PDA

View Full Version : Looking for ASI



Pages : 1 2 [3]

parasar
09-07-2015, 07:11 PM
There's some kind of coherent "Neolithic India" signal I got by mixing half ASE and Mideastern ENF ("SW-Asian") which peaks just east of the Indus:

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=85

Everything here is almost half and half something (an experiment to see how the software would group the admixture... since the presence of ANE ruins the signal because there's a large amount of ANE in the subcontinent).

Neolithic India = ASE + ENF
West Asia = ANE + ENF
SE-Asia = ASE + East Eurasian
Arctic = ANE + East Eurasian
EHG = ANE + WHG
EEF = WHG + ENF
Near-East = 90% ENF, 10% East African
NE-Asian = 12-15% ASE, ~85% East Eurasian

The leftover S-SE-Asia and Oceanian represent the edges of the Indian subcontinent pretty much. They go up as you get closer to the southeast, or the Himalayas, or the southwestern coast. Mixing this stuff with some of "West Asia" would give you the difference between this "Neolithic India" signal and something like HAP S-Indian.

Mixing half and half Neolithic India with West Asia would give you something approaching the HAP S-Indian signal... which suggests this should never have been called "South Indian" to begin with, it's actually just the Indus region plus the extra India-specific East Eurasian-like admixture (extra SE-Asian, NE-Asian, and Oceanian around the fringes). So that was a hybrid component averaging together two opposite halves of India and not representing any actual real population.

I'm still experimenting to see if this can be teased into a coherent signal with more "real" components.

In the meantime, I also got a decent equivalent to HAP S-Indian without the Oceanian:

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=88 (This is still running, I'll fill out the sheet when it is... these components resemble their actual names better)

Another thing I noticed, the Haryana Jatt comes out ~20% EHG no matter what I do, and the rest of us around half that, going up a bit in the Nepal Brahmin. And that (the Haryana Jatt) is approximately half a Swede's EHG. Since Western Steppe is more or less half EHG and half Gedrosia-WestAsian-like, that's almost like 40-50% of the Haryana Jatt's admixture that could be represented as Steppe, going down to 25% in the rest of India, with a sharp gradient around Haryana (in eastern Punjab, Nepal, presumably bordering areas of Rajasthan and then other parts of North India).

Another side effect of the Neolithic India calculator is that it will split admixture into West Asia and EEF instead of EHG for some Southern Europeans and Caucasians. But not for Northern Europeans or South/SC Asians curiously enough.

I'm trying right now (mostly unsuccessfully) to get the Neolithic India component to coexist properly with a HAP Gedrosian-like component.

EDIT: This also suggests the South Asia or South India scores of Afghans and other populations further away are purely "Neolithic India" (Except for those getting boosted by East Eurasian).

The nulls are quite informative.
Your West Asia peaks in a Gujarati-Maharashtran who has 0 EHG, 0 EEF and 0 Near East.
Your Neolithic-India peaks in a Gujjar (from Punjabi Gujrat, right?) who has 4.08%, 0.22%, 2.13% respectively.
Both of these seem very intertwined. Every one who has West Asia also has Neolithic India.

Dr_McNinja
09-07-2015, 08:53 PM
The nulls are quite informative.
Your West Asia peaks in a Gujarati-Maharashtran who has 0 EHG, 0 EEF and 0 Near East.
Your Neolithic-India peaks in a Gujjar (from Punjabi Gujrat, right?) who has 4.08%, 0.22%, 2.13% respectively.
Both of these seem very intertwined. Every one who has West Asia also has Neolithic India.

Yup, and Neolithic India component has 0% ANE. It's basically tracking ENF that's tied up with ASE.

I wouldn't say they peak in the same area. ANE/ENF peaks in Balochistan, but ASE/ENF peaks in West Punjab or just east of the Indus... which makes a lot of sense if you consider the history of the land for farming.

EDIT: Since HAP-Baloch has like 9.3% ASE in it, consider that something like 18-19% "Neolithic India"... which makes sense going by the rest of these components.

Varun R
09-07-2015, 09:03 PM
McNinja,

Will you be releasing the DIY files for these calculators? Also, I'm presuming you believe the Rakhigarhi skeletons will turn out to be (predominantly) a mix of ASI and ENF.

Dr_McNinja
09-07-2015, 09:51 PM
McNinja,

Will you be releasing the DIY files for these calculators? Also, I'm presuming you believe the Rakhigarhi skeletons will turn out to be (predominantly) a mix of ASI and ENF.Once I get a combination that I think is useful. I'm still experimenting with different ideas atm.

No idea about the ancient Haryana DNA actually. They'll definitely be ANE/ENF/ASE (I only took out ANE here so we could track the South Asian ENF signal easier)... maybe WHG and East Eurasian.

EDIT: I'm excited about it... that 4500ya period will shed a lot of light.

paulgill
09-08-2015, 01:15 AM
Once I get a combination that I think is useful. I'm still experimenting with different ideas atm.

No idea about the ancient Haryana DNA actually. They'll definitely be ANE/ENF/ASE (I only took out ANE here so we could track the South Asian ENF signal easier)... maybe WHG and East Eurasian.

EDIT: I'm excited about it... that 4500ya period will shed a lot of light.

Rakhigarhi skeletons I bet are J2 and/or G.

What is your guess?

Varun R
09-08-2015, 02:43 AM
@ Paulgill,

I wouldn't be surprised were L (old L1a or L3) to show up.

Edit: I need to learn the new nomenclature for all these haplogroups.

Dr_McNinja
09-08-2015, 01:11 PM
I did a run with ANE/ASE too (possibly representing South Asian ANE... which has a different spread than ASE/ENF):

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=89

I think Arctic here is acting kind of like a Northeast Asian for South Asians.

tamilgangster
09-08-2015, 07:33 PM
Yup, and Neolithic India component has 0% ANE. It's basically tracking ENF that's tied up with ASE.

I wouldn't say they peak in the same area. ANE/ENF peaks in Balochistan, but ASE/ENF peaks in West Punjab or just east of the Indus... which makes a lot of sense if you consider the history of the land for farming.



EDIT: Since HAP-Baloch has like 9.3% ASE in it, consider that something like 18-19% "Neolithic India"... which makes sense going by the rest of these components.


I noticed that there are 2 distinct sources of ASE in south Asia, a western type which is affiliated with the neotholic India component ur talking about, and an eastern one associated with austroasiatics, and mixed with east eurasian. Does The neotholic India component also carry some degree of east african component. The enf found in it I'm assuming is very basal eurasian related.

bored
09-08-2015, 07:43 PM
I did a run with ANE/ASE too (possibly representing South Asian ANE... which has a different spread than ASE/ENF):

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=89

I think Arctic here is acting kind of like a Northeast Asian for South Asians.

Interesting. You say that Pahari areas have more Oceanian but I notice that I get relatively low amounts on every calculator.

Dr_McNinja
09-09-2015, 09:02 PM
Interesting. You say that Pahari areas have more Oceanian but I notice that I get relatively low amounts on every calculator.
Yeah, yours is usually lower, but you do have higher East Eurasian. The spread is a bit random, I'd say the average is higher in Pahari areas.

It may not even be real Oceanian, but something which appears as Oceanian sometimes and East Eurasian/ASE other times. Maybe a branch of ASI that just stayed in the Himalayas and was a little different than that in South/SE India which appears closest to a mix of ASE, East Eurasian, and Oceanian.

I'm trying right now to modify HarappaWorld to remove Oceanian from S-Indian. I actually did that, now seeing if I can remove NE-Asian from it too.

I think it'd be worth uploading at least the Oceanian version to Gedmatch.

bored
09-09-2015, 09:17 PM
Yeah, yours is usually lower, but you do have higher East Eurasian. The spread is a bit random, I'd say the average is higher in Pahari areas.

It may not even be real Oceanian, but something which appears as Oceanian sometimes and East Eurasian/ASE other times. Maybe a branch of ASI that just stayed in the Himalayas and was a little different than that in South/SE India which appears closest to a mix of ASE, East Eurasian, and Oceanian.

I'm trying right now to modify HarappaWorld to remove Oceanian from S-Indian. I actually did that, now seeing if I can remove NE-Asian from it too.

I think it'd be worth uploading at least the Oceanian version to Gedmatch.

Well a lot of people in this area are getting higher East Eurasian. Even some Punjabi Brahmins get higher East Eurasian than me. Not sure where it's coming from. Some Tarkhans seem to have it higher too. I'm from southeast Jammu so I thought maybe I have it higher than others in my region. I'm not sure about that. It could be random.. My mom says her paternal side is originally from Pahari areas in Pakistan.

Dr_McNinja
09-10-2015, 05:06 PM
What do you all think of this version of Harappa?

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=91

The Papuan values look close enough to the values in other calculators.

Sapporo
09-10-2015, 05:14 PM
What do you all think of this version of Harappa?

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=91

The Papuan values look close enough to the values in other calculators.

Is the South Indian component here more like a Neolithic Indian component with the excess ASI going into Papuan, SE Asian, etc. What would increased Siberian, Beringian, Amerindian, etc. represent? ANE that was hidden in the SI component?

Dr_McNinja
09-10-2015, 05:20 PM
What do you all think of this version of Harappa?

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=91

The Papuan values look close enough to the values in other calculators.

This brings up some of the good things about the Harappa calculator. The Caucasian component is very useful in South Asia. So is the SW-Asian component. Excess SW-Asian and Mediterranean noise is now gone. Any appreciable amounts of SW-Asian probably are significant now. For example, for soulblighter (South Indian Brahmin), this is suggestive of the idea that his extra SW-Asian is coming out of his S-Indian since for actual South Indians, the Papuan is related to their ASI input. But for everyone else without excess SW-Asian spilling it out, the Papuan is probably associated with the higher East Eurasian components in the north (Himalayas).

For everyone in the northwest, SW-Asian probably points to a tiny bit of Baloch and/or foreign ancestry.

The "corrected" Caucasian amounts are more relevant now, and they reveal the slight foreign ancestry of my friend, the UP/Mumbai Muslim (who does have Afghan relatives in his matches and in his not-too-distant family tree... plus he actually looks Afghan (he's European white, not just light skinned by South Asian standards)). Now his results are more like other UP Muslims (though his ancestry includes Mumbai).


Is the South Indian component here more like a Neolithic Indian component with the excess ASI going into Papuan, SE Asian, etc. What would increased Siberian, Beringian, Amerindian, etc. represent? ANE?No, this would be like a more relevant version of actual/original HAP S-Indian. It's the same component... just without the Papuan. The excess SE-Asian/NE-Asian and probably the Arctic components represent the plentiful and near-indigenous East Eurasian within South Asia. We should have been expecting that across the North, so close to the Himalayas. There's SE-Asian as well as NE-Asian present there (SE-Asian actually better represents South China/Vietnam-area input which should historically have extended into the subcontinent, not the islands of actual Southeast Asia like Malaysia/Indonesia). The Arctic components could be East Eurasian noise tying up some ANE or representative of actual Arctic-like admixture which made its way into South Asia. There were probably multiple waves of Eastern Steppe input into South Asia, from ancient as well as recent (IE) times.

Dr_McNinja
09-10-2015, 05:26 PM
Btw, in harappa, NE-Euro and Caucasian probably represent a very closely related signal. The NE-Euro is boosted by Caucasian-like admixture and doesn't exactly track WHG from other calculators.

Since the two most relevant signals for the northwest subcontinent area are Gedrosian (Balochistan) and NE-Euro (Haryana/North India), it's possible that excess Caucasian inbetween these areas is a noisy intermediate (where the distinction between local Gedrosian and foreign Steppe isn't clear). Just one possible interpretation.

I say "Excess" because clearly there is a Caucasian signal in the S-Indian component. Probably Gedrosian in actuality that "comes apart" as that component breaks up a little.

bored
09-10-2015, 05:50 PM
Btw, in harappa, NE-Euro and Caucasian probably represent a very closely related signal. The NE-Euro is boosted by Caucasian-like admixture and doesn't exactly track WHG from other calculators.

Since the two most relevant signals for the northwest subcontinent area are Gedrosian (Balochistan) and NE-Euro (Haryana/North India), it's possible that excess Caucasian inbetween these areas is a noisy intermediate (where the distinction between local Gedrosian and foreign Steppe isn't clear). Just one possible interpretation.

I say "Excess" because clearly there is a Caucasian signal in the S-Indian component. Probably Gedrosian in actuality that "comes apart" as that component breaks up a little.

Judging by the geographical distribution of high and low Caucasus in South Asia it seems more likely that it represents West Asian like input.

bored
09-10-2015, 05:56 PM
This brings up some of the good things about the Harappa calculator. The Caucasian component is very useful in South Asia. So is the SW-Asian component. Excess SW-Asian and Mediterranean noise is now gone. Any appreciable amounts of SW-Asian probably are significant now. For example, for soulblighter (South Indian Brahmin), this is suggestive of the idea that his extra SW-Asian is coming out of his S-Indian since for actual South Indians, the Papuan is related to their ASI input. But for everyone else without excess SW-Asian spilling it out, the Papuan is probably associated with the higher East Eurasian components in the north (Himalayas).

For everyone in the northwest, SW-Asian probably points to a tiny bit of Baloch and/or foreign ancestry.

The "corrected" Caucasian amounts are more relevant now, and they reveal the slight foreign ancestry of my friend, the UP/Mumbai Muslim (who does have Afghan relatives in his matches and in his not-too-distant family tree... plus he actually looks Afghan (he's European white, not just light skinned by South Asian standards)). Now his results are more like other UP Muslims (though his ancestry includes Mumbai).

No, this would be like a more relevant version of actual/original HAP S-Indian. It's the same component... just without the Papuan. The excess SE-Asian/NE-Asian and probably the Arctic components represent the plentiful and near-indigenous East Eurasian within South Asia. We should have been expecting that across the North, so close to the Himalayas. There's SE-Asian as well as NE-Asian present there (SE-Asian actually better represents South China/Vietnam-area input which should historically have extended into the subcontinent, not the islands of actual Southeast Asia like Malaysia/Indonesia). The Arctic components could be East Eurasian noise tying up some ANE or representative of actual Arctic-like admixture which made its way into South Asia. There were probably multiple waves of Eastern Steppe input into South Asia, from ancient as well as recent (IE) times.

Indigenous East Eurasian? Isn't it simpler to assume that it isn't indigenous? After all, it's higher in groups that live closer to the predominantly East Eurasian non-South Asians. At least this is the case in the North.

Dr_McNinja
09-10-2015, 06:17 PM
Judging by the geographical distribution of high and low Caucasus in South Asia it seems more likely that it represents West Asian like input.
Steppe admixture was also likely quite West Asian. Distinct, however, from Neolithic-era West Asian-like admixture, even though the software will have trouble telling the difference.


Indigenous East Eurasian? Isn't it simpler to assume that it isn't indigenous? After all, it's higher in groups that live closer to the predominantly East Eurasian non-South Asians. At least this is the case in the North.I said near indigenous because some, if not most, of the trace amounts were probably there in the north for thousands of years. Pre-IE even, I'd guess.

bored
09-10-2015, 07:34 PM
What do you all think of this version of Harappa?

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=91

The Papuan values look close enough to the values in other calculators.

Oh you mean Harappa DIY- Oceania. It looks decent IMO.

Dr_McNinja
09-10-2015, 11:05 PM
I'm doing another version of those "null" components (50/50 mix of ANE/ENF, ASE/ENF, etc). I added an "ASE/WHG" component for Indo-Aryan. It hits ~4% in the Haryana Jatt (who still has much greater EHG), but ~9% in Sweden and ~4% in other Europeans (Chad in this case).

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=93

It curiously seems to parallel the WHG amounts in West Eurasia K8 for the non-Haryana South Asians, in spite of the EHG/EEF components.

(I'm still running through the rest of the individuals, so the sheet isn't completed)

ashwinb
09-11-2015, 02:07 AM
So does the EEF and EHG get attracted towards west asia for some samples(like mine), or something like that?

Dr_McNinja
09-11-2015, 03:16 AM
So does the EEF and EHG get attracted towards west asia for some samples(like mine), or something like that?EHG/EEF are both WHG/European Hunter Gatherer derivatives so they just represent Steppe input probably. This calculator is based off a Eurogenes ANE K7 one and you don't get any WHG in that one IIRC.

ashwinb
09-11-2015, 01:18 PM
And what's the difference between "near east" and west asia? and are they supposed to be different from neolithic india"s non-ASI ?

Dr_McNinja
09-11-2015, 02:13 PM
And what's the difference between "near east" and west asia? and are they supposed to be different from neolithic india"s non-ASI ?
Near east is predominantly ENF (90% ENF, 10% East African).

"Neolithic India" is just a mix of ASE and ENF only (50/50).

West Asia is ANE/ENF (Gedrosia-Caucasus).

Anabasis
09-11-2015, 02:31 PM
Near east is predominantly ENF (90% ENF, 10% East African).

"Neolithic India" is just a mix of ASE and ENF only (50/50).

West Asia is ANE/ENF (Gedrosia-Caucasus).

I dont think that gedrosia respresent ANE. ANE seems much more north euroasian. But gedrosia seems a Caucaus component unique to South central Asia.

Kurd
09-11-2015, 03:42 PM
I dont think that gedrosia respresent ANE. ANE seems much more north euroasian. But gedrosia seems a Caucaus component unique to South central Asia.

Your statement is correct. I think McNinja does not mean the gedrosian component acted as an ANE vector for S Asians, rather I believe he means the populations of gedrosia, Baloch, Pathans, Kalash, etc. collectively acted as an ANE vector for S Asians.

Coldmountains
09-11-2015, 04:51 PM
EHG/EEF are both WHG/European Hunter Gatherer derivatives so they just represent Steppe input probably. This calculator is based off a Eurogenes ANE K7 one and you don't get any WHG in that one IIRC.

I am more and more inclined to ignore ANE for estimating Indo-Iranian steppe ancestry in Asia. Chad is soon releasing a calculator for EHG and with it we should get a good idea about how significant is Indo-Iranian steppe ancestry in Asia. Jatts almost look like Pamiri in terms of WHG ancestry, which is in this case probably just additional EHG ancestry. But the peak of EHG among Jatts is really interesting and needs to be explained. If I am honest I also don't believe anymore that Sintashta-like ancestry is so high in South Central Asia. Now I think it is rather around ~ 35% for most South Central Asians and just for some Jatts and Pamiri significantly above that. I also think thar it is higher among southern or central Pashtuns than among many northern Pashtuns even when they have more ANE but ANE peaks in the Hindukush among non-IE people (Burusho) so it is not a good indicator for Sintashta-like admixture in this case.

Dr_McNinja
09-11-2015, 07:40 PM
I am more and more inclined to ignore ANE for estimating Indo-Iranian steppe ancestry in Asia. Chad is soon releasing a calculator for EHG and with it we should get a good idea about how significant is Indo-Iranian steppe ancestry in Asia. Jatts almost look like Pamiri in terms of WHG ancestry, which is in this case probably just additional EHG ancestry. But the peak of EHG among Jatts is really interesting and needs to be explained. If I am honest I also don't believe anymore that Sintashta-like ancestry is so high in South Central Asia. Now I think it is rather around ~ 35% for most South Central Asians and just for some Jatts and Pamiri significantly above that. I also think thar it is higher among southern or central Pashtuns than among many northern Pashtuns even when they have more ANE but ANE peaks in the Hindukush among non-IE people (Burusho) so it is not a good indicator for Sintashta-like admixture in this case.

I'm inclined to agree with you on all counts. ANE in South Asia is old and seems to track towards both South India and the Indus region. South Asian ENF (representing probably a neolithic origin) peaks in Indus rather than both. So I think high ANE was already in the Indus region before even ENF showed up.

