PDA

View Full Version : More Punjabi Gujjar Results for David's Fst 4 Caulcator



Xehanort
01-15-2018, 05:40 AM
Here are a bit more models David recently did for me:

Andronovo:

First model:

Andronovo: 34%
Iran Neolithic: 45%
Lebbo: 20%

Second model:

Andronovo: 37%
Iran Neolithic: 43%
Onge: 21%

Poltavka:

First model:

Iran Neolithic: 43%
Lebbo: 20%
Poltavka: 37%

Second model (Best fit from all runs):

Iran Neolithic: 40%
Onge: 20%
Poltavka: 40%

Sarmatian:

First model

Iran Neolithic: 41%
Lebbo: 20%
Sarmatian: 39%

Second model

Iran Neolithic: 40%
Onge: 20%
Sarmatian: 40%

Any thoughts?

From this, it seems that I have around 37 to 40% Steppe ancestry, 40 to 45% Neolithic Iran ancestry, and 20 to 21% Lebbo/Onge ancestry. The Onge would have probably been lower had David included Han as an additional proxy in the runs.

How typical are these results for upper caste South Asians?

Based on this, I think it can be concluded that around 70% of my ancestry is derived from an ANI/Vedic Aryan/Iron Age Steppe source, and 30% derived from the Indus Valley people, but hopefully the upcoming study on Vedic Aryans and the Indus Valley Civilization peoples will confirm this to be the case.

Have any other South Asians tested with Davidski? I would appreciate it if you could post your results.

bmoney
01-15-2018, 05:50 AM
Here are a bit more models David recently did for me:

Andronovo:

First model:

Andronovo: 34%
Iran Neolithic: 45%
Lebbo: 20%

Second model:

Andronovo: 37%
Iran Neolithic: 43%
Onge: 21%

Poltavka:

First model:

Iran Neolithic: 43%
Lebbo: 20%
Poltavka: 37%

Second model (Best fit from all runs):

Iran Neolithic: 40%
Onge: 20%
Poltavka: 40%

Sarmatian:

First model

Iran Neolithic: 41%
Lebbo: 20%
Sarmatian: 39%

Second model

Iran Neolithic: 40%
Onge: 20%
Sarmatian: 40%

Any thoughts?

From this, it seems that I have around 37 to 40% Steppe ancestry, 40 to 45% Neolithic Iran ancestry, and 20 to 21% Lebbo/Onge ancestry. The Onge would have probably been lower had David included Han as an additional proxy in the runs.

How typical are these results for upper caste South Asians?

Based on this, I think it can be concluded that around 70% of my ancestry is derived from an ANI/Vedic Aryan/Iron Age Steppe source, and 30% derived from the Indus Valley people, but hopefully the upcoming study on Vedic Aryans and the Indus Valley Civilization peoples will confirm this to be the case.

Have any other South Asians tested with Davidski? I would appreciate it if you could post your results.

Onge would be a bad fit to ghost-ASI but I guess ASI should leak into Onge in a 3 way model as its further from ANI and Vedic Aryans than the Onge so largely works

Lebbo, interesting fair enough its consistent with Onge. Is he planning on uploading his calc to gedmatch? I'd like to run it for myself

Also where did you get 30% IVC, it would be 40% using Iran_Neolithic or potentially 60% if all of your ASI came via IVC ancestry - IVC people could be mixed Iran_N and ASI

poi
01-15-2018, 07:01 AM
I paid for it a while back in November('wired money' to some guy in Australia... never done it before)... got files for Yamnaya-Dai-Iran_N and Sarmatian-Lebbo-Iran_N.

