Gedmatch calculators on Vahaduo

**Leto** · 10-21-2020, 01:15 PM

Originally Posted by Amon

This avg.'s Gedrosia and Mediterranean is kinda low and SW Asian is a lil higher. Gedrosia should have been around 1.5, Mediterranean around 6-6.5 and SW Asian around 50.

Yeah, dat's what I was thinkin' as well.

But I got no data on them.

**~~Maguzanci~~** · 10-21-2020, 01:38 PM

Originally Posted by Leto

Cool. And the average?

Here is the average

Code:

Western Indonesian: 3.93,0.49,0.21,62.90,0.12,0.99,16.12,0.22,0.35,13.67,0.92,0.08

Distance to: Western_Indonesian
10.35946910 PANIYA (these samples are mislabelled, they seem to be SE Asian and not South Asian)
19.20949245 Lahu
20.78414299 Vietnamese
22.07255083 Malay_Singapore
23.99380337 Cambodians
26.36972886 Dai
31.17905547 Khasi
33.95818900 Burmanese
36.96647941 Garo
41.41437673 Miaozu
43.89232963 She

Here is my model of the Western_Indonesian average: Bonda (an Austroasiatic tribe from Eastern India) represented the South Asian ancestry in Indonesians as they are mostly AASI. Didn't use the mainstream South Asians because it could hide the West Eurasian admix of the Indoensian samples. Brahui+Iranian Mazandaran+Lithuanian is utilized to represent the West Eurasian admix in Indonesians (seems mostly Iran_N related due to Brahui with minor Steppe affinity represented by Lithuanian). Shows noise amounts of West African (Yoruba). Dai+Vietnamese is used to represent the predominant Mongoloid ancestry of Indos. Surprisingly shows minor Chinese-like admix (Han_Central) but could actually be an Taiwan_Aborigine/Austronesian signal instead as there are no Austronesian samples in this calculator. Sorry for the lenghty sentences; just my analysis.

Target: Western_Indonesian
Distance: 0.7471% / 0.74706758
63.4 Dai
23.0 Bonda
5.8 Han_central
4.6 Brahui
2.0 Iran_Mazandaran
1.2 Lithuanian

I will later do some samples from Sulawesi and Eastern Indonesia (Maluku, Sumba). It will be difficult for the Eastern Indonesian samples though as this calculator lacks the Oceanian/Papuan component (Eastern Indos have very high Oceanian), therefore the results might be misleading. But if you want them also, I can do it.

**~~Maguzanci~~** · 10-21-2020, 01:39 PM

Originally Posted by Lucas

All updates from last few pages added.

Can you edited the Western_Indonesian average please? I decided to add one more sample.

Code:

Western_Indonesian,3.93,0.49,0.21,62.90,0.12,0.99,16.12,0.22,0.35,13.67,0.92,0.08

**Leto** · 10-21-2020, 01:54 PM

Originally Posted by Maguzanci

Here is the average

Code:

Western Indonesian: 3.93,0.49,0.21,62.90,0.12,0.99,16.12,0.22,0.35,13.67,0.92,0.08

Closest pops
Distance to: Western_Indonesian
10.35946910 PANIYA (looks like it is mislabelled. Its actually a SE Asian population, not South Asian).
19.20949245 Lahu
20.78414299 KHV30
22.07255083 MAS30
23.99380337 Cambodians
25.38024035 CDX30
26.36972886 Dai
31.17905547 Khasi
33.95818900 Burmanese
36.96647941 Garo
41.41437673 Miaozu

Here is my model of the Western_Indonesian average: Bonda (an Austroasiatic tribe from Eastern India) represented the South Asian ancestry in Indonesians as they are mostly AASI. Didn't use the mainstream South Asians because it could hide the West Eurasian admix of the Indoensian samples. Brahui+Iranian Mazandaran+Lithuanian is utilized to represent the West Eurasian admix in Indonesians (seems mostly Iran_N related due to Brahui with minor Steppe affinity represented by Lithuanian). Shows noise amounts of West African (Yoruba). Dai+Vietnamese is used to represent the predominant Mongoloid ancestry of Indos. Surprisingly shows minor Chinese admix (Han_Central). Sorry for the lenghty sentences; just my analysis.