I think Steppe ancestry might peak at 40-50% in Haryana Jatts. If the 20% EHG figure is meaningful that is. Everyone else is probably about half that, with a sharp gradient near North India (the ASE/WHG 'Indo-Aryan' component shows the spread... curiously being a relatively stronger signal in Brahmins than EHG/EEF on its own).

But even then, I don't think that Steppe population was Sintashta. Or even Karasuk. It's curious that the SE-Asian the Europeans got immediately goes into ASE/WHG combo instead.

Coldmountains
09-11-2015, 10:22 PM
I'm inclined to agree with you on all counts. ANE in South Asia is old and seems to track towards both South India and the Indus region. South Asian ENF (representing probably a neolithic origin) peaks in Indus rather than both. So I think high ANE was already in the Indus region before even ENF showed up.

I think Steppe ancestry might peak at 40-50% in Haryana Jatts. If the 20% EHG figure is meaningful that is. Everyone else is probably about half that, with a sharp gradient near North India (the ASE/WHG 'Indo-Aryan' component shows the spread... curiously being a relatively stronger signal in Brahmins than EHG/EEF on its own).

But even then, I don't think that Steppe population was Sintashta. Or even Karasuk. It's curious that the SE-Asian the Europeans got immediately goes into ASE/WHG combo instead.

Sintashta represents an important part of the history of the Indo-Iranian migration but I agree we should not just look at them. Most of South Asian R1a is L657 ( I am myself a L657/Y6 carrier) and I wonder if the first L657/Y6/Y9 carriers were genetically just the same like Sintashta or slightly different so it would be interesting to see if their is some kind of "steppe" signal among populations rich in L657 or dominated by it which is unique for them. I expect them to be very similar but I am curious if Proto-L657/Y9/Y6 had different amounts of EHG or EEF compared to the few Sintashta folks tested yet or of they even had some kind of "exotic" ancestry absent in Sintashta. I would not exclude that it were Sintashta folks which migrated directly to South Asia and that L657 is even from there but there is definitely more research needed for proving or disproving this.

tamilgangster
09-12-2015, 03:17 AM
Judging by the geographical distribution of high and low Caucasus in South Asia it seems more likely that it represents West Asian like input.

Not necesarily On harappa DNA, tribes such as irulas who lack baloch score visible levels of Caucasian, i believe this is due to excess west eurasian/SW asian type ENF from Neotholic Indian.

tamilgangster
09-12-2015, 03:31 AM
Indigenous East Eurasian? Isn't it simpler to assume that it isn't indigenous? After all, it's higher in groups that live closer to the predominantly East Eurasian non-South Asians. At least this is the case in the North.

Its fairly indigenous, predates dravidians/gedrosians, that why east eurasian scores are directly proportianal to ASI(within the same region) and it peaks among tribals in central and eastern India. This strain probably as indigenous to south asia as the Neotholic India Dr. Mcninja is talking about and extendend into south asia from the North East. The austroasiatics and the neotholic India populations both carried ASI.

tamilgangster
09-12-2015, 03:36 AM
I'm doing another version of those "null" components (50/50 mix of ANE/ENF, ASE/ENF, etc). I added an "ASE/WHG" component for Indo-Aryan. It hits ~4% in the Haryana Jatt (who still has much greater EHG), but ~9% in Sweden and ~4% in other Europeans (Chad in this case).

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=93

It curiously seems to parallel the WHG amounts in West Eurasia K8 for the non-Haryana South Asians, in spite of the EHG/EEF components.

(I'm still running through the rest of the individuals, so the sheet isn't completed)

THe fact that the neotholic India component, is directly proportional to the SOuth asian ANE, means that the Neotholic India component mixed with ANE somewhere in the northwest predating gedrosian admixture

MonkeyDLuffy
09-12-2015, 04:25 AM
I dont thInk only jatts score high WHG, Nepali brahmins scored high WHG as well. Also I'm pretty sure there are other groups who will have high steppe admixture.

bored
09-12-2015, 04:40 AM
I dont thInk only jatts score high WHG, Nepali brahmins scored high WHG as well. Also I'm pretty sure there are other groups who will have high steppe admixture.

Hindu Jats of Haryana/UP/Rajasthan have high steppe admix. Muslim Jatts would score lower and Sikh Jatts probably at the same level as many Brahmins. You have it higher too.

MonkeyDLuffy
09-12-2015, 06:50 AM
Hindu Jats of Haryana/UP/Rajasthan have high steppe admix. Muslim Jatts would score lower and Sikh Jatts probably at the same level as many Brahmins. You have it higher too.

We don't have enough samples to have a conclusion anyway. Who knows maybe a chamar might end up having high steppe. Also to note we don't have samples from every community either.

everest59
09-12-2015, 01:03 PM
I would say that certain Brahmin groups are pretty much equal in terms of WHG. IMO Nepali brahmins and UP Brahmins as well as Gujarati-a. Let's suppose Nepali Brahmins average 12% NE Euro on Harappa, and that Sintashta score 50%. I would put total steppe Admixture that is Sintashta like at 12/50 or 24%. Now, if we had BMAC samples that would be pushed higher by quite a bit.

DMXX
09-12-2015, 02:39 PM
I would say that certain Brahmin groups are pretty much equal in terms of WHG. IMO Nepali brahmins and UP Brahmins as well as Gujarati-a. Let's suppose Nepali Brahmins average 12% NE Euro on Harappa, and that Sintashta score 50%. I would put total steppe Admixture that is Sintashta like at 12/50 or 24%. Now, if we had BMAC samples that would be pushed higher by quite a bit.

Agreed. The other day, I fiddled around with conversion rates as you've done above for some of these ADMIXTURE calculators (this is in effect the "Component Proportions" approach I elaborated upon in this entry based on Kurd's Eurasia K20 (http://vaedhya.blogspot.co.uk/2015/08/steppe-ancestry-estimations-in-west.html)).

For this example, I've used a Kashmiri Pandit sample and ran it through three calculators. Results (and working) are below. Omitted the Kashmiri's GEDmatch ID in keeping with forum rules.



Calculator Peak Component Sintashta (M690970) Kashmiri Pandit (M------) Multip.factor Projected Sintashta admix
HarappaWorld NE-Euro 57.80% 8.92% x1.73 15.43%
Gedrosia K12 SINTASHTA_STEPPE_HERDERS 98.56% 8.28% x1.01 8.38%
Eurogenes ANE K7 WHG-UHG 68.27% 4.21% x1.46 6.2%


There's an inverse correlation between the degree of peak component specificity to Sintashta with the extent of Sintashta admix in this individual. The "multiplication factor" is over 70% greater in HarappaWorld compared to Kurd's Gedrosia K12.

For the purposes of estimating Sintashta admix in Asia using ADMIXTURE, Gedrosia K12 should support us the best here for relative comparisons. That was Kurd's stated goal to begin with (focusing the steppe signal). 6-15% is quite a difference!

everest59
09-12-2015, 03:54 PM
Agreed. The other day, I fiddled around with conversion rates as you've done above for some of these ADMIXTURE calculators (this is in effect the "Component Proportions" approach I elaborated upon in this entry based on Kurd's Eurasia K20 (http://vaedhya.blogspot.co.uk/2015/08/steppe-ancestry-estimations-in-west.html)).

For this example, I've used a Kashmiri Pandit sample and ran it through three calculators. Results (and working) are below. Omitted the Kashmiri's GEDmatch ID in keeping with forum rules.



Calculator Peak Component Sintashta (M690970) Kashmiri Pandit (M------) Multip.factor Projected Sintashta admix
HarappaWorld NE-Euro 57.80% 8.92% x1.73 15.43%
Gedrosia K12 SINTASHTA_STEPPE_HERDERS 98.56% 8.28% x1.01 8.38%
Eurogenes ANE K7 WHG-UHG 68.27% 4.21% x1.46 6.2%


There's an inverse correlation between the degree of peak component specificity to Sintashta with the extent of Sintashta admix in this individual. The "multiplication factor" is over 70% greater in HarappaWorld compared to Kurd's Gedrosia K12.

For the purposes of estimating Sintashta admix in Asia using ADMIXTURE, Gedrosia K12 should support us the best here for relative comparisons. That was Kurd's stated goal to begin with (focusing the steppe signal). 6-15% is quite a difference!

I notice that Sintashta scores 58% NE Euro on Harappa.
Using your conversion rates, I get:
Harappa (Multiplication factor of 1.73):
21.59% steppe

Gedrosia K12:
11.42

Eurogenes ANE K7:
11.94

DMXX
09-12-2015, 04:12 PM
I notice that Sintashta scores 58% NE Euro on Harappa.


Yes. Harappa looks like the "worst" here for Sintashta specificity.

My own results are 11.37% (Harappa), 9.16% (Gedrosia K12) and 15.6% (ANE K7). Not as wide a margin as yours, though (11.4-21.6%).

everest59
09-12-2015, 04:28 PM
Yes. Harappa looks like the "worst" here for Sintashta specificity.

My own results are 11.37% (Harappa), 9.16% (Gedrosia K12) and 15.6% (ANE K7). Not as wide a margin as yours, though (11.4-21.6%).

If I recall, the Haak paper had Corded Ware-like ancestry peaking at 75% in Norwegians/Lithuanians. Corded Ware and Sintashta are supposedly similar with some differences. Now, Gedrosia K12 is not meant for Europeans per se, but I notice Norwegians on this board scoring around 48% of the Sintashta component. So maybe the multiplication factor needs to go up.

DMXX
09-12-2015, 04:48 PM
If I recall, the Haak paper had Corded Ware-like ancestry peaking at 75% in Norwegians/Lithuanians. Corded Ware and Sintashta are supposedly similar with some differences. Now, Gedrosia K12 is not meant for Europeans per se, but I notice Norwegians on this board scoring around 48% of the Sintashta component. So maybe the multiplication factor needs to go up.

The Gedrosia K12 calculator applied to the two Corded Ware genomes:



Corded Ware I0103 - kit M966366
Population
S_INDIAN 0.77%
SUB_SAHARAN 0.49%
EARLY_EUROPEAN_FARMERS 4.58%
SW_ASIAN 3.35%
W_SIBERIAN -
SE_ASIAN -
BALOCHI 7.09%
SINTASHTA_STEPPE_HERDERS 73.09%
INDO_TIBETAN -
CAUCASUS 10.08%
E_AFRICAN -
E_SIBERIAN 0.55%

Corded Ware I0104 M622615
Population
S_INDIAN 2.20%
SUB_SAHARAN 0.10%
EARLY_EUROPEAN_FARMERS 19.31%
SW_ASIAN -
W_SIBERIAN 0.07%
SE_ASIAN -
BALOCHI 14.44%
SINTASHTA_STEPPE_HERDERS 44.94%
INDO_TIBETAN -
CAUCASUS 18.94%
E_AFRICAN -
E_SIBERIAN -


The second Corded Ware sample (I0104) scores even less SINTASHTA_STEPPE_HERDERS than the Norwegians you've noticed here, yet the other approached 75%. Looks like there was greater diversity in component values among CW than Sintashta-Andronovo.

I'm not sure if this supports upping the multiplication factor for the purpose of Sintashta-related admixture in Asia, since SINTASHTA_STEPPE_HERDERS was already configured to be Sintashta-centric, which was always the ideal for us...

I do agree the value's probably an underestimation of sorts. The Fst distances support the regular complaint some (such as Sein) have had regarding ADMIXTURE; it's quite likely the close distances mean some of the actual ancestry is being absorbed in drifted components. It's a valid criticism. Until we have aDNA, how we can circumvent this issue remains an on-going thorn in our sides.

[Edit]: Eureka! Dienekes' old methodologies might help us here. :)

In 2012, he expressed his Dodecad K12b components as iterations of each other (http://dienekes.blogspot.qa/2012/08/inter-relationships-of-dodecad-k12b-and.html):

http://4.bp.blogspot.com/-G5HKtB9fZJs/UDcviIMZrLI/AAAAAAAAFvQ/cLoB5_uvXYQ/s1600/barplot.png

If Kurd could express the Balochi and Caucasus components as combinations of the others (including Sintashta_steppe) in a supervised run, we'd have a quantifiable means of calculating a good portion of this "hidden" Sintashta ancestry.

At that stage, all one would need to do is calculate the "hidden" Sintashta in their Caucasus and Balochi components (using appropriate multipliers) and combining it with the "pure" signal manifesting in the separate component.

everest59
09-12-2015, 04:56 PM
It would be great if Kurd could replicate what Dienekes did above. It would give a better picture.

Coldmountains
09-12-2015, 04:57 PM
If I recall, the Haak paper had Corded Ware-like ancestry peaking at 75% in Norwegians/Lithuanians. Corded Ware and Sintashta are supposedly similar with some differences. Now, Gedrosia K12 is not meant for Europeans per se, but I notice Norwegians on this board scoring around 48% of the Sintashta component. So maybe the multiplication factor needs to go up.

The Norwegians were between 42%-45% Sintashta. Lithuanians are certainly above 50% and Russians between 45%-50% in many cases probably above.

everest59
09-12-2015, 04:58 PM
The Norwegians were between 42%-45% Sintashta. Lithuanians are certainly above 50% and Russians between 45%-50% in many cases probably above.

I see. Where did you see that though? A different paper?

Coldmountains
09-12-2015, 05:09 PM
I see. Where did you see that though? A different paper?

That are just the Gedrosia K12 results I remember. I would say that" Sintashta" peaks in the Baltic States (50%+) and after that among East Slavs and Scandinavians (40%-50%). But it is probably slightly higher on average among the former.

Kurd
09-12-2015, 05:20 PM
At that stage, all one would need to do is calculate the "hidden" Sintashta in their Caucasus and Balochi components (using appropriate multipliers) and combining it with the "pure" signal manifesting in the separate component.

Alternatively, I could remove any modern pops acting as BA steppe proxies from the run, and only leave some modern pops acting as proxies for other assumed basal ancestral comps such as Bedouin, Ami, Dai, etc in the run. I actually have been tweeking such runs over the past couple of weeks time permitting

everest59
09-12-2015, 05:22 PM
That are just the Gedrosia K12 results I remember. I would say that" Sintashta" peaks in the Baltic States (50%+) and after that among East Slavs and Scandinavians (40%-50%). But it is probably slightly higher on average among the former.

Oh, okay. I was only looking at Evon's results that he posted in the K12 thread.
If Yamnaya is close to 50% in Norwegians, I think Sintashta-like ancestry has to be higher than that (by which I include Corded Ware as well, which is at close to 75% per Haak).
Now, they found it higher in Norwegians than in Lithuanians, which I don't think is accurate.

Kurd
09-12-2015, 05:28 PM
If I recall, the Haak paper had Corded Ware-like ancestry peaking at 75% in Norwegians/Lithuanians. Corded Ware and Sintashta are supposedly similar with some differences. Now, Gedrosia K12 is not meant for Europeans per se, but I notice Norwegians on this board scoring around 48% of the Sintashta component. So maybe the multiplication factor needs to go up.

The Gedrosia 12 has Sintashta peaking at ~70% in Lithuanians. It is around 50% for Norwegians.

Coldmountains
09-12-2015, 05:56 PM
The Gedrosia 12 has Sintashta peaking at ~70% in Lithuanians. It is around 50% for Norwegians.

All Norwegians I have seen so far were significantly under 50% Sintashta. Calamus (42%) and Evon (42%) posted here their results (http://www.anthrogenica.com/showthread.php?5174-Post-your-Gedmatch-calculator-Gedrosia-12-results/page5). Lithuanians are not 70% Sintashta I think. A half African (Afro-American?) half Baltic member posted here his results and he was 27% Sintashta so his Baltic side must be above 50% Sintashta but not anywhere close to 70%. Because of this inaccurate population averages used for the oracle all Norwegians, which posted their results so far, got Oracle results like 75% Norwegian 25% Albanian/Turkish.

Kurd
09-12-2015, 06:50 PM
All Norwegians I have seen so far were significantly under 50% Sintashta. Calamus (42%) and Evon (42%) posted here their results (http://www.anthrogenica.com/showthread.php?5174-Post-your-Gedmatch-calculator-Gedrosia-12-results/page5). Lithuanians are not 70% Sintashta I think there is something wrong with the population averages. A half African (Afro-American?) half Baltic member posted here his results and he was 27% Sintashta so his Baltic side must be above 50% Sintashta but not anywhere close to 70%. Because of this inaccurate population averages used for the oracle all Norwegians, which posted their results so far, got Oracle results like 75% Norwegian 25% Albanian/Turkish.

FIRST, did you read the disclosure on the calculator's page; "This calculator has been designed for individuals of predominately South Asian and West Asian ancestry for inferring gedrosian Balochi admixture. Since those populations were mostly used to source allele frequencies, individuals with majority ancestry from outside those regions will most likely find this calculator less accurate and informative.".....just want to make sure!


SECOND, The Norwegians were from the Haak dataset, below are their scores, if you have a problem with them, you can write Haak. Also, below you will find individual scores for various Baltic individuals and others. So what do mean by "Lithuanians are not 70% Sintashta"?? In fact, Everest had pointed out that Haak found Corded Ware peaked at around 75% in Lithuanians. So Haak's results are not inconsistent with the 70% Sintashta average for Lithuanians, even though this calcuator was not intended for Europeans.

IND ID SINTASHTA_STEPPE_HERDERS
BA_Sintashta RISE386 99.99%
BA_Sintashta RISE391 99.99%
BA_Sintashta RISE392 99.99%
Corded_Ware_LN I0103 99.99%
Lithuanian lithuania9 73.45%
Estonian Est391 72.86%
Estonian Est377 72.59%
Lithuanian LithuanianA1 71.55%
Lithuanian LithuanianF1 71.45%
Lithuanian lithuania10 71.24%
Lithuanian lithuania2 70.96%
Estonian Est394 70.03%
Lithuanian LithuanianD1 69.26%
Finnish HG00171 69.03%
Lithuanian lithuania3 68.48%
Estonian Est358 68.24%
Finnish HG00182 68.24%
Estonian Est393 67.81%
Lithuanian lithuania8 67.72%
Russian HGDP00892 67.71%
Estonian Est375 67.61%
Lithuanian LithuanianE2 67.39%
Estonian Est372 66.58%
Estonian Est397 65.02%
Estonian Est380 64.97%
Finnish HG00266 64.91%
Lithuanian lithuania1 64.55%
Estonian Est400 63.77%
Finnish HG00190 63.67%
Finnish HG00173 62.23%
Russian HGDP00901 62.22%
Finnish HG00183 62.21%
Russian HGDP00880 62.08%
Russian HGDP00887 61.97%
Russian HGDP00897 61.73%
Russian HGDP00891 61.68%
Finnish HG00174 61.49%
Russian HGDP00889 61.00%
Russian HGDP00882 60.88%
Russian HGDP00895 60.85%
Russian HGDP00902 60.60%
Russian HGDP00896 60.32%
Russian HGDP00900 60.05%
Russian HGDP00898 60.00%
Russian HGDP00903 59.85%
Russian HGDP00884 59.77%
Russian HGDP00883 59.39%
Russian HGDP00890 59.25%
Russian HGDP00899 59.24%
Russian HGDP00879 58.45%
Russian HGDP00893 58.26%
Russian HGDP00894 58.22%
Russian HGDP00888 56.66%
Norwegian NOR107 55.72%
Norwegian NOR111 54.66%
Norwegian NOR146 53.03%
Norwegian NOR148 52.71%
Norwegian NOR119 52.60%
Norwegian NOR109 52.50%
Norwegian NOR126 52.26%
Norwegian NOR108 52.15%
Norwegian NOR124 51.94%
Norwegian NOR101 49.34%
Norwegian NOR106 47.16%

THIRD, it is obvious that the oracles will not be best fits for Europeans, since a very limited number of European pops were used in the run.