I can dig up the files from the email, but I vividly remember the Sarmatian was almost 50%(Lebbo and Iran_N 25% each) and Yamnaya was a bit under 40%(Iran was mid-30% and Dai was 25%). My data was 23andme v5, although I didn't ask him for the genotype rate. I will likely wait for aDNA from S/C Asia before going through this again. :biggrin1:

Xehanort
01-15-2018, 07:18 AM
Onge would be a bad fit to ghost-ASI but I guess ASI should leak into Onge in a 3 way model as its further from ANI and Vedic Aryans than the Onge so largely works

Lebbo, interesting fair enough its consistent with Onge. Is he planning on uploading his calc to gedmatch? I'd like to run it for myself

Also where did you get 30% IVC, it would be 40% using Iran_Neolithic or potentially 60% if all of your ASI came via IVC ancestry - IVC people could be mixed Iran_N and ASI

Oh, I just assumed that IVC was mostly Onge-like by the late phase, with a descent portion being Iran Neolithic, and that the Vedic Aryans themselves were mostly Steppe plus some additional Neolithic Iranian, and I think 65 and 35 is a better estimate, IMO. I don't think all of the Iran_Neolithic came from ASI.

Xehanort
01-15-2018, 07:22 AM
I paid for it a while back in November('wired money' to some guy in Australia... never done it before)... got files for Yamnaya-Dai-Iran_N and Sarmatian-Lebbo-Iran_N.

I can dig up the files from the email, but I vividly remember the Sarmatian was almost 50%(Lebbo and Iran_N 25% each) and Yamnaya was a bit under 40%(Iran was mid-30% and Dai was 25%). My data was 23andme v5, although I didn't ask him for the genotype rate. I will likely wait for aDNA from S/C Asia before going through this again. :biggrin1:

Interesting, so you seem to have more Sarmatian than me, but our Yamnaya is around the same. When David modeled me on Yamnaya I got around 36% Steppe, 45% Iran Neolithic, and 18% Dai.

As you can see, I have less Lebbo ancestry than yourself, but less Steppe and more Iran Neolithic. This complicate matters if we are to believe that the Vedic Aryans were partially Neolithic Iranian when they entered India. Thoughts?

bmoney
01-15-2018, 08:34 AM
Oh, I just assumed that IVC was mostly Onge-like by the late phase, with a descent portion being Iran Neolithic, and that the Vedic Aryans themselves were mostly Steppe plus some additional Neolithic Iranian, and I think 65 and 35 is a better estimate, IMO. I don't think all of the Iran_Neolithic came from ASI.

I didn't say any of it came from the ASI, i'm saying the IVC was majority Iran_N or even Iran_N and ASI - so you should use Iran_N as the best proxy for IVC

bmoney
01-15-2018, 08:39 AM
Interesting, so you seem to have more Sarmatian than me, but our Yamnaya is around the same. When David modeled me on Yamnaya I got around 36% Steppe, 45% Iran Neolithic, and 18% Dai.

As you can see, I have less Lebbo ancestry than yourself, but less Steppe and more Iran Neolithic. This complicate matters if we are to believe that the Vedic Aryans were partially Neolithic Iranian when they entered India. Thoughts?

Why doesn't it make sense - Iran_N % of the Sintashta steppe herders wouldn't be very high, they had the ancestry via CHG - your high Iran_N is mostly due to your proximity to the IVC, the admixture drops the further east and south you go from Baluchistan/Mehrgarh/Sindh

High R1a pops like Nepali Brahmins would be expected to have high steppe/Sarmatian/Scythian due to founder effects

Compare that to Baloch/Brahui/Makrani who have low steppe levels

pegasus
01-15-2018, 08:56 AM
I didn't say any of it came from the ASI, i'm saying the IVC was majority Iran_N or even Iran_N and ASI - so you should use Iran_N as the best proxy for IVC

I strongly feel urban IVC people will be a lot like the pan South Asian type you see in many groups, from the Patels, Punjabi Chamars, many South Indians. Also my hunch tells me groups like the Toda of Kerala, would likely be direct IVC isolates, their unique and fascinating culture stands out, no genome data on them surprisingly. They are kind of like the South Indian version of the Kalash IMO.

bmoney
01-15-2018, 10:12 AM
I strongly feel urban IVC people will be a lot like the pan South Asian type you see in many groups, from the Patels, Punjabi Chamars, many South Indians. Also my hunch tells me groups like the Toda of Kerala, would likely be direct IVC isolates, their unique and fascinating culture stands out, no genome data on them surprisingly. They are kind of like the South Indian version of the Kalash IMO.

