nganasan, but it could be because I used different samples:

Are you even tried qpGraph only with modern samples? I want to see such model, for some VURs could be.

**Zanzibar** · 10-22-2021, 09:03 PM

Originally Posted by Komintasavalta

It's not qpAdm but qpGraph. And there was no Iran_N or CHG source in their qpGraph model.When I recreated the qpGraph model of Tambets 2018 in ADMIXTOOLS 2, Udmurts only got 26% Nganasan, but it could be because I used different samples:

Code:

library(admixtools)

t=read.table(text="R Yoruba.DG
R A
A O
A AA
O E
O CH
E EU
E L1
EU WH
EU EH
AA EH
EH EH2
EH EH1
EH AAA
AA AAA
WH CCE
WH Italy_North_Villabruna_HG
CH YM
CH CH1
CH AAN
EH1 YM
EH1 EH0
AAA AAN
CH1 EH0
CH1 Georgia_Kotias.SG
L1 Turkey_N_published
L1 LB
L1 YM1
L1 CW
EH2 LB
EH2 CWC
YM YM1
YM CW
EH0 Russia_HG_Karelia
AAN Nganasan
AAN CCA
LB Germany_EN_LBK
LB CCW
YM1 Russia_Samara_EBA_Yamnaya
CW CCW
CW Estonia_CordedWare
CCW CWC
CWC CCE
CCE CCA
CCA Udmurt")

pop=c("Udmurt","Germany_CordedWare","Estonia_CordedWare","Germany_EN_LBK","Italy_North_Villabruna_HG","Nganasan","Russia_HG_Karelia","Russia_Samara_EBA_Yamnaya","Turkey_N_published","Yoruba.DG","Georgia_Kotias.SG","Chimp.REF")

f2=f2_from_geno("v50.0_HO_public",pops=pop)
gr=qpgraph(f2,t)

plot_graph(gr$edges)
ggsave("1.png",width=9,height=9)
system("mogrify -trim -border 64 -bordercolor white t/1.png")

# plotly_graph(gr$edges)

write_dot(gr$edges) # generate black-and-white graph at graphviz.it

Edit: But when I used `maxmiss=1`, Udmurts got 32% Nganasan. (By default, ADMIXTOOLS 2 discards each SNP which is missing from every sample in a population, but `maxmiss=1` discards no SNPs.)

Well here is qpAdm model of Udmurts by David showing them close to 30% Mongoloid: https://eurogenes.blogspot.com/2017/...lithic-to.html

Udmurt
chisq 7.095
tail prob 0.419052
Barcin_Neolithic 0.234
Caucasus_HG 0.133
Eastern_HG 0.298
Nganasan 0.295
Onge 0.01
Western_HG 0.028
Yoruba 0

Or this data, showing Udmurts around 30% Eastern Eurasian: https://genomebiology.biomedcentral....22-1/figures/6

These two models from this Anthrogenica thread: https://anthrogenica.com/showthread....lic-qpAdm-runs

Udmurt
Nganasan: 0.307±0.007
Estonia_MN_CCC: 0.150±0.033
Latvia_LN_CordedWare: 0.387±0.045
Ukraine_Globular_Amphora: 0.155±0.029
tail: 0.099657
chisq: 15.999

Udmurt
Nganasan: 0.287±0.009
Estonia_MN_CCC: 0.176±0.045
Germany_CordedWare: 0.450±0.068
Hungary_LBK_MN.SG: 0.088±0.027
tail: 0.659253
chisq: 7.688