Coldmountains
09-12-2015, 09:22 PM
FIRST, did you read the disclosure on the calculator's page; "This calculator has been designed for individuals of predominately South Asian and West Asian ancestry for inferring gedrosian Balochi admixture. Since those populations were mostly used to source allele frequencies, individuals with majority ancestry from outside those regions will most likely find this calculator less accurate and informative.".....just want to make sure!


SECOND, The Norwegians were from the Haak dataset, below are their scores, if you have a problem with them, you can write Haak. Also, below you will find individual scores for various Baltic individuals and others. So what do mean by "Lithuanians are not 70% Sintashta"?? In fact, Everest had pointed out that Haak found Corded Ware peaked at around 75% in Lithuanians. So Haak's results are not inconsistent with the 70% Sintashta average for Lithuanians, even though this calcuator was not intended for Europeans.

IND ID SINTASHTA_STEPPE_HERDERS
BA_Sintashta RISE386 99.99%
BA_Sintashta RISE391 99.99%
BA_Sintashta RISE392 99.99%
Corded_Ware_LN I0103 99.99%
Lithuanian lithuania9 73.45%
Estonian Est391 72.86%
Estonian Est377 72.59%
Lithuanian LithuanianA1 71.55%
Lithuanian LithuanianF1 71.45%
Lithuanian lithuania10 71.24%
Lithuanian lithuania2 70.96%
Estonian Est394 70.03%
Lithuanian LithuanianD1 69.26%
Finnish HG00171 69.03%
Lithuanian lithuania3 68.48%
Estonian Est358 68.24%
Finnish HG00182 68.24%
Estonian Est393 67.81%
Lithuanian lithuania8 67.72%
Russian HGDP00892 67.71%
Estonian Est375 67.61%
Lithuanian LithuanianE2 67.39%
Estonian Est372 66.58%
Estonian Est397 65.02%
Estonian Est380 64.97%
Finnish HG00266 64.91%
Lithuanian lithuania1 64.55%
Estonian Est400 63.77%
Finnish HG00190 63.67%
Finnish HG00173 62.23%
Russian HGDP00901 62.22%
Finnish HG00183 62.21%
Russian HGDP00880 62.08%
Russian HGDP00887 61.97%
Russian HGDP00897 61.73%
Russian HGDP00891 61.68%
Finnish HG00174 61.49%
Russian HGDP00889 61.00%
Russian HGDP00882 60.88%
Russian HGDP00895 60.85%
Russian HGDP00902 60.60%
Russian HGDP00896 60.32%
Russian HGDP00900 60.05%
Russian HGDP00898 60.00%
Russian HGDP00903 59.85%
Russian HGDP00884 59.77%
Russian HGDP00883 59.39%
Russian HGDP00890 59.25%
Russian HGDP00899 59.24%
Russian HGDP00879 58.45%
Russian HGDP00893 58.26%
Russian HGDP00894 58.22%
Russian HGDP00888 56.66%
Norwegian NOR107 55.72%
Norwegian NOR111 54.66%
Norwegian NOR146 53.03%
Norwegian NOR148 52.71%
Norwegian NOR119 52.60%
Norwegian NOR109 52.50%
Norwegian NOR126 52.26%
Norwegian NOR108 52.15%
Norwegian NOR124 51.94%
Norwegian NOR101 49.34%
Norwegian NOR106 47.16%

THIRD, it is obvious that the oracles will not be best fits for Europeans, since a very limited number of European pops were used in the run.

This still does not explain why the results of the Norwegians were so different from this. There is not such a huge difference between Norwegians and they are quite homogeneous. None of the many Norwegian results posted in the in the K12 thread were around 50% Sintashta. It is unlikely that there are any unmixed Norwegians, who are are 25% Albanian-like compared to average Norwegians. So we need to explain why your results for Norwegians are so different from that of the Norwegian members here. A North Russian posted also his results here and he had just 48% Sintashta and not 60% (http://www.anthrogenica.com/showthread.php?5174-Post-your-Gedmatch-calculator-Gedrosia-12-results/page34&highlight=roaring). North Russians are among Russians the most EHG/WHG shifted and except of Siberian/Uralic admixture similar to Balts. Your calculator is great but this results don't fit with the "real" results posted in the Gedrosia K12 thread. Something is off here and I just want to know the reason for that. Individual samples will of course often be different from the average for the ethnic group but not so much in my in opinion. Lithuanians can not be 30% more Sintashta-like than many Norwegians when they have just few percentages higher ANE/EHG than them. I would like to see the results of an unmixed Lithuanian but I don't think they will score significantly above 60% Sintashta and probably in many cases under 60%.

Coldmountains
09-12-2015, 10:11 PM
I run another Sintashta sample on the Gedrosia K12 calculator and i got this results.

Sintashta RISE395 Kit Number: M277797 (Bol'shekaragans)

5898

Sintashta RISE395 itself is just 50% Sintashta. So having around 50% Sintashta, what Lithuanians score in my opinion, is maybe close to being almost identical to Sintashta. I agree that Lithuanians are extremely close to Sintashta if not basically identical but all this different results are really confusing me. :confused:

Kurd
09-12-2015, 11:53 PM
This still does not explain why the results of the Norwegians were so different from this. There is not such a huge difference between Norwegians and they are quite homogeneous. None of the many Norwegian results posted in the in the K12 thread were around 50% Sintashta. It is unlikely that there are any unmixed Norwegians, who are are 25% Albanian-like compared to average Norwegians. So we need to explain why your results for Norwegians are so different from that of the Norwegian members here. A North Russian posted also his results here and he had just 48% Sintashta and not 60% (http://www.anthrogenica.com/showthread.php?5174-Post-your-Gedmatch-calculator-Gedrosia-12-results/page34&highlight=roaring). North Russians are among Russians the most EHG/WHG shifted and except of Siberian/Uralic admixture similar to Balts. Your calculator is great but this results don't fit with the "real" results posted in the Gedrosia K12 thread. Something is off here and I just want to know the reason for that. Individual samples will of course often be different from the average for the ethnic group but not so much in my in opinion. Lithuanians can not be 30% more Sintashta-like than many Norwegians when they have just few percentages higher ANE/EHG than them. I would like to see the results of an unmixed Lithuanian but I don't think they will score significantly above 60% Sintashta and probably in many cases under 60%.

Aside from the calculator effect margin of error of +/- 2%, It is quite possible that Haak used certain criteria for their samples, whether they be from certain areas, or meet certain admixture thresholds. I am not privy to any of that.

Coldmountains
09-12-2015, 11:57 PM
Aside from the calculator effect margin of error of +/- 2%, It is quite possible that Haak used certain criteria for their samples, whether they be from certain areas, or meet certain admixture thresholds. I am not privy to any of that.

Itīs a mystery i think but do you have an idea why RISE395 (Sintahshta) scores just 51% Sintashta ?. You not included this Sintashta sample in your run or ? Maybe this lower Sintashta score of RISE395 is caused by the same effect which made the Sintashta scores of the Russian and Norwegian members of this forum lower than among the Haak samples.

Kurd
09-13-2015, 12:04 AM
I run another Sintashta sample on the Gedrosia K12 calculator and i got this results.

Sintashta RISE395 Kit Number: M277797 (Bol'shekaragans)

http://www.anthrogenica.com/attachment.php?attachmentid=5898&stc=1

Sintashta RISE395 itself is just 50% Sintashta. So having around 50% Sintashta, what Lithuanians score in my opinion, is maybe close to being almost identical to Sintashta. I agree that Lithuanians are extremely close to Sintashta if not basically identical but all this different results are really confusing me. :confused:


I am not sure what criteria where used to label ancient samples Sintashta, Corded Ware, Yamnaya, etc., but I can tell you from experience with Admixture, there is sometimes huge variability in admixture profiles between ancient samples such as Sintashta, Corded Ware, etc. in fact, sometimes I wonder if samples have been mislabeled, so dont think for a moment that all Sintashta have similar admixture profiles.

In fact while doing Admixture runs I have frequently seen some Sintashta and Corded Ware split off into different clusters. We use our judgement as to which samples are problematic, and screen them out of the run. Sometimes, Inadequate marker coverage is to blame, but often the explanation is very different genetic makeup

everest59
09-13-2015, 12:08 AM
These ancient DNA samples tend to be very "erratic". It's due to the quality. Not sure about the quality of this particular Sintashta sample.

Coldmountains
09-13-2015, 12:11 AM
I am not sure what criteria where used to label ancient samples Sintashta, Corded Ware, Yamnaya, etc., but I can tell you from experience with Admixture, there is sometimes huge variability in admixture profiles between ancient samples such as Sintashta, Corded Ware, etc. in fact, sometimes I wonder if samples have been mislabeled, so dont think for a moment that all Sintashta have similar admixture profiles.

In fact while doing Admixture runs I have frequently seen some Sintashta and Corded Ware split off into different clusters. We use our judgement as to which samples are problematic, and screen them out of the run. Sometimes, Inadequate marker coverage is to blame, but often the explanation is very different genetic makeup

Sintashta 395 is actually more "northern" and "baltic" shifted than the other Sintashta samples except of RISE392. (http://www.anthrogenica.com/showthread.php?5286-Autosomal-PCA-charts-of-ancient-IE-samples-compared-to-modern-populations). You not included them in your run if i am not wrong and maybe this is reason why this Sintashta sample and the results of the forum members are so different from your Haak samples. But it is really a mystery for me

Coldmountains
09-13-2015, 12:20 AM
Aside from the calculator effect margin of error of +/- 2%, It is quite possible that Haak used certain criteria for their samples, whether they be from certain areas, or meet certain admixture thresholds. I am not privy to any of that.

I have not seen such kind of huge discrepancies between population averages and the result of forum mebers here in other oracles. This is really unusual and as far as i know Norwegians are quite homogenous so i dont expect that Haak sampled a Norwegian population very different from the Norwegian forum members here. All North Europeans so far on this forum (Poles, Norwegians,...) get this comibination 70%-75% Norwegian and 25% Anatolian/Albanian according to the oracles. This can be true for some individuals of mixed origin but not for so many forum memebers of more or less unmixed origin. I also run my Russian and Finnish gedmatch matches on the K12 calculator and the highest Sintashta score among them was just 48%

Andronovo RISE505 had also just 43% Sintashta
5902

Kurd
09-13-2015, 12:24 AM
With regards to Lithuanians very similar to Sintashta, the similarity becomes even larger if you subtract drift. A Dstat of the form D(X,Y, Sintashta,Chimp) where X represents various Lithuanians, and Y various Sintashta samples would be useful.

everest59
09-13-2015, 04:59 PM
Little bit OT, but here is a Nepalese Sherpa sample, which I analyzed from the Reich et al Indian dataset. Isn't it curious that Sherpas don't have any of the Baloch component? . The Indian admixture seems to be virtually all South Indian!
9.23% S-Indian
0.00% Baloch
0.51% Caucasian
0.00% NE-Euro
0.06% SE-Asian
11.45% Siberian
76.97% NE-Asian
1.47% Papuan
0.27% American
0.00% Beringian
0.00% Mediterranean
0.06% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African

I wonder if the South Indian is all ASI?
So the ASI did not come from let's say hill groups like bahuns or chettris.

I guess Sherpas genetically have similarity to NE Asian groups besides Tibetans?

Shaikorth
09-14-2015, 05:31 PM
I would like to see the results of an unmixed Lithuanian but I don't think they will score significantly above 60% Sintashta and probably in many cases under 60%.

There are several Gedmatch kits (from Finland and Baltic states) exceeding RISE395's Sintashta numbers. Doesn't get near 60% though. F211572 is a good example of someone with a very high baltic drift (exceeding Lithuanian Baltic component averages noticeably in Eurogenes K13 and K15) and has 54.63%.

That's an extreme case though, for instance M585533 is a Lithuanian closer to average Lithuanian Eurogenes K15 numbers and gets 49.63%.

Something like D-stat comparison between Lithuanians and RISE395 to Sintashta average would be the easiest way to figure out whether there are people more Sintashta than Sintashta alive today or if this is just calculator effect or other issues of ADMIXTURE. I reckon they won't be closer than RISE395 in formal tests.

Kurd
09-14-2015, 06:36 PM
There are several Gedmatch kits (from Finland and Baltic states) exceeding RISE395's Sintashta numbers. Doesn't get near 60% though. F211572 is a good example of someone with a very high baltic drift (exceeding Lithuanian Baltic component averages noticeably in Eurogenes K13 and K15) and has 54.63%.

That's an extreme case though, for instance M585533 is a Lithuanian closer to average Lithuanian Eurogenes K15 numbers and gets 49.63%.

Something like D-stat comparison between Lithuanians and RISE395 to Sintashta average would be the easiest way to figure out whether there are people more Sintashta than Sintashta alive today or if this is just calculator effect or other issues of ADMIXTURE. I reckon they won't be closer than RISE395 in formal tests.

Most of the Haak Lithuanians, Estonians and Russians also got over 50% as per post 551. I certainly doubt that they were all mislabeled by Haak. A more plausible explanation is that Haak screened their samples according to some geographical or admixture threshold criteria.

With regards to the calculator being off for members and not the reference samples or visa versa, that would be impossible, because that's not how ADMIXTURE works. Calculator effect of +/-2% would be the only discrepancy.

Everest mentions that Lithuanians also scored 70% CW in Haak. I used Haak Lithuanias

anyways, I'll be running some Dstats in the next couple of days to shed some light

Shaikorth
09-14-2015, 06:57 PM
Most of the Haak Lithuanians, Estonians and Russians also got over 50% as per post 551. I certainly doubt that they were all mislabeled by Haak. A more plausible explanation is that Haak screened their samples according to some geographical or admixture threshold criteria.

With regards to the calculator being off for members and not the reference samples or visa versa, that would be impossible, because that's not how ADMIXTURE works. Calculator effect of +/-2% would be the only discrepancy.

Everest mentions that Lithuanians also scored 70% CW in Haak. I used Haak Lithuanias

anyways, I'll be running some Dstats in the next couple of days to shed some light

I checked the Gedmatch Lithuanian kits against Eurogenes K13/K15 Lithuanian averages, which are based on Behar et al. 2010 Lithuanians (not included in the initial run to remove calculator effect discrepancy) and the Behar Lithuanians are exactly the same samples as Haak's Lithuanians. I don't think there is any mislabeling here either, the issue is more likely with ADMIXTURE itself.

Kurd
09-14-2015, 07:22 PM
I don't think there is any mislabeling here either, the issue is more likely with ADMIXTURE itself.

Do you mean that the Haak Lithuanian samples scored too much Sintashta on Gedrosia 12?

Shaikorth
09-14-2015, 07:38 PM
Do you mean that the Haak Lithuanian samples scored too much Sintashta on Gedrosia 12?

Not really, and besides all North Europeans on Gedmatch score less Sintashta than their counterparts in the spreadsheet so it isn't just a Lithuanian issue. I know this has been mentioned before, but the whole thing really looks like calculator effect to me. Dodecad K12b's North European component can also have over 10% swings between Gedmatch version and the spreadsheet, with the latter having higher numbers.

Kurd
09-14-2015, 08:05 PM
Not really, and besides all North Europeans on Gedmatch score less Sintashta than their counterparts in the spreadsheet so it isn't just a Lithuanian issue. I know this has been mentioned before, but the whole thing really looks like calculator effect to me. Dodecad K12b's North European component can also have over 10% swings between Gedmatch version and the spreadsheet, with the latter having higher numbers.

Do you have an example of a 10% swing. I have never seen more than a 2% calculator effect with any of my calculators

Shaikorth
09-14-2015, 08:51 PM
Do you have an example of a 10% swing. I have never seen more than a 2% calculator effect with any of my calculators

Behar Lithuanian averages of Dodecad K12b (included in the initial run so calculator effect exists in Gedmatch)
Atlantic Med 13.7
North European 77.1
South Asian 0.1
Southwest Asian 1
Caucasus 8

Gedmatch M232340
Gedrosia 1.06%
Siberian 0.35%
Southeast_Asian 0.51%
Atlantic_Med 20.34%
North_European 66.92%
South_Asian 1.96%
Caucasus 8.86%
11.52 distance to Lithuanians (Behar) in Oracle, closest match are Poles at 5

However in Eurogenes K15 (no calculator effect) M232340's distance to closest population in Oracle is under 4, and they are Lithuanians as they should be.

Kurd
09-14-2015, 09:24 PM
Behar Lithuanian averages of Dodecad K12b (included in the initial run so calculator effect exists in Gedmatch)
Atlantic Med 13.7
North European 77.1
South Asian 0.1
Southwest Asian 1
Caucasus 8

Gedmatch M232340
Gedrosia 1.06%
Siberian 0.35%
Southeast_Asian 0.51%
Atlantic_Med 20.34%
North_European 66.92%
South_Asian 1.96%
Caucasus 8.86%
11.52 distance to Lithuanians (Behar) in Oracle, closest match are Poles at 5

However in Eurogenes K15 (no calculator effect) M232340's distance to closest population in Oracle is under 4, and they are Lithuanians as they should be.

when we talk about calculator effect on admixture results we mean a difference between the sample's official results from the ADMIXTURE run vs the sample's result when ran through the calculator itself.

So for example if individual X had an official Gedrosian score of 10% from the ADMIXTURE run and a Gedrosian score of 11% when run through the calculator, the calculator effect would be 1%. The highest I've seen with my calculators is 2% . An effect of 10% is extremely unlikely

Shaikorth
09-14-2015, 09:48 PM
when we talk about calculator effect on admixture results we mean a difference between the sample's official results from the ADMIXTURE run vs the sample's result when ran through the calculator itself.

So for example if individual X had an official Gedrosian score of 10% from the ADMIXTURE run and a Gedrosian score of 11% when run through the calculator, the calculator effect would be 1%. The highest I've seen with my calculators is 2% . An effect of 10% is extremely unlikely

The effect is smaller with components an individual has less of. The 10% difference happens with components that should be very high, as you can see in the case of that Lithuanian. It's only the North European component and the Mediterranean component that show high changes in my example. I think this is what happens with North Europeans and Sintashta component.

Kurd
09-14-2015, 10:57 PM
The effect is smaller with components an individual has less of. The 10% difference happens with components that should be very high, as you can see in the case of that Lithuanian. It's only the North European component and the Mediterranean component that show high changes in my example. I think this is what happens with North Europeans and Sintashta component.

Thanks, I'll run a couple through the calculator

Dr_McNinja
09-14-2015, 11:50 PM
Little bit OT, but here is a Nepalese Sherpa sample, which I analyzed from the Reich et al Indian dataset. Isn't it curious that Sherpas don't have any of the Baloch component? . The Indian admixture seems to be virtually all South Indian!
9.23% S-Indian
0.00% Baloch
0.51% Caucasian
0.00% NE-Euro
0.06% SE-Asian
11.45% Siberian
76.97% NE-Asian
1.47% Papuan
0.27% American
0.00% Beringian
0.00% Mediterranean
0.06% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African

I wonder if the South Indian is all ASI?
So the ASI did not come from let's say hill groups like bahuns or chettris.

I guess Sherpas genetically have similarity to NE Asian groups besides Tibetans?That S-Indian could be all Papuan for all we know. I'd love to run this sample through the Oceanian version of HarappaWorld I posted in a previous page, or some of the other calculators which had split it up (HarappaWorld with no S-Indian).

Shaikorth
09-14-2015, 11:51 PM
Pinega Russian (the academic samples from that recent Balto-Slav paper) results. Looks like there is no difference in Sintashta component between South Russians and them, even though the Pinega samples are outliers compared to other North Russians such as Kargopol/HGDP. If anything, the Pinega samples average a bit more than ones in the south.