Yeah I agree, Gujarati B, C or D would probably be the best proximate for IVC. Only the Baloch/Brahui are still relatively unmixed with ASI which to me suggests that as soon as you touch the Indian basin in Punjab you start coming into contact with ASI which then suggests that the Paleolithic Indians were pretty widespread considering even Pashtuns have the admixture

Never thought of the Toda that way but yes that clicks - they are missing the recent steppe ancestry the Kalash have absorbed however

poi
01-15-2018, 01:23 PM
Why doesn't it make sense - Iran_N % of the Sintashta steppe herders wouldn't be very high, they had the ancestry via CHG - your high Iran_N is mostly due to your proximity to the IVC, the admixture drops the further east and south you go from Baluchistan/Mehrgarh/Sindh

High R1a pops like Nepali Brahmins would be expected to have high steppe/Sarmatian/Scythian due to founder effects

Compare that to Baloch/Brahui/Makrani who have low steppe levels

I understand how founder effect could elevate R1a frequency, but why would autosomal/admixture be affected?

poi
01-15-2018, 01:29 PM
As you can see, I have less Lebbo ancestry than yourself, but less Steppe and more Iran Neolithic. This complicate matters if we are to believe that the Vedic Aryans were partially Neolithic Iranian when they entered India. Thoughts?

I now think, given no aDNA as of yet, that both you and I showng very high Karasuk-Scythian in AncientK12 and high steppe in general could be our steppe ancestry coming from the Indo Scythians. We just have to wait and see if the estimate from the Moorjani paper from 2013 is validated or not by future actual aDNA.

bmoney
01-15-2018, 01:41 PM
I understand how founder effect could elevate R1a frequency, but why would autosomal/admixture be affected?

higher % of ancestry from this pop as compared to others from different paternal lineages as most West Eurasian admixture in SA is male-mediated

poi
01-15-2018, 02:34 PM
higher % of ancestry from this pop as compared to others from different paternal lineages as most West Eurasian admixture in SA is male-mediated

Is that another way of saying that the 'founding' admixture composition was sort of preserved due to the founder effect?

Xehanort
01-15-2018, 09:02 PM
I now think, given no aDNA as of yet, that both you and I showng very high Karasuk-Scythian in AncientK12 and high steppe in general could be our steppe ancestry coming from the Indo Scythians. We just have to wait and see if the estimate from the Moorjani paper from 2013 is validated or not by future actual aDNA.

Yeah, perhaps this could indeed be the case. However, in your case it would be quite interesting as to how Scythian descendants would gain a status of Brahmin. I still do believe that this most certainly could be what we are dealing with here. Indo-Scythians, Hepthalites, Kushans, etc., were all very similar, and we could derive our ancestry from any one of these peoples, but having spoken to David as well, he thinks that our high steppe in from the Bronze Age, not from later invaders. I guess we will just have to wait and see.

bmoney
01-16-2018, 01:17 AM
Is that another way of saying that the 'founding' admixture composition was sort of preserved due to the founder effect?

thats right

deuterium_1
01-16-2018, 10:08 AM
Yeah, perhaps this could indeed be the case. However, in your case it would be quite interesting as to how Scythian descendants would gain a status of Brahmin. I still do believe that this most certainly could be what we are dealing with here. Indo-Scythians, Hepthalites, Kushans, etc., were all very similar, and we could derive our ancestry from any one of these peoples, but having spoken to David as well, he thinks that our high steppe in from the Bronze Age, not from later invaders. I guess we will just have to wait and see.