Target: Western_Indonesian
Distance: 0.7689% / 0.76887318
50.4 Dai
22.4 Bonda
17.0 Vietnamese
4.6 Brahui
2.2 Han_central
2.0 Iran_Mazandaran
1.2 Lithuanian
0.2 Yoruba

I will later do some samples from Sulawesi and Eastern Indonesia (Maluku, Sumba). It will be difficult for the Eastern Indonesian samples though as this calculator lacks the Oceanian/Papuan component (Eastern Indos have very high Oceanian), therefore the results might be misleading. But if you want them also, I can do it.

Well, we probably do not need such a detailed division for Southeast Asia on Dodecad. I'd say 3-4 major groups would be enough. Also make sure to overhaul the existing pops, some of them may be seriously inaccurate. As you pointed out, Dodecad lacks certain components necessary for that region (just like it lacks Amerindian).

**~~Maguzanci~~** · 10-21-2020, 01:59 PM

Originally Posted by Leto

Well, we probably do not need such a detailed division for Southeast Asia on Dodecad. I'd say 3-4 major groups would be enough. Also make sure to overhaul the existing pops, some of them may be seriously inaccurate. As you pointed out, Dodecad lacks certain components necessary for that region (just like it lacks Amerindian).

Alright. What does overhaul mean?

Yup it lack necessary components. Even though it lacks Amerindian, I feel it can gives sort of glimpse of how much East Eurasian and West Eurasian ancestries Native Americans do have.

For example, Anzick, a Paleo-Indian individual from North America. The North Euro+Gedrosia suggest ANE-related West Eurasian ancestry. While the Siberian+East Asian+SE Asian suggests East Eurasian ancestry or even some Onge/Australoid-like affinity (in SE Asian component). Not sure about South Asian but is probably ANE or (Onge-related afifnity) as well.

Admix Results (sorted):

# Population Percent
1 Siberian 33.5
2 East_Asian 29.08
3 North_European 17.79
4 Gedrosia 7.82
5 Southeast_Asian 6.33
6 South_Asian 4.42
7 Sub_Saharan 1.07

But anyway still take it will a grain of salt.

**Lucas** · 10-21-2020, 02:36 PM

@MAzgunaci
Done

=-==========

I found on my computer some Polish averages for K23b based on Gedmatch kits (made by Michał from Anthrogenica).
Why not to add them to http://vahaduo.genetics.ovh/mdlp_k23...ed-vahaduo.htm

Masovia - 18 kits
Kashubian - 8 kits
South Poland - 24 kits

Now my results are better:

Target: Luk
Distance: 1.5813% / 1.58130667 | ADC: 0.25x
70.0 PL_Masovian
21.2 Russian_Meshtchyora
6.6 PL_Kashubian
2.2 BedouinB

Target: Luk
Distance: 1.1083% / 1.10831508
65.2 PL_Masovian
13.6 Latvian
13.6 PL_Kashubian
3.0 BedouinB
2.2 Dargin_Urkarah
1.4 Udmurd
0.8 Frisian
0.2 Pygmy

without Polish averages it was disaster

Target: Luk
Distance: 1.9697% / 1.96966173 | ADC: 0.25x
46.2 Kashub (original, rather strange reference)
25.0 Ukrainian_Center
14.8 Russian-West
11.2 Russian-Upper-Volga
2.0 BedouinB
0.8 Russian_Meshtchyora

Target: Luk
Distance: 1.5198% / 1.51977394
39.0 Latvian
28.8 Ukrainian_Center
15.6 Frisian
7.2 Pole
4.2 Greek_Phokaia
2.0 BedouinB
1.8 Udmurd
1.2 Georgian_Svan
0.2 Pygmy

**Leto** · 10-21-2020, 02:50 PM

Originally Posted by Maguzanci

Alright. What does overhaul mean?

To revise, to revisit. I mean check if they are correct or realistic. We've had to delete many of them because they were completely impossible to be like real people. The calculator effect.

**~~Maguzanci~~** · 10-21-2020, 04:07 PM

Originally Posted by Leto

To revise, to revisit. I mean check if they are correct or realistic. We've had to delete many of them because they were completely impossible to be like real people. The calculator effect.

Ok I can check and scrutinize them. But I cannot delete them though. I will have to tell you and Lukasz?

Forget to ask when making averages, its not good to use family members or relatives as the samples right?

**Leto** · 10-21-2020, 04:11 PM

Originally Posted by Maguzanci

Ok I can check and scrutinize them. But I cannot delete them though. I will have to tell you and Lukasz?

Forget to ask when making averages, its not good to use family members or relatives as the samples right?

Yeah, just leave a message on this thread or submit the averages to him directly.

I would avoid immediate family members like parents/children/siblings. More distant ones are okay.