M116026 RusPinega1
S_INDIAN 0.18%
SUB_SAHARAN 0.36%
EARLY_EUROPEAN_FARMERS 26.31%
SW_ASIAN -
W_SIBERIAN 7.94%
SE_ASIAN 1.27%
BALOCHI 3.16%
SINTASHTA_STEPPE_HERDERS 48.87%
INDO_TIBETAN 0.55%
CAUCASUS 8.78%
E_AFRICAN -
E_SIBERIAN 2.58%

M515902 RusPinega9
S_INDIAN 1.22%
SUB_SAHARAN 0.25%
EARLY_EUROPEAN_FARMERS 27.12%
SW_ASIAN 0.14%
W_SIBERIAN 8.12%
SE_ASIAN 1.64%
BALOCHI 5.93%
SINTASHTA_STEPPE_HERDERS 47.47%
INDO_TIBETAN -
CAUCASUS 5.94%
E_AFRICAN -
E_SIBERIAN 2.18%

M143348 RusPinega17
S_INDIAN 1.26%
SUB_SAHARAN 0.42%
EARLY_EUROPEAN_FARMERS 25.71%
SW_ASIAN -
W_SIBERIAN 6.73%
SE_ASIAN 1.82%
BALOCHI 4.37%
SINTASHTA_STEPPE_HERDERS 49.59%
INDO_TIBETAN 0.28%
CAUCASUS 7.72%
E_AFRICAN -
E_SIBERIAN 2.11%

M462211 RusPinega20
S_INDIAN -
SUB_SAHARAN 0.13%
EARLY_EUROPEAN_FARMERS 24.20%
SW_ASIAN -
W_SIBERIAN 7.55%
SE_ASIAN 1.86%
BALOCHI 5.68%
SINTASHTA_STEPPE_HERDERS 47.10%
INDO_TIBETAN 0.73%
CAUCASUS 9.45%
E_AFRICAN -
E_SIBERIAN 3.29%

FenriR
09-14-2015, 11:56 PM
The "calculator effect" was quite an issue like two years ago, especially regarding the Oracle-results: since the scores were a bit off, people of Eastern European origin were often pulled towards more western populations (they got less in their "basic", northeastern, components), e.g. Russians got "Sorbs" as their closest matches in some calcs =)
And the only way to understand your results was to compare them with other people, who werenīt part of a project.

tamilgangster
09-15-2015, 06:44 AM
Little bit OT, but here is a Nepalese Sherpa sample, which I analyzed from the Reich et al Indian dataset. Isn't it curious that Sherpas don't have any of the Baloch component? . The Indian admixture seems to be virtually all South Indian!
9.23% S-Indian
0.00% Baloch
0.51% Caucasian
0.00% NE-Euro
0.06% SE-Asian
11.45% Siberian
76.97% NE-Asian
1.47% Papuan
0.27% American
0.00% Beringian
0.00% Mediterranean
0.06% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African

I wonder if the South Indian is all ASI?
So the ASI did not come from let's say hill groups like bahuns or chettris.

I guess Sherpas genetically have similarity to NE Asian groups besides Tibetans?

THere are many SE asian populations that have SE asian but lack baloch. But what especially strange about the sherpa sample is that it ALSO lacks the SE asian component. Nysha and tibetans also have South Indian, but compleltey lack either baloch or SE asian, both of which are components usually found admixed with South Indian

bored
09-15-2015, 07:03 AM
THere are many SE asian populations that have SE asian but lack baloch. But what especially strange about the sherpa sample is that it ALSO lacks the SE asian component. Nysha and tibetans also have South Indian, but compleltey lack either baloch or SE asian, both of which are components usually found admixed with South Indian

What. Why would they have SE Asian? There are many NE Asian populations that have NE Asian but lack Baloch as well. The "strangest" thing about it is that Sherpas live right beside Indic people and yet lack the Baloch component.

tamilgangster
09-15-2015, 03:32 PM
What. Why would they have SE Asian? There are many NE Asian populations that have NE Asian but lack Baloch as well. The "strangest" thing about it is that Sherpas live right beside Indic people and yet lack the Baloch component.
So do many adivasi populations, they also lack baloch and are surrounded by other indid populations, Sherpas are fringe groups. The lack of SE asian is strange because Even northern han Chinese have SE asian.

tamilgangster
09-17-2015, 07:16 AM
Pinega Russian (the academic samples from that recent Balto-Slav paper) results. Looks like there is no difference in Sintashta component between South Russians and them, even though the Pinega samples are outliers compared to other North Russians such as Kargopol/HGDP. If anything, the Pinega samples average a bit more than ones in the south.


M116026 RusPinega1
S_INDIAN 0.18%
SUB_SAHARAN 0.36%
EARLY_EUROPEAN_FARMERS 26.31%
SW_ASIAN -
W_SIBERIAN 7.94%
SE_ASIAN 1.27%
BALOCHI 3.16%
SINTASHTA_STEPPE_HERDERS 48.87%
INDO_TIBETAN 0.55%
CAUCASUS 8.78%
E_AFRICAN -
E_SIBERIAN 2.58%

M515902 RusPinega9
S_INDIAN 1.22%
SUB_SAHARAN 0.25%
EARLY_EUROPEAN_FARMERS 27.12%
SW_ASIAN 0.14%
W_SIBERIAN 8.12%
SE_ASIAN 1.64%
BALOCHI 5.93%
SINTASHTA_STEPPE_HERDERS 47.47%
INDO_TIBETAN -
CAUCASUS 5.94%
E_AFRICAN -
E_SIBERIAN 2.18%

M143348 RusPinega17
S_INDIAN 1.26%
SUB_SAHARAN 0.42%
EARLY_EUROPEAN_FARMERS 25.71%
SW_ASIAN -
W_SIBERIAN 6.73%
SE_ASIAN 1.82%
BALOCHI 4.37%
SINTASHTA_STEPPE_HERDERS 49.59%
INDO_TIBETAN 0.28%
CAUCASUS 7.72%
E_AFRICAN -
E_SIBERIAN 2.11%

M462211 RusPinega20
S_INDIAN -
SUB_SAHARAN 0.13%
EARLY_EUROPEAN_FARMERS 24.20%
SW_ASIAN -
W_SIBERIAN 7.55%
SE_ASIAN 1.86%
BALOCHI 5.68%
SINTASHTA_STEPPE_HERDERS 47.10%
INDO_TIBETAN 0.73%
CAUCASUS 9.45%
E_AFRICAN -
E_SIBERIAN 3.29%


Im surprised that SOuth Russians have that much siberian admixture. WHich group was used as the proxy for the west siberian component,

Coldmountains
09-17-2015, 07:29 AM
Im surprised that SOuth Russians have that much siberian admixture. WHich group was used as the proxy for the west siberian component,

Pinega is in Northwest Russia (Arkhangelsk Oblast). Their high west Siberian component is not surprising because North Russia was mainly populated by Finno-Ugrian tribes prior to the Slavic migration. Many of this Finno-Ugrian tribes were assimilated so North Russians are outliners among Slavs/East Slavs.

tamilgangster
09-17-2015, 05:16 PM
Pinega is in Northwest Russia (Arkhangelsk Oblast). Their high west Siberian component is not surprising because North Russia was mainly populated by Finno-Ugrian tribes prior to the Slavic migration. Many of this Finno-Ugrian tribes were assimilated so North Russians are outliners among Slavs/East Slavs.

That makes sense, I wouldn't say theY are outliers, Its very likely that many people who identify as ethnic Russians have admixture from minority groups that live in their respective areas

everest59
09-18-2015, 12:30 AM
Are Subbas in the Reich dataset Nepalese Limbus? I would have thought Limbus would have more South Asian.

8.52% S-Indian
0.00% Baloch
0.00% Caucasian
0.00% NE-Euro
3.50% SE-Asian
10.38% Siberian
75.98% NE-Asian
1.55% Papuan
0.00% American
0.06% Beringian
0.00% Mediterranean
0.00% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African

everest59
09-18-2015, 12:49 AM
Kashmiri Pandits in sheet 4:

https://docs.google.com/spreadsheets/d/1yroLfxb7NvkZ63tMJcRm-4yNH8PcXtzhJWm1Hpfr5-8/edit#gid=1750165129

BMG
09-18-2015, 02:09 AM
Hi Everest can you do more ITU and STU samples if you have time .

Skyfall
09-28-2015, 07:45 PM
Everest, could you answer the following questions for me, I'd greatly appreciate your input here, if you have time:

I know you are an expert on these things, so I'd greatly appreciate it if you could answer a few questions for me. I am kinda lost and don't really understand where our ancestry (as a Jat) comes from. All I know is that we are Bhatti Jats. I'd greatly appreciate your help with the following:

As I understand it, ANI is a hybrid population that is mostly composed of West Eurasian population groups, namely Europeans and West Asians. ASI, on the other hand, has an indigenous component. My questions are the following:

1) Does ASI have anything to do with Australoid populations/the Onge/Negritos? I really need this clarified. From what I have read, only isolated tribal groups have some input from Onge-like populations, but the mainland caste populations have no such input. Then why do certain agenda-drive websites keep parroting the same nonsense of all South Asians being admixed with Onge/Australoid/Negrito populations? If ASI is so far separated from the Onge that is is laughable to suggest a link between the two. Lets take a look at the evidence:

There is no evidence whatsoever to suggest that ASI or the South Indian component in Indians is related to modern-day Australoids or even Negritos. These are the fst distances, the most widely used measure of genetic distance between populations, between ASI and other populations:

Caucasian: 0.077
Baloch: 0.08
NE Asian: 0.081
NE Euro: 0.082
SE Asian: 0.084
SW Asian: 0.091
Siberian: 0.093
Mediterranean: 0.095
Beringian: 0.116
E African: 0.122
American: 0.128
W African: 0.142
Papuan: 0.145
Pygmy: 0.188
San: 0.203
BTW, Here are the Fst distances for your perusal:

https://docs.google.com/spreadsheet/...WZzNBMEE#gid=2

If one actually reads this fst distance spreadsheet I posted above (with data from Reich et.al and other Harvard geneticists), it is clear that the South Indian/ASI component is closest to Gedrosia (at 0.081) followed by Caucasus (at 0.082) and East Asian (at 0.085) and Northern European at (0.086). This clearly shows that it’s actually closer to Gedrosia and Caucasus than the East Asian components. Again, the component is closest to Caucasian, Baloch, NE Asian, NE Euro and SE Asian in that order. So its closer to Caucasian populations, followed by a Mongoloid South-East Asian population, followed by a North-Eastern European population.

In other words, the ASI/South Indian component is actually closer to Caucasian populations than even Mongoloid populations, and it is nowhere near close to Australoid populations. In fact, it’s even closer to North Eastern Europeans than it is to Australoids and closer to West and East African than the Papuan component! So how can one link ASI to Australoid? I keep reading that South Asians are all admixed with Australoid, and it is really frustrating to read that when it is patently false. Even the so-called tribes of India are not all admixed with the Australoid component, and have more in common with each other than they do with the Onge: http://www.ncbi.nlm.nih.gov/pubmed/6479999
So what is the truth here?

2) What % of ASI is Caucasian West Eurasian in nature? On the Harappa Ancestry Project, the "South Indian" component is 50% West Eurasian, and according to this spreasheet: https://docs.google.com/spreadsheets...dd4/edit?pli=1 castes like the "Bhatia" are only 25% South Indian, which would mean that they are only 12.5% ASI -- the question is what % of this ASI is Caucasian in nature? I just want a breakdown of the South Indian and ASI into Caucasian/Non-Caucasian components. And is this "South Indian" breakdown correct?

3) Is ANI completely Caucasian in nature, or does it also possess some non-Caucasian elements? I'd appreciate a breakdown for ANI as well.

4) Do upper-caste NW Indians and South Asians have enough of a Caucasian genetic component that they can be classified as "Caucasian" in terms of which racial branch they belong to, in the grand sense?

5) If you had to give me a rough % breakdown of how Caucasian a Jatt or a Bhatia that is 25-27% "South Indian" is what would be a rough estimate at this point in time? I know there is a lot of speculation at this point, but we can still hypothesize to a certain extent, and I'd greatly appreciate it if you could perhaps give me a rough number/% for how Caucasian an upper-caste NW Indian really is.

6) On 23andme, there are many Upper-caste NW Indians who score in the range of 60-95% European, and their Central West Eurasian Caucasoid component is 70-90% in NW India and 50-70 in south india. Like this man: http://i.imgur.com/jVT8gxb.jpg -- if this is the case, then why aren't more people acknowledging this fact, and why aren't upper-caste NW Indians considered Caucasoid? Are these 23andme results anomalies, or incorrect? I'd appreciate an explanation.

7) ASI is just like the paleolithic ANE influence in Europeans, correct? And since ANE is 45% East-Asian/Native American in composition, and 10% ASE, then wouldn't that mean that Europeans are also admixed with non-Caucasoid populations? Or are Europeans pure Caucasians? I know that Europeans like Northern Russians and Finns have substantial recent East Eurasian admixture (>10%) but wouldn't their ANE genetic component make them even more admixed/Non-Caucasoid? I'd appreciate some information on this.

Thanks a bunch!

ashwinb
09-28-2015, 10:20 PM
skyfall, have you done the K6-K8 test, you get % in terms of ANE,ENF,ASE etc components. In case you havent seen any results, https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=58

Ignis90
09-29-2015, 06:05 AM
...

I didn't bother reading seriously when I saw "caucasoid", "mongoloid", "autraloid", "negroid". These terms have nothing to do with this topic and are actually an effective mental handicap when trying to understand any related topic in human population genetics.

Also, there is no need to make a "who's more caucasoid" contest by blatantly making West Eurasians and North Africans multiracial. I guess from a South Asian point of view it makes sense because South Asians have always been seen as a mixed and intermediate group of populations (between West and East Eurasia). The latter point is now completely obsolete since Lazaridis et al. and the Mal'ta Buret sample put a word on the long known Amerindian-like influence in Europeans and West Asians - Ancient North Eurasians - and introduced a new and mysterious population - Basal Eurasians- to explain why modern West Eurasians share less drift with East Eurasians than ancient West Eurasians do.

So yes, South Asians are as mixed as any other modern populations because the Holocene (and earlier) has been the scene of many migrations and mixes of genetically divergent populations. And these ancient populations didn't necessarily follow a modern racial pattern, in other terms, using modern humans and the vague racial classification applied to them by some people and trying to make ancient populations fit into this racial boxes is highly inaccurate.

Skyfall
09-29-2015, 03:31 PM
Wait, are you stating that North Africans and Middle Easterners and West Asians are not multiracial? Correct me if I'm mistaken, but I believe that many studies have shown Northern Africans to be highly admixed with SSA, and in countries like Tunisia, this admix reaches >20%, not to mention Egypt and others. In the Middle East, I've read studies that indicate much the same thing. And I recently read a study that the average Turkish individual was 15% East Asian, etc. So why is it a problem to point out the fact that all populations are highly admixed, not just South Asians? I'm not trying to make this into a contest, I just want the facts.

I expected a more open-minded approach on this forum, since on previous forums, people have refused to acknowledge the fact that all populations these days are more or less mixed, not just South Asians. It was particularly a bone of contention for them when I pointed out the admixture in Western Eurasian and Northern African populations. I just want a clear discussion of just how admixed we all are, because I don't want to be in the dark about these things. We as humans need to recognize that racial purity is a myth. So I apologize if I offended you, but I really would like to know about these things.

Also, since I'm a novice at these things and you are much more experienced, I would greatly appreciate it if you could expand on this statement:

"The latter point is now completely obsolete since Lazaridis et al. and the Mal'ta Buret sample put a word on the long known Amerindian-like influence in Europeans and West Asians - Ancient North Eurasians - and introduced a new and mysterious population - Basal Eurasians- to explain why modern West Eurasians share less drift with East Eurasians than ancient West Eurasians do."

This is similar to what I stated in my long essays above. (I apologize if they sounded anti-intellectual in any way, I am learning the ropes here) But essentially, what you are stating is that Basal Eurasians explain why Western Eurasians share less drift with Eastern Eurasians, and that because of this fact, South Asians are not that divergent from/ not intermediates between/ or much more admixed than Western Eurasians after all? I would greatly appreciate it if you could give me a more parsimonious explanation of this.

Ignis90
09-29-2015, 03:41 PM
I think you should read my text again. Basically, get rid of modern and obsolete racial classifcation terms and understand that 1) the ancestral populations did not often look like modern populations which descend from them and 2) that closely related populations can look drastically different while more genetically distant populations can look similar.

It should be enough to understand now.

Skyfall
09-29-2015, 03:49 PM
Sorry again, but could you expand on this particular statement: "that closely related populations can look drastically different while more genetically distant populations can look similar."

and also this:

"The latter point is now completely obsolete since Lazaridis et al. and the Mal'ta Buret sample put a word on the long known Amerindian-like influence in Europeans and West Asians - Ancient North Eurasians - and introduced a new and mysterious population - Basal Eurasians- to explain why modern West Eurasians share less drift with East Eurasians than ancient West Eurasians do."

I corresponded with Razib about the Amerindian-like influence in Europeans and West Asians, and he stated that this wasn't really an admixture with an Eastern Eurasian population, but something entirely different. Essentially, would it be safe to say that this influence is equivalent to East Asian admixture? He even wrote an article about the same pointing out how NEuros got to 10-18% Amerinidian: http://blogs.discovermagazine.com/gnxp/2012/09/across-the-sea-of-grass-how-northern-europeans-got-to-be-10-northeast-asian/#.Vgq7N844LYs

Also, where does that leave South Asians? And what you are stating is that Basal Eurasians explain why Western Eurasians share less drift with Eastern Eurasians, and that because of this fact, South Asians are not that divergent from/ not intermediates between/ or much more admixed than Western Eurasians after all? I would greatly appreciate it if you could give me a more parsimonious explanation of this.

Dr_McNinja
10-02-2015, 01:18 PM
ASI and "South Indian" are two entirely separate things.

ASI stands for Ancestral South Indian and refers to the first residents of South India (including the coasts presumably). This would have branched off very early and be somewhat related to (having branched off from a point near to) Papuan/Oceanian, and most likely, Onge. Perhaps someone can link you to some of the treemix analyses which show their relative positions.

"South Indian" in calculators is a hybrid component resulting from modern genetic drift and a lack of ancient South Eurasian genetic samples. It's not real, it doesn't really mean anything except perhaps "Ancient India" (both North and South). The reason this "Ancient India" component is high in the South and thus dubbed "South Indian" is because the other parts of the Indian subcontinent have more recent foreign admixture from West and Central Asia.

It (the "South Indian" component you see in calculators) is actually a combination of ASI (represented by Eurogenes K8 components "ASE", "Oceanian", and "East Eurasian") and Baloch (mostly Caucasus-type), on top of whatever Baloch/Gedrosian is registering separately.

All populations are highly admixed when looked at through ancient DNA calculators like Eurogenes K7 or K8. The closest to "pure" I can think of might be Bedouin Arabs who have 80-90% of the "ENF" component there, and Japanese who have 80+% of the "East Eurasian" component there (but to be fair, the calculator was not meant to distinguish between that branch of the Eurasian tree as much).

"ANI" needs to be defined first. I prefer ANI, due to the "Ancestral" part of the name, to refer to the ancient West/Southwest Asian contribution to the region. So that would just be Gedrosia-Caucasus admixture. More recent North Indian admixture comes from Indo-Europeans not that long ago, who did carry some Caucasian admixture with them too, but usually registers as European on calculators.

redifflal
10-02-2015, 03:28 PM
ASI and "South Indian" are two entirely separate things.

ASI stands for Ancestral South Indian and refers to the first residents of South India (including the coasts presumably). This would have branched off very early and be somewhat related to (having branched off from a point near to) Papuan/Oceanian, and most likely, Onge. Perhaps someone can link you to some of the treemix analyses which show their relative positions.