I would have thought that these nomadic groups became Kshatriyas?.

khanabadoshi
01-16-2018, 10:35 AM
He ran some tests for me as well, however, they are a bit different and revolve around finding some pattern involving my y-HG.

All D-Stat Runs:

Afanasievo (https://pastebin.com/raw/8uj1iqrR)
Andronovo (https://pastebin.com/raw/ZfdKcfyD)
Corded Ware Germany (https://pastebin.com/raw/fJfPRXAF)
Poltavka (https://pastebin.com/raw/MzQjPmDx)
Sintashta (https://pastebin.com/raw/Ep6JLPe4)
Srubnaya (https://pastebin.com/raw/3dA574tJ)
Yamnaya Kalmykia (https://pastebin.com/raw/zyNbsYEF)
Yamnaya Samara

(https://pastebin.com/raw/dviGmrzL)
Ranked by probability of fit:

(https://pastebin.com/raw/dviGmrzL)

Afanasievo
0.93389


Poltavka
0.933709


Yamnaya Samara
0.843282


Yamnaya Kalmykia
0.622361


Srubnaya
0.363324


Corded Ware Germany
0.3587


Andronovo
0.245824


Shintashta
0.0858584


Leftpop percentage breakdown sorted by tail probability:



tail prob.

Iran_LN

Afanasievo

Adamanese_Onge
Han




0.93389

0.363
0.419
0.169

0.049














tail prob.

Iran_LN

Poltavka
Adamanese_Onge

Han





0.933709

0.372
0.4

0.166

0.062













tail prob.

Iran_LN

Yamnaya_Samara

Adamanese_Onge
Han




0.843282

0.397
0.389

0.138

0.076













tail prob.

Iran_LN

Yamnaya_Kalmykia

Adamanese_Onge

Han



0.622361

0.364
0.4

0.166

0.07














tail prob.

Iran_N

Srubnaya

Adamanese_Onge
Han
AG3-MA1



0.363324

0.357
0.329
0.203
0.053
0.058











tail prob.

Iran_N

Corded_Ware_Germany

Adamanese_Onge
Han
AG3-MA1


0.3587

0.363
0.333

0.206
0.054
0.044












tail prob.

Iran_N

Andronovo
Adamanese_Onge
Han
AG3-MA1


0.245824

0.371

0.316

0.201
0.048
0.064












tail prob.

Iran_N

Shintashta

Adamanese_Onge
Han
AG3-MA1


0.0858584

0.353

0.306

0.191
0.08
0.07





(https://pastebin.com/raw/dviGmrzL)

bmoney
01-16-2018, 01:13 PM
He ran some tests for me as well, however, they are a bit different and revolve around finding some pattern involving my y-HG.

All D-Stat Runs:

Afanasievo (https://pastebin.com/raw/8uj1iqrR)
Andronovo (https://pastebin.com/raw/ZfdKcfyD)
Corded Ware Germany (https://pastebin.com/raw/fJfPRXAF)
Poltavka (https://pastebin.com/raw/MzQjPmDx)
Sintashta (https://pastebin.com/raw/Ep6JLPe4)
Srubnaya (https://pastebin.com/raw/3dA574tJ)
Yamnaya Kalmykia (https://pastebin.com/raw/zyNbsYEF)
Yamnaya Samara

(https://pastebin.com/raw/dviGmrzL)
Ranked by probability of fit:

(https://pastebin.com/raw/dviGmrzL)

Afanasievo
0.93389


Poltavka
0.933709


Yamnaya Samara
0.843282


Yamnaya Kalmykia
0.622361


Srubnaya
0.363324


Corded Ware Germany
0.3587


Andronovo
0.245824


Shintashta
0.0858584


Leftpop percentage breakdown sorted by tail probability:



tail prob.

Iran_LN

Afanasievo

Adamanese_Onge
Han




0.93389

0.363
0.419
0.169

0.049














tail prob.