"South Indian" in calculators is a hybrid component resulting from modern genetic drift and a lack of ancient South Eurasian genetic samples. It's not real, it doesn't really mean anything except perhaps "Ancient India" (both North and South). The reason this "Ancient India" component is high in the South and thus dubbed "South Indian" is because the other parts of the Indian subcontinent have more recent foreign admixture from West and Central Asia.

It (the "South Indian" component you see in calculators) is actually a combination of ASI (represented by Eurogenes K8 components "ASE", "Oceanian", and "East Eurasian") and Baloch (mostly Caucasus-type), on top of whatever Baloch/Gedrosian is registering separately.

All populations are highly admixed when looked at through ancient DNA calculators like Eurogenes K7 or K8. The closest to "pure" I can think of might be Bedouin Arabs who have 80-90% of the "ENF" component there, and Japanese who have 80+% of the "East Eurasian" component there (but to be fair, the calculator was not meant to distinguish between that branch of the Eurasian tree as much).

"ANI" needs to be defined first. I prefer ANI, due to the "Ancestral" part of the name, to refer to the ancient West/Southwest Asian contribution to the region. So that would just be Gedrosia-Caucasus admixture. More recent North Indian admixture comes from Indo-Europeans not that long ago, who did carry some Caucasian admixture with them too, but usually registers as European on calculators.

Is there a word on whether ANI was present in the subcontinent before Indo-Europeanization?

ashwinb
10-02-2015, 06:43 PM
I read somewhere on the forum that ANE is present in significant amounts in tribals(And that most indians score about 30% on ANE), have the Euroegenes K6-K8 tests been done on tribal populations? Any links ?

Varun R
10-02-2015, 08:14 PM
Ashwin,
If I remember correctly, David did test Paniya using K8. I seem to remember a figure of somewhere on the order of 25 percent ANE

ashwinb
10-03-2015, 12:59 PM
Oh,so all south asians have on an average 25%+ ANE, with an average of 30%? And none seemed to have anything more than 31% I guess.

parasar
10-03-2015, 02:41 PM
Ashwin,
If I remember correctly, David did test Paniya using K8. I seem to remember a figure of somewhere on the order of 25 percent ANE

I have my doubts on the Paniya being 25% ANE. They look like a pre-Ust Ishim split, remaining almost Basal like due to isolation.
Though perhaps I am being misled by the samples as everest59 had mentioned there was some potential mislabeling.

Varun R
10-03-2015, 02:52 PM
Parasar,
Is that right? Hm, well, I don't remember anything strange about the Paniya sample results apart from high ANE...
I am increasingly starting to believe that a very ANE rich (perhaps ANE exclusive?) population migrated to the subcontinent prior to the Neolithic. Of course, there is no concrete evidence for this as of yet, so it's merely a speculation based off admixture and D-Stat runs....

Megalophias
10-03-2015, 03:58 PM
I have my doubts on the Paniya being 25% ANE. They look like a pre-Ust Ishim split, remaining almost Basal like due to isolation.
Though perhaps I am being misled by the samples as everest59 had mentioned there was some potential mislabeling.
South Asians have a tendency to be equidistant between WHG and ANE, or nearly so, like East Asians. But unlike East Asians they are also share lots of drift with WHG/UHG/ANE.

Equally related to WHG + ANE but closer to them than East Asians = ancestry from a "West Eurasian" branch basal to WHG+ANE. (Not basal to Ust'-Ishim though.)

Or just happen to be mixed with WHG/UHG and ANE in roughly equal proportions, of course.

But I haven't seen the stats for many South Indians or tribals, so this might only apply to the northwest + Gujaratis and Bengalis.

Dr_McNinja
10-03-2015, 06:06 PM
Is there a word on whether ANI was present in the subcontinent before Indo-Europeanization?It had to have been since there's that Gedrosian admixture layer centered in Balochistan, which is in the "north", so to speak, and radiates to the entire subcontinent and was there since probably the Neolithic or earlier.

Kale
10-05-2015, 06:20 PM
I think the ANE in South Asians may be inflated by a third branch of Eurasians that contributed to MA1, Central Asians, and South Asians. In reality South Asians may be only 15-25% 'West Eurasian'.

MonkeyDLuffy
10-05-2015, 06:39 PM
I think the ANE in South Asians may be inflated by a third branch of Eurasians that contributed to MA1, Central Asians, and South Asians. In reality South Asians may be only 15-25% 'West Eurasian'.

So what's the other 85%? East asian?

Megalophias
10-05-2015, 07:39 PM
So what's the other 85%? East asian?
We could call it CHG or ACE maybe. But it isn't East Eurasian, it would either be Basal West Eurasian or a sister branch to both East and West Eurasian.

Sapporo
10-05-2015, 08:37 PM
So what's the other 85%? East asian?

Yes, I'm perplexed as well? If we are 85% non West Eurasian and either South Eurasian or East Eurasian, we would either cluster with East Eurasians or off in some isolated position on PCA plots. Not on a cline toward Central Asia/Europe.

Whatever it is, is has to be on a branch closer to one of East or West Eurasia. I don't think our ancestry is that independent of both West and East Eurasia.

Kale
10-06-2015, 01:09 PM
So what's the other 85%? East asian?

Some is the aforementioned third branch. Some basal Eurasian. Definitely some East Eurasian (be it from ASI or more recent arrivals).

MonkeyDLuffy
10-06-2015, 05:16 PM
Some is the aforementioned third branch. Some basal Eurasian. Definitely some East Eurasian (be it from ASI or more recent arrivals).

ASI won't be east eurasian to be exact. More like South eurasian. Enf and whg alone makes south asians approximately 40-50% west eurasian (if you don't count the ANE). The three main components of South asian genetic mixture are ENF, ASI and ANE. East eurasian is not much (4% in my case). ASI is what makes SAs cluster a lil asay from West eurasians. Like SSA pulls North africans a little away from europeans.

kenji.aryan
10-06-2015, 05:19 PM
ASI won't be east eurasian to be exact. More like South eurasian. Enf and whg alone makes south asians approximately 40-50% west eurasian (if you don't count the ANE). The three main components of South asian genetic mixture are ENF, ASI and ANE. East eurasian is not much (4% in my case). ASI is what makes SAs cluster a lil asay from West eurasians. Like SSA pulls North africans a little away from europeans.

On 23andme East eurasian does pulls some south asians further south and there is also no way of telling how much south eurasian we have from 23andme.

MonkeyDLuffy
10-06-2015, 05:30 PM
On 23andme East eurasian does pulls some south asians further south and there is also no way of telling how much south eurasian we have from 23andme.

23&me plot is garbage for SAs. David's plot is much better. I'm relying on K7 and K8 right now.

bored
10-06-2015, 05:49 PM
23&me plot is garbage for SAs. David's plot is much better. I'm relying on K7 and K8 right now.

23andme's PCA plot is apparently extra sensitive to East Eurasian and pulls you too far east if you have even a little bit.

Kale
10-06-2015, 06:03 PM
ASI won't be east eurasian to be exact. More like South eurasian. Enf and whg alone makes south asians approximately 40-50% west eurasian (if you don't count the ANE). The three main components of South asian genetic mixture are ENF, ASI and ANE. East eurasian is not much (4% in my case). ASI is what makes SAs cluster a lil asay from West eurasians. Like SSA pulls North africans a little away from europeans.

Look at the sharing between Sakilli and other groups though...
MbutiPygmy Sakilli Stuttgart Ust'-Ishim -0.0184 -4.260
MbutiPygmy Sakilli La_Brana-1 Ust'-Ishim -0.0211 -4.456
MbutiPygmy Sakilli Papuan Ust'-Ishim -0.0244 -5.877
MbutiPygmy Sakilli Motala12 Ust'-Ishim -0.0317 -7.016
MbutiPygmy Sakilli Loschbour Ust'-Ishim -0.0318 -7.281
MbutiPygmy Sakilli Dai Ust'-Ishim -0.0467 -12.846
MbutiPygmy Sakilli MA-1 Ust'-Ishim -0.0501 -9.609

Obviously MA-1 ancestry is pretty significant. But not terribly, considering even Bedouins and Yemenite Jews score -0.519 and -0.524 respectively. Dai is very close behind (granted they probably have a bit of ASI to inflate that). But I mean, Stuttgart and La Brana behind Papuan? ENF and WHG ancestry should be much more recent than any Papuan relation as well, which makes their contributions even more pitifully miniscule.

MonkeyDLuffy
10-06-2015, 06:44 PM
Look at the sharing between Sakilli and other groups though...
MbutiPygmy Sakilli Stuttgart Ust'-Ishim -0.0184 -4.260
MbutiPygmy Sakilli La_Brana-1 Ust'-Ishim -0.0211 -4.456
MbutiPygmy Sakilli Papuan Ust'-Ishim -0.0244 -5.877
MbutiPygmy Sakilli Motala12 Ust'-Ishim -0.0317 -7.016
MbutiPygmy Sakilli Loschbour Ust'-Ishim -0.0318 -7.281
MbutiPygmy Sakilli Dai Ust'-Ishim -0.0467 -12.846
MbutiPygmy Sakilli MA-1 Ust'-Ishim -0.0501 -9.609

Obviously MA-1 ancestry is pretty significant. But not terribly, considering even Bedouins and Yemenite Jews score -0.519 and -0.524 respectively. Dai is very close behind (granted they probably have a bit of ASI to inflate that). But I mean, Stuttgart and La Brana behind Papuan? ENF and WHG ancestry should be much more recent than any Papuan relation as well, which makes their contributions even more pitifully miniscule.

ASI is the oldest component of South Asians, but there is no way ENF and WHG are recent, pretty sure ENF and WHG were already present in IVC population. Maybe in lesser percentage than the modern population, as there were lots of waves of migration in SA. ASI hangs between East and west eurasian, pulling more towards East Eurasian for sure.

Here is my 23&me plot, the ASI is enough to pull me away from West eurasian, but its still much closer to West eurasians than East asian population, stating that SA have much more than 15% west eurasian.

http://i.imgur.com/R6L4vTZ.jpg

Here is the PCA david made for me

http://i.imgur.com/Brw1Q54.png

Megalophias
10-06-2015, 09:42 PM
Look at the sharing between Sakilli and other groups though...
MbutiPygmy Sakilli Stuttgart Ust'-Ishim -0.0184 -4.260
MbutiPygmy Sakilli La_Brana-1 Ust'-Ishim -0.0211 -4.456
MbutiPygmy Sakilli Papuan Ust'-Ishim -0.0244 -5.877
MbutiPygmy Sakilli Motala12 Ust'-Ishim -0.0317 -7.016
MbutiPygmy Sakilli Loschbour Ust'-Ishim -0.0318 -7.281
MbutiPygmy Sakilli Dai Ust'-Ishim -0.0467 -12.846
MbutiPygmy Sakilli MA-1 Ust'-Ishim -0.0501 -9.609

Obviously MA-1 ancestry is pretty significant. But not terribly, considering even Bedouins and Yemenite Jews score -0.519 and -0.524 respectively. Dai is very close behind (granted they probably have a bit of ASI to inflate that). But I mean, Stuttgart and La Brana behind Papuan? ENF and WHG ancestry should be much more recent than any Papuan relation as well, which makes their contributions even more pitifully miniscule.

Thanks for those stats, Kale.

Significantly closer to MA-1 than to Loschbour, which indicates that the relation can't be solely due to a basal West Eurasian branch.

For Chamar, who are also scheduled caste but from North India instead of South India:

MbutiPygmy Chamar Stuttgart Ust'-Ishim -0.0233 -5.999
MbutiPygmy Chamar Papuan Ust'-Ishim -0.0244 -6.400
MbutiPygmy Chamar Loschbour Ust'-Ishim -0.0370 -9.244
MbutiPygmy Chamar Dai Ust'-Ishim -0.0481 -14.298
MbutiPygmy Chamar MA-1 Ust'-Ishim -0.0599 -12.722

Overall rather close, but they are slightly closer to almost everyone than the Sakilli are.

For comparison
MbutiPygmy Ket Stuttgart Ust'-Ishim -0.0213 -4.666
MbutiPygmy Ket Papuan Ust'-Ishim -0.0241 -5.496
MbutiPygmy Karitiana Loschbour Ust'-Ishim -0.0357 -6.228
MbutiPygmy Papuan Dai Ust'-Ishim -0.0476 -9.469
MbutiPygmy BedouinB MA-1 Ust'-Ishim -0.0519 -11.485

They are about as close to EEF as Kets from Central Siberia, about as close to WHG as Amazonian Indians, about as close to East Asian as Papuans, and about as close to ANE as Bedouins.

Pretty much their own thing!

bored
10-06-2015, 09:47 PM
U
Thanks for those stats, Kale.

Significantly closer to MA-1 than to Loschbour, which indicates that the relation can't be solely due to a basal West Eurasian branch.

For Chamar, who are also scheduled caste but from North India instead of South India:

MbutiPygmy Chamar Stuttgart Ust'-Ishim -0.0233 -5.999
MbutiPygmy Chamar Papuan Ust'-Ishim -0.0244 -6.400
MbutiPygmy Chamar Loschbour Ust'-Ishim -0.0370 -9.244
MbutiPygmy Chamar Dai Ust'-Ishim -0.0481 -14.298
MbutiPygmy Chamar MA-1 Ust'-Ishim -0.0599 -12.722

Overall rather close, but they are slightly closer to almost everyone than the Sakilli are.

There were probably all kinds of movements both into and out of India leading to quite a tangle of different kinds of ancestry. We need more sophisticated methods to sort it out, or better yet aDNA.

How do you read these?

Megalophias
10-06-2015, 10:02 PM
Do you mean how do I interpret them, or how do you read D-statistics in general?

bored
10-06-2015, 10:47 PM
Do you mean how do I interpret them, or how do you read D-statistics in general?

Yes, I mean how do I interpret them.

Megalophias
10-07-2015, 12:57 AM
Yes, I mean how do I interpret them.
The first number is the actual D score, which tells you how close the populations are to each other; the second number is the Z score, which tells you how significant the result is. A Z score of more than 2 (or less than -2 if it is negative) usually means something, a Z score less than 1 (or more than -1 if negative) is probably noise.

If you have a D stat (MbutiPygmy, Chamar; Dai, Ust'-Ishim) it means that the computer looked at each site in the genome (that you have coverage for) and checks if Mbuti and Chamar are different there, and Dai and Ust'-Ishim are different there. Then if Mbuti and Dai go one way (they both have more of one allele), and Chamar and Ust'-Ishim go the other way (they both have less of that allele), then it adds to the D score. If Mbuti and Ust'-Ishim go one way, and Chamar and Dai go the other way, then it subtracts from the D score. So the closer Chamar is to Ust'-Ishim, and the closer Mbuti is to Dai, the more positive, and the closer Chamar is to Dai, and the closer Mbuti is to Ust'-Ishim, the more negative the score.

We are using Mbuti because they split off from Eurasians a long time ago and have had very little contact with them since then, so they are a good neutral outgroup. They have had genetic drift since then, which at some sites will make them randomly more like Dai, and at other sites randomly more like Ust'-Ishim, but because it is random over thousands of sites it all evens out to very nearly zero. So we don't have to worry about Mbuti being closer to anyone (hopefully) and we are just looking at who Chamar is closest to.

We are using Ust'-Ishim because he is so old that he is about equally related to all Eurasians (except he is a bit further from people with Basal Eurasian), so we are looking at more of whatever makes the different branches of Eurasians distinct and less of what they all share.

The nice thing about D statistics is that they don't care about drift that is very specific to a population. That gives high Fst and populations that hog ADMIXTURE components (like Kalash), but to a D statistic it all cancels out, so we only see the drift that is shared between the populations being compared. So when we compare (Mbuti, Chamar; Dai, Ust'-Ishim) we are ignoring the drift that Mbuti and Chamar share (the ancestry shared by all modern humans) and the drift that Dai and Ust'-Ishim share (the ancestry shared by all non-Basal Eurasians), and all the drift specific to the individual populations. We just looking at the drift that Dai experienced after splitting from its most recent common ancestor with Ust'-Ishim (general East Eurasian + East Asian + Dai specific), and seeing how much of that Chamar has (probably mostly just the general East Eurasian part).

Kale
10-07-2015, 01:43 PM
1) ASI is the oldest component of South Asians, but there is no way ENF and WHG are recent, pretty sure ENF and WHG were already present in IVC population.
2) ASI hangs between East and west eurasian, pulling more towards East Eurasian for sure.
2a) Here is my 23&me plot, the ASI is enough to pull me away from West eurasian, but its still much closer to West eurasians than East asian population, stating that SA have much more than 15% west eurasian.

Here is the PCA david made for me...

1) Considering ASI may be ~50,000 years old...I'd say neolithic (say what, 8,000 years or so) is comparatively recent :P
2) ASI may be just a tad younger than the East Eurasian - West Eurasian split. Maybe drifting 5,000 years or less with the East Eurasian branch.
2a) Because of 2, South Asian should be in it's own dimension. You can plot 3 divergent branches on a 2 dimensional plot, but to plot 4 populations, you need 3 dimensions. Any Eurasian plot you see, will have Han/Dai/Japanese in one corner, Europeans in corner two, and South Asians in corner 3...with the tip of the south Asian corner just baaarrreeellly closer to West Eurasians.

MonkeyDLuffy
10-08-2015, 02:22 AM
1) Considering ASI may be ~50,000 years old...I'd say neolithic (say what, 8,000 years or so) is comparatively recent :P
2) ASI may be just a tad younger than the East Eurasian - West Eurasian split. Maybe drifting 5,000 years or less with the East Eurasian branch.
2a) Because of 2, South Asian should be in it's own dimension. You can plot 3 divergent branches on a 2 dimensional plot, but to plot 4 populations, you need 3 dimensions. Any Eurasian plot you see, will have Han/Dai/Japanese in one corner, Europeans in corner two, and South Asians in corner 3...with the tip of the south Asian corner just baaarrreeellly closer to West Eurasians.

Can you post the plot?

tamilgangster
10-08-2015, 02:34 AM
Ashwin,
If I remember correctly, David did test Paniya using K8. I seem to remember a figure of somewhere on the order of 25 percent ANE

There hasn't been a proper krun on paniya by eurogenes k7, pirimalai kallars who have 63% south indian and 30% baloch have 30% ANE on eurogenes k7.

tamilgangster
10-08-2015, 02:39 AM
Parasar,
Is that right? Hm, well, I don't remember anything strange about the Paniya sample results apart from high ANE...
I am increasingly starting to believe that a very ANE rich (perhaps ANE exclusive?) population migrated to the subcontinent prior to the Neolithic. Of course, there is no concrete evidence for this as of yet, so it's merely a speculation based off admixture and D-Stat runs....

It has to be because how else would you explain the fact that their is higher ANE among ASI shifted populations than among baloch. It could possibly be a proxy error since the proxy for ANE is the "non east eurasian" portion of karitiana, which might actually contain oceanic/australoid admixture.

tamilgangster
10-08-2015, 02:52 AM
I think the ANE in South Asians may be inflated by a third branch of Eurasians that contributed to MA1, Central Asians, and South Asians. In reality South Asians may be only 15-25% 'West Eurasian'.

Most south asians score 25-40% ANE, so that has to be wrong. South asians plot closest to west eurasian, compared to either east asians or ocenians

tamilgangster
10-08-2015, 02:55 AM
On 23andme East eurasian does pulls some south asians further south and there is also no way of telling how much south eurasian we have from 23andme.

ASI shifted populations in general are more east eurasian shifted but 23 and me as we all know is shit for determining archaic admixture

tamilgangster
10-08-2015, 04:27 AM
THis is just speculation, but its possible that the elevated East African, East eurasian, and much of the ANE could be from an ancient population similar to ust ishim. This is based on the results for ust ishim, being scattered among different components.

Dr_McNinja
11-03-2015, 10:26 PM
Sample run I did:

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=97

Can't quite figure out what went wrong with East-India. It's like 20% ANE, 25% ASE, 30% East Eurasian, 25% ENF... I'm guessing it's acting kind of like a Western China type signal perhaps... or I'm not getting it.

tamilgangster
11-03-2015, 10:36 PM
Sample run I did:

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=97

Can't quite figure out what went wrong with East-India. It's like 20% ANE, 25% ASE, 30% East Eurasian, 25% ENF... I'm guessing it's acting kind of like a Western China type signal perhaps... or I'm not getting it.

THe east India sample, im guessing is supposed to represent austroasiatic tribals from central india. WHich im guessing would be somthing like 30% East eurasian, 30% ASE 30%ENF 5%Ocenian 5% East African, I think you will get less central asian related noise

Dr_McNinja
11-08-2015, 03:23 AM
So I wound up going off on a tangent and making a calculator similar to Harappa but where Gedrosian is higher and Caucasus lower in South Asia:

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=99

Interestingly with the Pashtun, the Caucasus signal picks up as separate again.

Dr_McNinja
01-11-2016, 03:47 AM
HarappaWorld calculator with CHG instead of Baloch/Caucasian:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=358502708

HAP Baloch in K7/K8 comes out ~35% ANE, ~9.3% ASE and I think Caucasian (peaking in Georgians) is like 25-26% ANE, ~2% ASE. Not sure since I didn't request Everest run the zombies for the Caucasian component through K7/K8 (that's the values for a Caucasian component from another calculator). If anyone can get that info, I can remake it more exactly.

So if you average the two components together, they average out to basically CHG/Kotias proportions (~31% ANE, ~5% ASE) which I'm not sure is a coincidence since both components are CHG derived.

It's close to Baloch+Caucasian values but not exactly. Most South Asians' noise in SW-Asian/Mediterranean disappears as a result.

The reason it's still so high in Iran/Levant/S-Caucasus/Anatolia is because there's no proper Anatolian Neolithic Farmer or Early Neolithic Farmer component. It's just SW-Asian and Mediterranean (anyone know the ANE/ENF/WHG breakdown of those Anatolian NFs?)

An EHG component would pull some admixture away from CHG (since some CHG is stuck in "S-Indian"). But aside from dropping or mixing components, can't do much more manipulation of existing calculators.

Dr_McNinja
01-11-2016, 03:51 AM
I'm also doing the Harappa-CHG with S-Indian removed. Results so far:

Me: 9.88% NE-Euro
6.73% SE-Asian
4.08% Siberian
3.84% NE-Asian
5.82% Papuan
2.41% American
2.84% Beringian
0.00% Mediterranean
1.15% SW-Asian
0.04% San
0.92% E-African
0.76% Pygmy
0.49% W-African
61.03% CHG

Sapporo: 11.28% NE-Euro
5.32% SE-Asian
4.06% Siberian
2.33% NE-Asian
3.71% Papuan
2.45% American
2.40% Beringian
0.00% Mediterranean
1.09% SW-Asian
0.00% San
1.37% E-African
0.13% Pygmy
0.15% W-African
65.72% CHG HRP0393 Haryana: 17.87% NE-Euro
5.70% SE-Asian
3.89% Siberian
2.91% NE-Asian
4.08% Papuan
2.88% American
0.33% Beringian
0.02% Mediterranean
0.06% SW-Asian
0.72% San
0.34% E-African
0.03% Pygmy
1.18% W-African
59.99% CHG NE-Euro goes back up and there's remarkable closeness in the remaining components. Sapporo isn't that much more in CHG and there's a little SW-Asian. The Siberian/Amerindian/Beringian being close or the same is also interesting.

Parasar (Bihar Brahmin): 10.87% NE-Euro
5.75% SE-Asian
1.60% Siberian
8.36% NE-Asian
6.87% Papuan
3.13% American
1.07% Beringian
0.00% Mediterranean
0.41% SW-Asian
0.01% San
0.96% E-African
0.07% Pygmy
1.00% W-African
59.90% CHG Ashwin (Gujarati/Maharashtra Brahmin): 6.87% NE-Euro
5.95% SE-Asian
4.37% Siberian
9.11% NE-Asian
7.53% Papuan
1.89% American
0.88% Beringian
0.00% Mediterranean
0.74% SW-Asian
0.03% San
1.32% E-African
0.01% Pygmy
1.77% W-African
59.54% CHG

Dr_McNinja
01-11-2016, 04:05 AM
HarappaWorld calculator with CHG instead of Baloch/Caucasian:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=358502708

HAP Baloch in K7/K8 comes out ~35% ANE, ~9.3% ASE and I think Caucasian (peaking in Georgians) is like 25-26% ANE, ~2% ASE. Not sure since I didn't request Everest run the zombies for the Caucasian component through K7/K8 (that's the values for a Caucasian component from another calculator). If anyone can get that info, I can remake it more exactly.

So if you average the two components together, they average out to basically CHG/Kotias proportions (~31% ANE, ~5% ASE) which I'm not sure is a coincidence since both components are CHG derived.

It's close to Baloch+Caucasian values but not exactly. Most South Asians' noise in SW-Asian/Mediterranean disappears as a result.

The reason it's still so high in Iran/Levant/S-Caucasus/Anatolia is because there's no proper Anatolian Neolithic Farmer or Early Neolithic Farmer component. It's just SW-Asian and Mediterranean (anyone know the ANE/ENF/WHG breakdown of those Anatolian NFs?)

An EHG component would pull some admixture away from CHG (since some CHG is stuck in "S-Indian"). But aside from dropping or mixing components, can't do much more manipulation of existing calculators.
The SW-Asian is mostly gone, but what little's left might be significant in South Asians. It's only in any appreciable amount in Baloch or Pashtun. So the tiny traces in South Asians could represent Baloch or Pashtun-admixture (basically west of the Indus). Interestingly my father and that one cousin of my mom (who show this pattern in other calculators and share a bunch of DNA to the exclusion of my mom and I) both have a little coming up. My friend from UP/Mumbai has just a trace also appearing. I think the reason the SW-Asian component here is becoming like this is because all the ENF-type admixture is sucked up by CHG leaving only that which has a strong affinity to that component. So even 0.10% might be significant.

So this could represent Baloch-type ancestors several generations back in Punjabis (since my working theory is Punjabi Jatts are basically a combo of Baloch/Sindhi-like and Haryana-like groups from >1000ya).

In the run without S-Indian, my extra SE-Asian is interesting. Could just be the same ANE-starvation phenomena again. I've found other mixed Pakistani Punjabis who have Eurogenes K7 WHG values over 5% but have a lot of ASE and East Eurasian. They do not resemble in admixture any other Indian group. It looks like a mix of Jatt/Brahmin type with populations high in East Asian, but the Oracles can't find suitable populations. One I just added in punDNAL had an Oracle showing 70% UP Muslim and 30% Haryana Jatt. That sounds like something that could have fit us, but my Oracle didn't give any Haryana results.

EDIT: Rest of CHG run without S-Indian:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1105414506

Mellifluous
01-11-2016, 04:48 AM
HarappaWorld calculator with CHG instead of Baloch/Caucasian:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=358502708

HAP Baloch in K7/K8 comes out ~35% ANE, ~9.3% ASE and I think Caucasian (peaking in Georgians) is like 25-26% ANE, ~2% ASE. Not sure since I didn't request Everest run the zombies for the Caucasian component through K7/K8 (that's the values for a Caucasian component from another calculator). If anyone can get that info, I can remake it more exactly.

So if you average the two components together, they average out to basically CHG/Kotias proportions (~31% ANE, ~5% ASE) which I'm not sure is a coincidence since both components are CHG derived.

It's close to Baloch+Caucasian values but not exactly. Most South Asians' noise in SW-Asian/Mediterranean disappears as a result.

The reason it's still so high in Iran/Levant/S-Caucasus/Anatolia is because there's no proper Anatolian Neolithic Farmer or Early Neolithic Farmer component. It's just SW-Asian and Mediterranean (anyone know the ANE/ENF/WHG breakdown of those Anatolian NFs?)

An EHG component would pull some admixture away from CHG (since some CHG is stuck in "S-Indian"). But aside from dropping or mixing components, can't do much more manipulation of existing calculators.

Will we be able to use this calculator or are you just experimenting with it? :)

Dr_McNinja
01-11-2016, 04:56 AM
Will we be able to use this calculator or are you just experimenting with it? :)I can upload this one if you'd like. It requires DIYDodecad

Dr_McNinja
01-11-2016, 05:32 AM
Here's the link: https://drive.google.com/open?id=0B-XBmvmgdkfVbXMxUDNNbXgxNjA

Mellifluous
01-11-2016, 06:43 AM
Here's the link: https://drive.google.com/open?id=0B-XBmvmgdkfVbXMxUDNNbXgxNjA

Thanks!

khanabadoshi
01-11-2016, 06:44 AM
Here's the link: https://drive.google.com/open?id=0B-XBmvmgdkfVbXMxUDNNbXgxNjA

I'll run mi familia now.

Helgenes50
01-11-2016, 07:11 AM
Here's the link: https://drive.google.com/open?id=0B-XBmvmgdkfVbXMxUDNNbXgxNjA

Thanks Dc_McNinja,

Harappa is one of the best calculators.
I know that I'm not South Indian, I post my results to compare.
These are very close to those of David and Chad

47.11% NE-Euro
0.00% SE-Asian
0.00% Siberian
0.00% NE-Asian
0.03% Papuan
0.59% American
0.01% Beringian
34.05% Mediterranean
2.09% SW-Asian
0.01% San
0.07% E-African
0.27% Pygmy
0.02% W-African
15.75% CHG

My mother's results

45.32% NE-Euro
0.00% SE-Asian
0.01% Siberian
0.00% NE-Asian
0.16% Papuan
0.67% American
0.18% Beringian
32.71% Mediterranean
2.15% SW-Asian
0.00% San
0.03% E-African
0.28% Pygmy
0.00% W-African
18.49% CHG

MonkeyDLuffy
01-11-2016, 08:42 AM
This calculator is shaping up nicely mcninja. I'm surprised to see my east euro is higher and chg is lower. Probably the calculator did a good job seperating the euro from chg.

khanabadoshi
01-11-2016, 08:55 AM
Myself

Sister



8.34% NE-Euro
4.89% SE-Asian
5.76% Siberian
4.44% NE-Asian
5.03% Papuan
1.12% American
0.57% Beringian
0.00% Mediterranean
1.41% SW-Asian
0.23% San
0.12% E-African
0.03% Pygmy
1.70% W-African
66.36% CHG

10.88% NE-Euro
3.64% SE-Asian
4.31% Siberian
5.22% NE-Asian
4.26% Papuan
2.17% American
0.28% Beringian
0.02% Mediterranean
2.34% SW-Asian
0.09% San
0.21% E-African
0.00% Pygmy
2.00% W-African
64.57% CHG



Grandmother (M)

Grandmother (M) Brother



16.74% NE-Euro
1.56% SE-Asian
4.73% Siberian
9.07% NE-Asian
2.94% Papuan
2.77% American
1.13% Beringian
0.11% Mediterranean
0.99% SW-Asian
0.00% San
0.03% E-African
0.11% Pygmy
0.96% W-African
58.86% CHG

18.25% NE-Euro
2.18% SE-Asian
5.35% Siberian
10.75% NE-Asian
2.72% Papuan
3.60% American
1.07% Beringian
0.22% Mediterranean
0.83% SW-Asian
0.04% San
0.71% E-African
0.04% Pygmy
0.10% W-African
54.15% CHG



Uncle (P)

Aunt (M)



7.81% NE-Euro
5.18% SE-Asian
3.16% Siberian
3.96% NE-Asian
4.45% Papuan
1.86% American
1.24% Beringian
0.00% Mediterranean
1.76% SW-Asian
0.29% San
0.16% E-African
0.01% Pygmy
2.16% W-African
67.96% CHG

12.17% NE-Euro
3.78% SE-Asian
4.75% Siberian
8.11% NE-Asian
4.36% Papuan
2.55% American
0.48% Beringian
0.04% Mediterranean
0.42% SW-Asian
0.36% San
0.74% E-African
0.14% Pygmy
0.96% W-African
61.13% CHG



Grandfather (M) Sister

1/2 Legahri 1/4 Lahori -- 2nd Cousin



7.63% NE-Euro
3.06% SE-Asian
5.07% Siberian
5.79% NE-Asian
5.60% Papuan
1.32% American
1.36% Beringian
0.00% Mediterranean
1.65% SW-Asian
0.38% San
1.52% E-African
0.05% Pygmy
1.39% W-African
65.16% CHG

6.32% NE-Euro
4.74% SE-Asian
3.57% Siberian
4.12% NE-Asian
4.60% Papuan
1.58% American
1.24% Beringian
0.00% Mediterranean
2.61% SW-Asian
0.04% San
1.44% E-African
0.23% Pygmy
1.29% W-African
68.20% CHG



Mohmand Pashtun -- Pakistan




12.15% NE-Euro
2.61% SE-Asian
4.85% Siberian
3.25% NE-Asian
3.27% Papuan
2.40% American
0.52% Beringian
0.41% Mediterranean
1.90% SW-Asian
0.36% San
0.30% E-African
0.06% Pygmy
0.74% W-African
67.18% CHG

Dr_McNinja
01-11-2016, 10:12 AM
I just did another where I seem to be getting consistent ~30-40% Papuan for the East Eurasian part of S-Indian. The rest is a varying mixture of NE-Asian and SE-Asian (mostly SE-Asian).

It looks like something like Karasuk makes everything neat and clean for South Asians. Karasuk alone isn't enough, Andronovo is needed. For some Afghans, even that isn't enough which signifies to me either Sintashta or recent Tajik-like ancestry.

Normally this would just be the equivalent of a fancy Oracle within the calculator. Except these Bronze Age populations are recent enough that the component parts (Baloch, NE-Euro, Med, Siberian, etc), which in this calculator are from real populations, and their allele frequencies could actually be relevant. Especially since HarappaWorld seems to do inordinately well apart from the old S-Indian/Baloch. It's remarkably stable. The mixed South American/European individual I have gets consistent Amerindian results no matter how much I tinker with it. Nothing "leaks".

And for some reason, the Tajik from Tajikstan does not have the same affinity to Karasuk (but does to Andronovo). He has the constituent parts instead (more in line with recent mixing of East Asian-like admixture with Tajiks).

Interestingly, the same happens for Rukha. And he's half Tajik (the other half Tajik I had showed up with Karasuk like everyone else, but Rukha's Tajik half is supposed to be more Pamiri-like IIRC). This does not happen for the Haryana Jatt, all of his Euro admixture goes into Andronovo/Karasuk.

Here's Andronovo+Karasuk+CHG:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1040594487

Here's Karasuk+CHG:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1810383904

I'll post Andronovo tomorrow, but it's as expected... ~30% in Haryana Jatt and below, leaving NE-Euro, Amerindian, Beringian, Siberian, etc left over (suggesting Andronovo alone isn't enough to account for the Steppe admixture, needs something Karasuk-like as well).

The Karasuk runs seem to indicate an almost basal ~20% Karasuk-like layer across South Asia...

Did they ever find out what branch of R1a those Karasuk remains belonged to? (Note: This wouldn't mean Karasuk is the Steppe ancestor of India, just that something very much like it fits nicely)

EDIT: For calculating the East Eurasian part of S-Indian, or "ASI", if we go with Khetran Baloch's 38-40% as a baseline, that would presumably mean any number under this is dilution due to genuine extra NE-Asian (like with Bored/Jammu-Brahmin).

Also, total Steppe admixture in the Afghans at the far end would be at least mid-30s.

anglesqueville
01-11-2016, 10:20 AM
Me:

52.30% NE-Euro
0.19% SE-Asian
0.23% Siberian
0.08% NE-Asian
0.50% Papuan
1.10% American
0.02% Beringian
32.45% Mediterranean
0.36% SW-Asian
0.05% San
0.00% E-African
0.00% Pygmy
0.27% W-African
12.45% CHG

My mother:

53.20% NE-Euro
0.22% SE-Asian
0.40% Siberian
0.01% NE-Asian
0.08% Papuan
0.57% American
0.00% Beringian
31.99% Mediterranean
0.63% SW-Asian
0.28% San
0.00% E-African
0.00% Pygmy
0.02% W-African
12.60% CHG

My wife:

41.51% NE-Euro
0.03% SE-Asian
0.01% Siberian
0.03% NE-Asian
0.12% Papuan
0.06% American
0.08% Beringian
36.71% Mediterranean
3.10% SW-Asian
0.14% San
0.02% E-African
0.00% Pygmy
0.00% W-African
18.17% CHG

Reza
01-11-2016, 11:43 AM
I just did another where I seem to be getting consistent ~30-40% Papuan for the East Eurasian part of S-Indian. The rest is a varying mixture of NE-Asian and SE-Asian (mostly SE-Asian).

It looks like something like Karasuk makes everything neat and clean for South Asians. Karasuk alone isn't enough, Andronovo is needed. For some Afghans, even that isn't enough which signifies to me either Sintashta or recent Tajik-like ancestry.

Normally this would just be the equivalent of a fancy Oracle within the calculator. Except these Bronze Age populations are recent enough that the component parts (Baloch, NE-Euro, Med, Siberian, etc), which in this calculator are from real populations, and their allele frequencies could actually be relevant. Especially since HarappaWorld seems to do inordinately well apart from the old S-Indian/Baloch. It's remarkably stable. The mixed South American/European individual I have gets consistent Amerindian results no matter how much I tinker with it. Nothing "leaks".

And for some reason, the Tajik from Tajikstan does not have the same affinity to Karasuk (but does to Andronovo). He has the constituent parts instead (more in line with recent mixing of East Asian-like admixture with Tajiks).

Interestingly, the same happens for Rukha. And he's half Tajik (the other half Tajik I had showed up with Karasuk like everyone else, but Rukha's Tajik half is supposed to be more Pamiri-like IIRC). This does not happen for the Haryana Jatt, all of his Euro admixture goes into Andronovo/Karasuk.

Here's Andronovo+Karasuk+CHG:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1040594487

Here's Karasuk+CHG:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1810383904

I'll post Andronovo tomorrow, but it's as expected... ~30% in Haryana Jatt and below, leaving NE-Euro, Amerindian, Beringian, Siberian, etc left over (suggesting Andronovo alone isn't enough to account for the Steppe admixture, needs something Karasuk-like as well).

The Karasuk runs seem to indicate an almost basal ~20% Karasuk-like layer across South Asia...

Did they ever find out what branch of R1a those Karasuk remains belonged to? (Note: This wouldn't mean Karasuk is the Steppe ancestor of India, just that something very much like it fits nicely)

EDIT: For calculating the East Eurasian part of S-Indian, or "ASI", if we go with Khetran Baloch's 38-40% as a baseline, that would presumably mean any number under this is dilution due to genuine extra NE-Asian (like with Bored/Jammu-Brahmin).

Also, total Steppe admixture in the Afghans at the far end would be at least mid-30s.

Nice work Dr_McNinja.

Question about Karasuk .. what would account for its variation between S Asian members? And does it have any E Asian affinity? I score 22% which is on the higher end of the range. My SE and NE Asian scores don't appear to have bled much into it..

Varun R
01-11-2016, 01:12 PM
23andMe 0.997917 genotype rate
8.98% NE-Euro
7.90% SE-Asian
3.67% Siberian
6.67% NE-Asian
6.01% Papuan
2.00% American
0.67% Beringian
0.00% Mediterranean
0.15% SW-Asian
0.67% San
1.45% E-African
0.96% Pygmy
0.05% W-African
60.82% CHG

Dr_McNinja
01-11-2016, 04:26 PM
Nice work Dr_McNinja.