Iran_LN

Poltavka
Adamanese_Onge

Han





0.933709

0.372
0.4

0.166

0.062













tail prob.

Iran_LN

Yamnaya_Samara

Adamanese_Onge
Han




0.843282

0.397
0.389

0.138

0.076













tail prob.

Iran_LN

Yamnaya_Kalmykia

Adamanese_Onge

Han



0.622361

0.364
0.4

0.166

0.07














tail prob.

Iran_N

Srubnaya

Adamanese_Onge
Han
AG3-MA1



0.363324

0.357
0.329
0.203
0.053
0.058











tail prob.

Iran_N

Corded_Ware_Germany

Adamanese_Onge
Han
AG3-MA1


0.3587

0.363
0.333

0.206
0.054
0.044












tail prob.

Iran_N

Andronovo
Adamanese_Onge
Han
AG3-MA1


0.245824

0.371

0.316

0.201
0.048
0.064












tail prob.

Iran_N

Shintashta

Adamanese_Onge
Han
AG3-MA1


0.0858584

0.353

0.306

0.191
0.08
0.07





(https://pastebin.com/raw/dviGmrzL)

Very interesting - so as everyone says Yamna is the better fit

I guess the missing puzzle is just finding some eastern Yamna R1a sample

poi
01-16-2018, 05:23 PM
Rechecking the result files from Davidski, my 23andme v5's coverage is rather very poor (barely over 10%!). I might re-test when my ftdna is done.

khanabadoshi
01-16-2018, 05:46 PM
Very interesting - so as everyone says Yamna is the better fit

I guess the missing puzzle is just finding some eastern Yamna R1a sample

Yeah, because all the Afanasievo were R1b (but there were only 3 tested for yHG). I'm not sure about the Poltavka, but it's interesting that the best fits are apart in time and location (one east, one west). Afanasievo are basically eastern Siberian contemporaries of Yamna and, evidently, genetically the same as them. Poltavka is about 1000 years later and pretty similar to Yamna culture, but centered in the Don-Volga region and had different pottery, I think? I think it is the direct precursor to Srubnaya and a contributor to Shintashta. Afanasievo + Andronovo contributes to the Tarim Basin culture if I remember, and they had R1a Z93- results (which is relevant for myself). I assume from there contribution go to to Mittani and Vedic mediated by some group in some fashion or another. I'm sure others can parse this out much better. I only understand Steppe cultures in rudimental terms via little Wikipedia skim sessions.

https://upload.wikimedia.org/wikipedia/commons/d/db/Indo-European_migrations.gif (https://upload.wikimedia.org/wikipedia/commons/d/db/Indo-European_migrations.gif)


http://4.bp.blogspot.com/-tcGjUeTX2e4/VXtCOalNZJI/AAAAAAAAKIc/2OSeIs2vQ0M/s1600/nature14507-f1.jpg

Xehanort
01-16-2018, 06:06 PM
He ran some tests for me as well, however, they are a bit different and revolve around finding some pattern involving my y-HG.

All D-Stat Runs:

Afanasievo (https://pastebin.com/raw/8uj1iqrR)
Andronovo (https://pastebin.com/raw/ZfdKcfyD)
Corded Ware Germany (https://pastebin.com/raw/fJfPRXAF)
Poltavka (https://pastebin.com/raw/MzQjPmDx)
Sintashta (https://pastebin.com/raw/Ep6JLPe4)
Srubnaya (https://pastebin.com/raw/3dA574tJ)
Yamnaya Kalmykia (https://pastebin.com/raw/zyNbsYEF)
Yamnaya Samara

(https://pastebin.com/raw/dviGmrzL)
Ranked by probability of fit:

(https://pastebin.com/raw/dviGmrzL)

Afanasievo
0.93389


Poltavka
0.933709


Yamnaya Samara
0.843282


Yamnaya Kalmykia
0.622361


Srubnaya
0.363324


Corded Ware Germany
0.3587


Andronovo
0.245824


Shintashta
0.0858584


Leftpop percentage breakdown sorted by tail probability:



tail prob.