Question about Karasuk .. what would account for its variation between S Asian members? And does it have any E Asian affinity? I score 22% which is on the higher end of the range. My SE and NE Asian scores don't appear to have bled much into it..
Everyone's pretty much around that mark. There's a slight dip near Balochistan. Some people have more affinity to Karasuk specifically if they've had significant Siberian/Beringian/Amerindian.

evon
01-11-2016, 04:37 PM
Norway clan..
Harappa usually underestimates non-European components, and I see this is no different from before in this regard. The SE Asian is more correct here then in many other calculators, but it gave my maternal uncle more then my grandmother, which is clearly wrong. CHG looks inline with most other calculators, no big difference really, as most land between 20 and 13%..


Me (23andme v3):

56.53% NE-Euro
0.72% SE-Asian
0.11% Siberian
0.46% NE-Asian
0.14% Papuan
0.05% American
0.02% Beringian
28.35% Mediterranean
0.47% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
13.15% CHG

Mother (FTDNA):

56.57% NE-Euro
0.64% SE-Asian
0.58% Siberian
0.33% NE-Asian
0.08% Papuan
0.20% American
0.00% Beringian
28.13% Mediterranean
0.15% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
13.31% CHG

Grandmother (23andme V3):

55.71% NE-Euro
0.81% SE-Asian
0.07% Siberian
0.04% NE-Asian
0.13% Papuan
0.05% American
0.06% Beringian
27.02% Mediterranean
0.47% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
15.64% CHG

Grand Uncle (FTDNA):

55.05% NE-Euro
0.24% SE-Asian
0.85% Siberian
0.08% NE-Asian
0.00% Papuan
0.34% American
0.00% Beringian
27.48% Mediterranean
0.12% SW-Asian
0.00% San
0.01% E-African
0.00% Pygmy
0.00% W-African
15.83% CHG

Paternal Aunt (23andme V2):

54.07% NE-Euro
0.02% SE-Asian
0.40% Siberian
0.04% NE-Asian
0.33% Papuan
0.04% American
1.02% Beringian
29.38% Mediterranean
0.18% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
14.52% CHG

Maternal Uncle (23andme V2):

57.15% NE-Euro
1.48% SE-Asian
0.12% Siberian
0.11% NE-Asian
0.43% Papuan
0.01% American
0.00% Beringian
27.37% Mediterranean
0.20% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
13.12% CHG

Dr_McNinja
01-11-2016, 07:02 PM
Damn, can't get an ASI component to stick. The constituent components (Papuan, SE-Asian, etc) have a stronger pull than the composite. Papuan is coming out to 50% of the Papuan+SE-Asian number, quite consistently. Could be the NE-Asian and African isn't related at all.

Guess that would be something like an Onge population.

Dr_McNinja
01-12-2016, 03:36 AM
So Karasuk shouldn't be a surprise I guess, try running HRP0418's kit through Gedmatch's Archaic matches calculator: M003843

You might have to squint a little but you can eyeball it easily. The same kits are showing up on everyone's (I checked myself, HRP0393 and Sapporo before HRP0418). RISE497 (Karasuk), 505 (Andronovo), 493 (Karasuk), 495 (Karasuk), 496 (Karasuk), 523 (Mez), 511 (Afansievo). 98 (baSca) showed up well on HRP0393's but was less strong on ours (but enough to be noted).

I noticed traces of 548 (Yamnaya), 150 (baUne), and 395 (Sintashta) as well but not as much as the others.

randwulf
01-12-2016, 05:08 AM
My tribe with Dr. McNinja's Harappaworld with CHG:

Me:

44.72% NE-Euro
0.12% SE-Asian
0.85% Siberian
0.07% NE-Asian
0.04% Papuan
0.06% American
0.20% Beringian
34.22% Mediterranean
2.95% SW-Asian
0.22% San
0.00% E-African
0.16% Pygmy
0.01% W-African
16.39% CHG

Dad:

43.41% NE-Euro
0.14% SE-Asian
1.06% Siberian
0.03% NE-Asian
0.56% Papuan
0.51% American
0.21% Beringian
32.20% Mediterranean
2.13% SW-Asian
0.26% San
0.01% E-African
0.06% Pygmy
0.08% W-African
19.34% CHG

Mom:

48.76% NE-Euro
0.03% SE-Asian
0.02% Siberian
0.02% NE-Asian
0.02% Papuan
0.09% American
0.18% Beringian
32.47% Mediterranean
1.37% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
17.04% CHG

Wife:

45.55% NE-Euro
0.07% SE-Asian
0.37% Siberian
0.06% NE-Asian
0.04% Papuan
0.12% American
0.03% Beringian
33.06% Mediterranean
2.73% SW-Asian
0.04% San
0.32% E-African
0.03% Pygmy
0.03% W-African
17.56% CHG

Father-in-law:

48.91% NE-Euro
0.07% SE-Asian
0.28% Siberian
0.04% NE-Asian
0.00% Papuan
0.08% American
0.21% Beringian
31.03% Mediterranean
1.58% SW-Asian
0.01% San
0.40% E-African
0.20% Pygmy
0.44% W-African
16.74% CHG

Mellifluous
01-12-2016, 07:25 AM
14.48% NE-Euro
1.10% SE-Asian
3.59% Siberian
6.47% NE-Asian
3.71% Papuan
2.31% American
0.98% Beringian
0.30% Mediterranean
2.59% SW-Asian
0.00% San
0.57% E-African
0.30% Pygmy
0.28% W-African
63.31% CHG

anglesqueville
01-12-2016, 04:44 PM
Some PCAs (non normalized)
7246
72477248

psaglav
01-12-2016, 06:02 PM
This is another one of those DIY calculators, isn't it?

kerbal
01-12-2016, 06:15 PM
FINAL ADMIXTURE PROPORTIONS:
----------------------------

40.61% NE-Euro
0.38% SE-Asian
0.01% Siberian
0.25% NE-Asian
0.35% Papuan
0.57% American
0.01% Beringian
38.94% Mediterranean
1.84% SW-Asian
0.00% San
0.03% E-African
0.06% Pygmy
0.10% W-African
16.86% CHG
0.980002 genotype rate

evon
01-12-2016, 06:54 PM
This is another one of those DIY calculators, isn't it?

Poor thing, I just imagine a confused stuffed animal trying to grapple with unfamiliar software... :behindsofa:

If you want I can run your data, but I remember you sent it to someone else already, so maybe that person could do it? if not you can send it to me, I will not share it with anyone etc..

psaglav
01-12-2016, 07:32 PM
Poor thing, I just imagine a confused stuffed animal trying to grapple with unfamiliar software... :behindsofa:

If you want I can run your data, but I remember you sent it to someone else already, so maybe that person could do it? if not you can send it to me, I will not share it with anyone etc..

The avatar picture is me, a stuffed owl going WAT? O_O

The damn program didn't run properly althought I did everything it said. I wasn't even able to change the directory. boo

evon
01-12-2016, 07:36 PM
The avatar picture is me, a stuffed owl going WAT? O_O

The damn program didn't run properly althought I did everything it said. I wasn't even able to change the directory. boo

If you have access to a Windows machine I would try that, as it is much easier then using Mac/Linux..You could try a crossover software which would allow you to use Windows software on a Mac machine..I am not familiar with Mac, but I use Wine on my Linux machine, which is listed here:
http://www.howtogeek.com/187359/5-ways-to-run-windows-software-on-a-mac/

psaglav
01-12-2016, 08:18 PM
Thanks to evon:

28.36% NE-Euro
0.20% SE-Asian
1.37% Siberian
0.54% NE-Asian
0.50% Papuan
0.17% American
2.35% Beringian
19.92% Mediterranean
9.19% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
37.40% CHG

MfA
01-12-2016, 08:37 PM
.....

Mestace
01-13-2016, 10:29 AM
Mine
43.08% NE-Euro
0.07% SE-Asian
0.01% Siberian
0.01% NE-Asian
0.00% Papuan
0.02% American
0.17% Beringian
36.05% Mediterranean
2.39% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
18.20% CHG



Mother
44.36% NE-Euro
0.02% SE-Asian
0.01% Siberian
0.01% NE-Asian
0.00% Papuan
0.01% American
0.13% Beringian
36.33% Mediterranean
1.93% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
17.22% CHG



Father
41.49% NE-Euro
0.03% SE-Asian
0.01% Siberian
0.01% NE-Asian
0.05% Papuan
0.16% American
0.04% Beringian
35.94% Mediterranean
3.93% SW-Asian
0.00% San
0.03% E-African
0.01% Pygmy
0.01% W-African
18.29% CHG

anglesqueville
01-13-2016, 11:59 AM
update
7269

anglesqueville
01-13-2016, 02:02 PM
For europeans ony, not normed variables, and normed variables. The "clan" points are familial kits.
72707271

psaglav
01-13-2016, 02:55 PM
update
7269

There I am! Right in the middle, by myself. I wonder if the Ashkenazi Jews would place in a similar spot? Or, say, someone from Romania.

psaglav
01-13-2016, 02:57 PM
update
7269

What is the 0 point on that graph? Is it everyone's median or the median of all the populations in the calculator? (or smth else?)

anglesqueville
01-13-2016, 03:16 PM
What is the 0 point on that graph? Is it everyone's median or the median of all the populations in the calculator? (or smth else?)

"median" is not the right word. Nothing else than the zero vector of the 2-dimensional vectorial space where the covariances of the individual variables are projected. In itself this point has no signification.

psaglav
01-13-2016, 04:27 PM
"median" is not the right word. Nothing else than the zero vector of the 2-dimensional vectorial space where the covariances of the individual variables are projected. In itself this point has no signification.

I only understood the last bit, but thanks, since it has no significance.

Dr_McNinja
01-22-2016, 06:13 AM
Slight detour from the previous runs:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1084480849

Useful for South Asians only.

I'm working on another version with a "Bengal" component, except it's showing up in Afghans, Pahari Jatts, J&K, Nepal, me and my family... probably because of the SE-Asian in it which isn't covered adequately by the N-Indian Brahmin component (which has a little NE-Asian, but no extra SE-Asian).

Weird, because there is an SE-Asian component but it seems that it likes to come packaged with other South Asian admixture. Obviously, this admixture isn't from all the way over there, so what population can I use for such a component? I could go with Nepal, but I doubt it came from Nepal to Kashmir, Pakistan, and Afghanistan. I can't find anything suitable in the official Harappa spreadsheet. I could just try to make up an "East Pahari" pop from Kashmir.

Were it not for that, you could try to actually plot these results on a map (perhaps you still can with this run).

Also, has anyone tried using Pacific Islander populations like Tongan or Samoan in place of Dai/SE-Asian as an ASI proxy in South Asians?

paulgill
01-22-2016, 06:24 AM
Slight detour from the previous runs:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1084480849

Useful for South Asians only.

I'm working on another version with a "Bengal" component, except it's showing up in Afghans, Pahari Jatts, J&K, Nepal, me and my family... probably because of the SE-Asian in it which isn't covered adequately by the N-Indian Brahmin component (which has a little NE-Asian, but no extra SE-Asian).

Weird, because there is an SE-Asian component but it seems that it likes to come packaged with other South Asian admixture. Obviously, this admixture isn't from all the way over there, so what population can I use for such a component? I could go with Nepal, but I doubt it came from Nepal to Kashmir, Pakistan, and Afghanistan. I can't find anything suitable in the official Harappa spreadsheet. I could just try to make up an "East Pahari" pop from Kashmir.

Were it not for that, you could try to actually plot these results on a map (perhaps you still can with this run).

Also, has anyone tried using Pacific Islander populations like Tongan or Samoan in place of Dai/SE-Asian as an ASI proxy in South Asians?

Use the name of the population [Ethnicity] which has highest of that component.

bored
01-22-2016, 07:54 AM
Slight detour from the previous runs:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1084480849

Useful for South Asians only.

I'm working on another version with a "Bengal" component, except it's showing up in Afghans, Pahari Jatts, J&K, Nepal, me and my family... probably because of the SE-Asian in it which isn't covered adequately by the N-Indian Brahmin component (which has a little NE-Asian, but no extra SE-Asian).

Weird, because there is an SE-Asian component but it seems that it likes to come packaged with other South Asian admixture. Obviously, this admixture isn't from all the way over there, so what population can I use for such a component? I could go with Nepal, but I doubt it came from Nepal to Kashmir, Pakistan, and Afghanistan. I can't find anything suitable in the official Harappa spreadsheet. I could just try to make up an "East Pahari" pop from Kashmir.

Were it not for that, you could try to actually plot these results on a map (perhaps you still can with this run).

Also, has anyone tried using Pacific Islander populations like Tongan or Samoan in place of Dai/SE-Asian as an ASI proxy in South Asians?

"NW Punjab" component peaks in which populations? Mine is high.

Looks like it's a Pahari thing since bol_nat scores high on it too. "Pahari" as in Western Pahari not Nepal etc.

tamilgangster
01-22-2016, 08:43 AM
Slight detour from the previous runs:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1084480849

Useful for South Asians only.

I'm working on another version with a "Bengal" component, except it's showing up in Afghans, Pahari Jatts, J&K, Nepal, me and my family... probably because of the SE-Asian in it which isn't covered adequately by the N-Indian Brahmin component (which has a little NE-Asian, but no extra SE-Asian).

Weird, because there is an SE-Asian component but it seems that it likes to come packaged with other South Asian admixture. Obviously, this admixture isn't from all the way over there, so what population can I use for such a component? I could go with Nepal, but I doubt it came from Nepal to Kashmir, Pakistan, and Afghanistan. I can't find anything suitable in the official Harappa spreadsheet. I could just try to make up an "East Pahari" pop from Kashmir.

Were it not for that, you could try to actually plot these results on a map (perhaps you still can with this run).

Also, has anyone tried using Pacific Islander populations like Tongan or Samoan in place of Dai/SE-Asian as an ASI proxy in South Asians?

they are too denisovan shifted

Dr_McNinja
01-25-2016, 05:41 AM
When removing S-Indian from Harappa, this is where the admixture went:

Me (37.18% S-Indian):

Baloch - 11.51%
Caucasian - 18.29%
NE-Euro - 6.48%
SE-Asian - 17.43%
Siberian - 5.92%
NE-Asian - 11.43%
Papuan - 12.93%
Amerindian - 2.64%
Med - 0.0%
Beringian - 1.99%
SW-Asian - 5.16%
San - 0.38%
E-African - 3.23%
Pygmy - 1.48%
W-African - 1.10%

Sapporo (30.15% S-Indian)

Baloch - 11.67%
Caucasian - 16.75%
NE-Euro - 6.67%
SE-Asian - 16.29%
Siberian - 8.13%
NE-Asian - 9.62%
Papuan - 11.70%
Amerindian - 2.95%
Beringian - 2.82%
Med - 2.12%
SW-Asian - 5.17%
San - 0.0%
E-African - 5.04%
Pygmy - 0.50%
W-African - 0.53%

For kicks, I put that into Harappa Oracle.

Mine:

[1,] "31.5% tongan_xing_13 + 68.5% turkmen_yunusbayev_11" "10.9646"
[2,] "31.3% tongan_xing_13 + 68.7% uzbek_behar_15" "12.1198"
[3,] "30.9% samoan_xing_13 + 69.1% turkmen_yunusbayev_11" "12.1334"
[4,] "66.4% nogai_yunusbayev_16 + 33.6% tongan_xing_13" "12.399"
[5,] "30.7% samoan_xing_13 + 69.3% uzbek_behar_15" "13.1783"
[6,] "67% nogai_yunusbayev_16 + 33% samoan_xing_13" "13.5296"
[7,] "34% tongan_xing_13 + 66% turk-aydin_hodoglugil_20" "14.7725"
[8,] "66.6% tajik_yunusbayev_15 + 33.4% tongan_xing_13" "15.0488"
[9,] "32.1% tongan_xing_13 + 67.9% uyghur_hgdp_10" "15.2025"
[10,] "67.1% hazara_hgdp_22 + 32.9% tongan_xing_13" "15.4729"
[11,] "33.4% samoan_xing_13 + 66.6% turk-aydin_hodoglugil_20" "15.766"
[12,] "32.8% samoan_xing_13 + 67.2% tajik_yunusbayev_15" "15.9418"
[13,] "31.4% samoan_xing_13 + 68.6% uyghur_hgdp_10" "16.0797"
[14,] "67.7% hazara_hgdp_22 + 32.3% samoan_xing_13" "16.3061"
[15,] "36.2% tongan_xing_13 + 63.8% turk-istanbul_hodoglugil_20" "16.3094"
[16,] "63.9% iranian_harappa_9 + 36.1% tongan_xing_13" "16.407"
[17,] "35.8% tongan_xing_13 + 64.2% turkish_harappa_4" "16.4204"
[18,] "63.2% azeri_harappa_2 + 36.8% tongan_xing_13" "16.4356"
[19,] "36.4% tongan_xing_13 + 63.6% turk-kayseri_hodoglugil_23" "16.8103"
[20,] "36.9% tongan_xing_13 + 63.1% turk_behar_19" "16.819"
[21,] "63.6% kurd_harappa_11 + 36.4% tongan_xing_13" "16.9057"
[22,] "63.6% iranian_behar_19 + 36.4% tongan_xing_13" "16.9844"
[23,] "62.2% kumyk_yunusbayev_14 + 37.8% tongan_xing_13" "17.0419"
[24,] "22.4% melanesian_hgdp_10 + 77.6% turkmen_yunusbayev_11" "17.105"
[25,] "63.9% stalskoe_xing_5 + 36.1% tongan_xing_13" "17.2112"

Sapporo:

[1,] "29.6% tongan_xing_13 + 70.4% uzbek_behar_15" "11.6242"
[2,] "30.1% tongan_xing_13 + 69.9% turkmen_yunusbayev_11" "11.658"
[3,] "29.1% samoan_xing_13 + 70.9% uzbek_behar_15" "12.5095"
[4,] "29.6% samoan_xing_13 + 70.4% turkmen_yunusbayev_11" "12.5572"
[5,] "67.7% nogai_yunusbayev_16 + 32.3% tongan_xing_13" "13.319"
[6,] "68.2% nogai_yunusbayev_16 + 31.8% samoan_xing_13" "14.187"
[7,] "68.2% tajik_yunusbayev_15 + 31.8% tongan_xing_13" "15.1353"
[8,] "68.6% hazara_hgdp_22 + 31.4% tongan_xing_13" "15.6222"
[9,] "30.6% tongan_xing_13 + 69.4% uyghur_hgdp_10" "15.7134"
[10,] "32.8% tongan_xing_13 + 67.2% turk-aydin_hodoglugil_20" "15.7475"
[11,] "31.3% samoan_xing_13 + 68.7% tajik_yunusbayev_15" "15.8493"
[12,] "69.1% hazara_hgdp_22 + 30.9% samoan_xing_13" "16.2806"
[13,] "30.1% samoan_xing_13 + 69.9% uyghur_hgdp_10" "16.4021"
[14,] "32.3% samoan_xing_13 + 67.7% turk-aydin_hodoglugil_20" "16.5181"
[15,] "21.1% melanesian_hgdp_10 + 78.9% turkmen_yunusbayev_11" "17.4238"
[16,] "20.5% melanesian_hgdp_10 + 79.5% uzbek_behar_15" "17.4987"
[17,] "27.3% singapore-malay_sgvp_89 + 72.7% turkmen_yunusbayev_11" "17.571"
[18,] "29.7% khmer-cambodian_xing_3 + 70.3% turkmen_yunusbayev_11" "17.669"
[19,] "65% iranian_harappa_9 + 35% tongan_xing_13" "17.67"
[20,] "64.2% azeri_harappa_2 + 35.8% tongan_xing_13" "17.7618"
[21,] "26.4% dai_hgdp_10 + 73.6% turkmen_yunusbayev_11" "17.7758"
[22,] "26.7% dai-chinese_1000genomes_91 + 73.3% turkmen_yunusbayev_11" "17.8136"
[23,] "35.2% tongan_xing_13 + 64.8% turk-istanbul_hodoglugil_20" "17.8183"
[24,] "29.9% lahu_hgdp_8 + 70.1% turkmen_yunusbayev_11" "17.8439"
[25,] "34.7% tongan_xing_13 + 65.3% turkish_harappa_4" "17.8445"

FWIW, Tongan/Samoan are ~70% SE-Asian, ~30% Papuan. So I'm guessing that's something similar to Onge on a PCA.