Iran_LN

Afanasievo

Adamanese_Onge
Han




0.93389

0.363
0.419
0.169

0.049














tail prob.

Iran_LN

Poltavka
Adamanese_Onge

Han





0.933709

0.372
0.4

0.166

0.062













tail prob.

Iran_LN

Yamnaya_Samara

Adamanese_Onge
Han




0.843282

0.397
0.389

0.138

0.076













tail prob.

Iran_LN

Yamnaya_Kalmykia

Adamanese_Onge

Han



0.622361

0.364
0.4

0.166

0.07














tail prob.

Iran_N

Srubnaya

Adamanese_Onge
Han
AG3-MA1



0.363324

0.357
0.329
0.203
0.053
0.058











tail prob.

Iran_N

Corded_Ware_Germany

Adamanese_Onge
Han
AG3-MA1


0.3587

0.363
0.333

0.206
0.054
0.044












tail prob.

Iran_N

Andronovo
Adamanese_Onge
Han
AG3-MA1


0.245824

0.371

0.316

0.201
0.048
0.064












tail prob.

Iran_N

Shintashta

Adamanese_Onge
Han
AG3-MA1


0.0858584

0.353

0.306

0.191
0.08
0.07





(https://pastebin.com/raw/dviGmrzL)

Thanks, he did not include Han in my runs, but he told me that it would reduce my Onge. I don't know why he didn't include it.

bmoney
01-16-2018, 11:36 PM
Thanks, he did not include Han in my runs, but he told me that it would reduce my Onge. I don't know why he didn't include it.

Han, interesting - the Han are 25 to 40% SE Asian (probs from the Aboriginal Taiwanese) and the rest NE Asian like Japanese and Koreans

The common link with ASI would be the SE Asian component that is usually part of the ghost-ASI when you look at calcs such as K9 ASI

pegasus
01-18-2018, 09:22 AM
Han, interesting - the Han are 25 to 40% SE Asian (probs from the Aboriginal Taiwanese) and the rest NE Asian like Japanese and Koreans

The common link with ASI would be the SE Asian component that is usually part of the ghost-ASI when you look at calcs such as K9 ASI

South Asians have their own special ghost populations, using Han and Onge are still poor proxies, will form their own distinctive cline because of latent ANE like ancestry or intrusive ANE, which was alluded to in that paper for mtdna U. U2 seems to suggest an archaic link with upper paleolithic populations and suggests. Kostenski , a nearly 40Kya man in Russia had it , he had basal C and mtdna U2. I know the only man upto now which has the same clade is a man in Nepal, there was a link to it somewhere .

bmoney
01-18-2018, 02:03 PM
South Asians have their own special ghost populations, using Han and Onge are still poor proxies, will form their own distinctive cline because of latent ANE like ancestry or intrusive ANE, which was alluded to in that paper for mtdna U. U2 seems to suggest an archaic link with upper paleolithic populations and suggests. Kostenski , a nearly 40Kya man in Russia had it , he had basal C and mtdna U2. I know the only man upto now which has the same clade is a man in Nepal, there was a link to it somewhere .

The Kostenki man had East Asian Y-dna C but had a lot of basal Eurasian

Whats a likely scenario for this?

pegasus
01-18-2018, 02:26 PM
The Kostenki man had East Asian Y-dna C but had a lot of basal Eurasian

Whats a likely scenario for this?

These people like him moved out of Central Asia into Eastern Europe, these populations are so old , only genomes from SC/S Asia will truly answer that. He is more SA shifted than MA1 even interestingly. No he is not East Asian at all, he is largely West Eurasian but with significant archaic SA related ancestry , C* originates from South Asia. These Kostenski people are mix of Basal and upper Paleolithic, with the later portion having significant archaic SA related ancestry.