Also interesting, the lots of East Eurasian is making it look like an East Asian-admixed Central Asian population. Karasuk did well in such a thing (admixture and IBS), but qpAdm doesn't seem to find an affinity. Or it could've been 50% of a population that's a cross between something from Oceania and further north.

Dr_McNinja
01-25-2016, 08:54 PM
Taking just the West Eurasian components (Baloch, Caucasian, NE-Euro, SW-Asian, Med, E-African), I get this:

[1,] "iranian_harappa_9" "12.5617"
[2,] "kurd_xing_24" "12.671"
[3,] "kurd_yunusbayev_6" "12.7512"
[4,] "kurd_harappa_11" "12.9147"
[5,] "iranian_behar_19" "13.4729"
[6,] "kumyk_yunusbayev_14" "13.8638"
[7,] "azeri_harappa_2" "14.0015"
[8,] "stalskoe_xing_5" "14.0373"
[9,] "lezgin_behar_18" "14.3815"
[10,] "urkarah_xing_18" "15.6192"
[11,] "turkish_harappa_4" "15.7879"
[12,] "uzbekistan-jew_behar_2" "16.4334"
[13,] "chechen_yunusbayev_20" "16.4771"
[14,] "turk-istanbul_hodoglugil_20" "16.9109"
[15,] "turk_behar_19" "17.0388"
[16,] "turk-kayseri_hodoglugil_23" "17.6351"
[17,] "armenian_harappa_2" "18.0836"
[18,] "turkmen_yunusbayev_11" "18.1456"
[19,] "iraqi-arab_harappa_4" "19.0502"
[20,] "balkar_yunusbayev_19" "19.9781"
[21,] "assyrian_harappa_6" "20.3139"
[22,] "adygei_hgdp_17" "20.3713"
[23,] "north-ossetian_yunusbayev_15" "20.4765"
[24,] "iraqi-mandaean_harappa_2" "20.5406"
[25,] "turk-aydin_hodoglugil_20" "20.8852"


[1,] "17.4% ethiopian-jew_behar_12 + 82.6% lezgin_behar_18" "5.9047"
[2,] "17.3% ethiopian_behar_19 + 82.7% lezgin_behar_18" "6.0329"
[3,] "81.9% lezgin_behar_18 + 18.1% tygray_pagani_21" "6.0753"
[4,] "17.7% amhara_pagani_25 + 82.3% lezgin_behar_18" "6.1977"
[5,] "17.2% afar_pagani_12 + 82.8% lezgin_behar_18" "6.4483"
[6,] "72.7% lezgin_behar_18 + 27.3% yemenese_behar_8" "6.7237"
[7,] "84.6% lezgin_behar_18 + 15.4% somali_harappa_2" "6.8115"
[8,] "83.7% lezgin_behar_18 + 16.3% oromo_pagani_21" "6.9443"
[9,] "79.3% lezgin_behar_18 + 20.7% qatari_henn2012_156" "6.9874"
[10,] "20.1% bedouin_hgdp_46 + 79.9% lezgin_behar_18" "7.1162"
[11,] "26.2% egyptian_behar_12 + 73.8% lezgin_behar_18" "7.2886"
[12,] "85.1% lezgin_behar_18 + 14.9% somali_pagani_23" "7.2912"
[13,] "14.9% esomali_pagani_17 + 85.1% lezgin_behar_18" "7.349"
[14,] "80.1% lezgin_behar_18 + 19.9% saudi_behar_20" "7.3626"
[15,] "85% lezgin_behar_18 + 15% somali_reich_6" "7.4261"
[16,] "71.2% urkarah_xing_18 + 28.8% yemenese_behar_8" "7.6485"
[17,] "43.8% iraqi-arab_harappa_4 + 56.2% urkarah_xing_18" "7.6579"
[18,] "85% lezgin_behar_18 + 15% wolayta_pagani_8" "7.7283"
[19,] "39.5% palestinian_harappa_2 + 60.5% urkarah_xing_18" "7.8199"
[20,] "27.9% egyptian_behar_12 + 72.1% urkarah_xing_18" "7.8264"
[21,] "78.7% lezgin_behar_18 + 21.3% yemen-jew_behar_15" "7.8653"
[22,] "24.8% egypt_henn2012_19 + 75.2% lezgin_behar_18" "7.8882"
[23,] "40.8% iraqi-arab_harappa_4 + 59.2% lezgin_behar_18" "8.0943"
[24,] "33.9% jordanian_behar_20 + 66.1% urkarah_xing_18" "8.1252"
[25,] "76.9% urkarah_xing_18 + 23.1% yemen-jew_behar_15" "8.173"


And,

[1,] "iranian_harappa_9" "14.4873"
[2,] "kurd_harappa_11" "14.6192"
[3,] "iranian_behar_19" "14.8024"
[4,] "kurd_xing_24" "15.805"
[5,] "stalskoe_xing_5" "15.9342"
[6,] "kurd_yunusbayev_6" "15.9534"
[7,] "azeri_harappa_2" "16.4366"
[8,] "turkish_harappa_4" "16.6373"
[9,] "turkmen_yunusbayev_11" "17.0223"
[10,] "kumyk_yunusbayev_14" "17.681"
[11,] "turk-istanbul_hodoglugil_20" "17.7344"
[12,] "turk-kayseri_hodoglugil_23" "18.4263"
[13,] "turk_behar_19" "18.4386"
[14,] "uzbekistan-jew_behar_2" "18.6355"
[15,] "iraqi-arab_harappa_4" "18.6444"
[16,] "lezgin_behar_18" "18.7055"
[17,] "urkarah_xing_18" "19.0265"
[18,] "turk-aydin_hodoglugil_20" "19.3282"
[19,] "armenian_harappa_2" "19.3483"
[20,] "chechen_yunusbayev_20" "20.9949"
[21,] "palestinian_harappa_2" "21.5653"
[22,] "tajik_yunusbayev_15" "22.1685"
[23,] "nogai_yunusbayev_16" "22.6389"
[24,] "iraqi-mandaean_harappa_2" "22.7308"
[25,] "syrian_behar_16" "22.9792"

My S-Indian had 44.67% of that and Sapporo had 47.42%. So it's close to half.

Dr_McNinja
01-25-2016, 09:05 PM
The East Eurasian part comes out as,

[1,] "44.1% mongolian_rasmussen_9 + 55.9% tongan_xing_13" "12.5352"
[2,] "41.6% daur_hgdp_9 + 58.4% tongan_xing_13" "12.5721"
[3,] "41.7% hezhen_hgdp_9 + 58.3% tongan_xing_13" "12.8817"
[4,] "46.1% kyrgyz_hodoglugil_16 + 53.9% tongan_xing_13" "13.0945"
[5,] "42.6% mongola_hgdp_10 + 57.4% tongan_xing_13" "13.5292"
[6,] "46.5% kyrgyz_xing_25 + 53.5% tongan_xing_13" "13.5817"
[7,] "58.5% tongan_xing_13 + 41.5% xibo_hgdp_9" "14.046"
[8,] "41.5% oroqen_hgdp_9 + 58.5% tongan_xing_13" "14.2564"
[9,] "36.2% aus_reich_3 + 63.8% lahu_hgdp_8" "14.3636"
[10,] "45.8% kazakh_harappa_5 + 54.2% tongan_xing_13" "14.4782"
[11,] "41.6% buryat_xing_25 + 58.4% tongan_xing_13" "15.1645"
[12,] "45% mongolian_rasmussen_9 + 55% samoan_xing_13" "15.468"
[13,] "44% altaian_rasmussen_13 + 56% tongan_xing_13" "15.5047"
[14,] "37% tibet_simonson_31 + 63% tongan_xing_13" "15.5322"
[15,] "42.3% daur_hgdp_9 + 57.7% samoan_xing_13" "15.6937"
[16,] "72.6% lahu_hgdp_8 + 27.4% papuan_hgdp_17" "15.827"
[17,] "47.1% kyrgyz_hodoglugil_16 + 52.9% samoan_xing_13" "15.8831"
[18,] "42.5% hezhen_hgdp_9 + 57.5% samoan_xing_13" "15.9843"
[19,] "37.4% aus_reich_3 + 62.6% vietnamese_xing_7" "16.0785"
[20,] "40.7% buryat_rasmussen_18 + 59.3% tongan_xing_13" "16.0936"
[21,] "53% tongan_xing_13 + 47% uyghur_hgdp_10" "16.1524"
[22,] "47.6% kyrgyz_xing_25 + 52.4% samoan_xing_13" "16.1988"
[23,] "42.4% nepalese-b_xing_6 + 57.6% tongan_xing_13" "16.4547"
[24,] "46.3% hazara_hgdp_22 + 53.7% tongan_xing_13" "16.4919"
[25,] "60.3% tongan_xing_13 + 39.7% tu_hgdp_10" "16.5315"


And,

[1,] "46.2% mongolian_rasmussen_9 + 53.8% tongan_xing_13" "10.5446"
[2,] "48.1% kyrgyz_hodoglugil_16 + 51.9% tongan_xing_13" "11.8045"
[3,] "44.1% buryat_xing_25 + 55.9% tongan_xing_13" "12.1699"
[4,] "48.6% kyrgyz_xing_25 + 51.4% tongan_xing_13" "12.2922"
[5,] "43.5% oroqen_hgdp_9 + 56.5% tongan_xing_13" "12.4484"
[6,] "46.5% altaian_rasmussen_13 + 53.5% tongan_xing_13" "12.8536"
[7,] "43.4% buryat_rasmussen_18 + 56.6% tongan_xing_13" "12.8912"
[8,] "47.9% kazakh_harappa_5 + 52.1% tongan_xing_13" "13.1217"
[9,] "46.9% mongolian_rasmussen_9 + 53.1% samoan_xing_13" "13.4126"
[10,] "42.6% daur_hgdp_9 + 57.4% tongan_xing_13" "13.4632"
[11,] "55.9% tongan_xing_13 + 44.1% tuvinian_rasmussen_14" "13.5322"
[12,] "42.7% hezhen_hgdp_9 + 57.3% tongan_xing_13" "13.9666"
[13,] "48.9% kyrgyz_hodoglugil_16 + 51.1% samoan_xing_13" "14.3961"
[14,] "49.5% kyrgyz_xing_25 + 50.5% samoan_xing_13" "14.7021"
[15,] "44.8% buryat_xing_25 + 55.2% samoan_xing_13" "14.881"
[16,] "44.2% oroqen_hgdp_9 + 55.8% samoan_xing_13" "15.2499"
[17,] "47.3% altaian_rasmussen_13 + 52.7% samoan_xing_13" "15.3252"
[18,] "44.1% buryat_rasmussen_18 + 55.9% samoan_xing_13" "15.4612"
[19,] "48.8% kazakh_harappa_5 + 51.2% samoan_xing_13" "15.5121"
[20,] "43.2% mongola_hgdp_10 + 56.8% tongan_xing_13" "15.6582"
[21,] "55.1% samoan_xing_13 + 44.9% tuvinian_rasmussen_14" "15.9935"
[22,] "43.3% daur_hgdp_9 + 56.7% samoan_xing_13" "16.0257"
[23,] "43.4% hezhen_hgdp_9 + 56.6% samoan_xing_13" "16.4947"
[24,] "58.2% tongan_xing_13 + 41.8% xibo_hgdp_9" "16.6247"
[25,] "47.8% hazara_hgdp_22 + 52.2% tongan_xing_13" "16.6801"

The issue is the ANE in the component, which is presumably why there's a shift towards the north.

Or that came in with the Caucasian-like component and it's 70-75% Karasuk-like and 25-30% Oceania/Polynesia.

surbakhunWeesste
01-25-2016, 09:34 PM
Taking just the West Eurasian components (Baloch, Caucasian, NE-Euro, SW-Asian, Med, E-African), I get this:

[1,] "iranian_harappa_9" "12.5617"
[2,] "kurd_xing_24" "12.671"
[3,] "kurd_yunusbayev_6" "12.7512"
[4,] "kurd_harappa_11" "12.9147"
[5,] "iranian_behar_19" "13.4729"
[6,] "kumyk_yunusbayev_14" "13.8638"
[7,] "azeri_harappa_2" "14.0015"
[8,] "stalskoe_xing_5" "14.0373"
[9,] "lezgin_behar_18" "14.3815"
[10,] "urkarah_xing_18" "15.6192"
[11,] "turkish_harappa_4" "15.7879"
[12,] "uzbekistan-jew_behar_2" "16.4334"
[13,] "chechen_yunusbayev_20" "16.4771"
[14,] "turk-istanbul_hodoglugil_20" "16.9109"
[15,] "turk_behar_19" "17.0388"
[16,] "turk-kayseri_hodoglugil_23" "17.6351"
[17,] "armenian_harappa_2" "18.0836"
[18,] "turkmen_yunusbayev_11" "18.1456"
[19,] "iraqi-arab_harappa_4" "19.0502"
[20,] "balkar_yunusbayev_19" "19.9781"
[21,] "assyrian_harappa_6" "20.3139"
[22,] "adygei_hgdp_17" "20.3713"
[23,] "north-ossetian_yunusbayev_15" "20.4765"
[24,] "iraqi-mandaean_harappa_2" "20.5406"
[25,] "turk-aydin_hodoglugil_20" "20.8852"


[1,] "17.4% ethiopian-jew_behar_12 + 82.6% lezgin_behar_18" "5.9047"
[2,] "17.3% ethiopian_behar_19 + 82.7% lezgin_behar_18" "6.0329"
[3,] "81.9% lezgin_behar_18 + 18.1% tygray_pagani_21" "6.0753"
[4,] "17.7% amhara_pagani_25 + 82.3% lezgin_behar_18" "6.1977"
[5,] "17.2% afar_pagani_12 + 82.8% lezgin_behar_18" "6.4483"
[6,] "72.7% lezgin_behar_18 + 27.3% yemenese_behar_8" "6.7237"
[7,] "84.6% lezgin_behar_18 + 15.4% somali_harappa_2" "6.8115"
[8,] "83.7% lezgin_behar_18 + 16.3% oromo_pagani_21" "6.9443"
[9,] "79.3% lezgin_behar_18 + 20.7% qatari_henn2012_156" "6.9874"
[10,] "20.1% bedouin_hgdp_46 + 79.9% lezgin_behar_18" "7.1162"
[11,] "26.2% egyptian_behar_12 + 73.8% lezgin_behar_18" "7.2886"
[12,] "85.1% lezgin_behar_18 + 14.9% somali_pagani_23" "7.2912"
[13,] "14.9% esomali_pagani_17 + 85.1% lezgin_behar_18" "7.349"
[14,] "80.1% lezgin_behar_18 + 19.9% saudi_behar_20" "7.3626"
[15,] "85% lezgin_behar_18 + 15% somali_reich_6" "7.4261"
[16,] "71.2% urkarah_xing_18 + 28.8% yemenese_behar_8" "7.6485"
[17,] "43.8% iraqi-arab_harappa_4 + 56.2% urkarah_xing_18" "7.6579"
[18,] "85% lezgin_behar_18 + 15% wolayta_pagani_8" "7.7283"
[19,] "39.5% palestinian_harappa_2 + 60.5% urkarah_xing_18" "7.8199"
[20,] "27.9% egyptian_behar_12 + 72.1% urkarah_xing_18" "7.8264"
[21,] "78.7% lezgin_behar_18 + 21.3% yemen-jew_behar_15" "7.8653"
[22,] "24.8% egypt_henn2012_19 + 75.2% lezgin_behar_18" "7.8882"
[23,] "40.8% iraqi-arab_harappa_4 + 59.2% lezgin_behar_18" "8.0943"
[24,] "33.9% jordanian_behar_20 + 66.1% urkarah_xing_18" "8.1252"
[25,] "76.9% urkarah_xing_18 + 23.1% yemen-jew_behar_15" "8.173"


And,

[1,] "iranian_harappa_9" "14.4873"
[2,] "kurd_harappa_11" "14.6192"
[3,] "iranian_behar_19" "14.8024"
[4,] "kurd_xing_24" "15.805"
[5,] "stalskoe_xing_5" "15.9342"
[6,] "kurd_yunusbayev_6" "15.9534"
[7,] "azeri_harappa_2" "16.4366"
[8,] "turkish_harappa_4" "16.6373"
[9,] "turkmen_yunusbayev_11" "17.0223"
[10,] "kumyk_yunusbayev_14" "17.681"
[11,] "turk-istanbul_hodoglugil_20" "17.7344"
[12,] "turk-kayseri_hodoglugil_23" "18.4263"
[13,] "turk_behar_19" "18.4386"
[14,] "uzbekistan-jew_behar_2" "18.6355"
[15,] "iraqi-arab_harappa_4" "18.6444"
[16,] "lezgin_behar_18" "18.7055"
[17,] "urkarah_xing_18" "19.0265"
[18,] "turk-aydin_hodoglugil_20" "19.3282"
[19,] "armenian_harappa_2" "19.3483"
[20,] "chechen_yunusbayev_20" "20.9949"
[21,] "palestinian_harappa_2" "21.5653"
[22,] "tajik_yunusbayev_15" "22.1685"
[23,] "nogai_yunusbayev_16" "22.6389"
[24,] "iraqi-mandaean_harappa_2" "22.7308"
[25,] "syrian_behar_16" "22.9792"

My S-Indian had 44.67% of that and Sapporo had 47.42%. So it's close to half.

So you and Sap are ‘iranic’ like without the SI? This is remarkable.

Dr_McNinja
01-25-2016, 09:55 PM
So you and Sap are ‘iranic’ like without the SI? This is remarkable.
No, more like South Indian admixture itself as represented by the S-Indian component can be construed as half like some kind of archaic Iranic.

Dr_McNinja
01-25-2016, 10:08 PM
Tried making a "Dravidian" component out of that West Eurasian part of HAP S-Indian, this is what it showed up in:

(Baloch/Caucasian components were still there from original calculator, so they went up a little too)

Me: 10.69%
HRP0393 Haryana: 9.38%
Sapporo: 11.12%
South Italian: 10.79%
Assyrian (Zeph): 4.28%
Sweden: 3.34%
Caucasus (karachay): 3.49%
Iranian (DMXX): 11.04%
longbowman: 10.98%
parasar: 9.46%
abbas (Iranian of some kind?): 13.07%
ashwin: 9.99%
Turk (dogukan): 3.73%
Turk (Hanna): 7.08%
Anabasis: 2.98%
DLuffy: 10.71%
Chad: 3.99%
Iranian (NK19191): 12.10%
Bored: 9.73%
Pashtun/Tajik (WN): 11.48%
Sein's friend: 10.36%
N-African Berber: 6.37%
Khetran Baloch: 12.38%
Everest: 8.91%
Rukha: 12.00%
Sein: 12.10%
HRP0370 Pashtun: 12.89%
Soulblighter (South Indian Brahmin): 9.88%
Reza (Bengali): 7.34%

Doesn't work if I put in a simulated CHG component instead of Baloch/Caucasian, everything gets taken in by CHG.

surbakhunWeesste
01-25-2016, 10:47 PM
No, more like South Indian admixture itself as represented by the S-Indian component can be construed as half like some kind of archaic Iranic.

Ah cool and thanks, can you do it for Rukha? or just any other Afghan Pashtun whose data you have hands on?

tamilgangster
01-25-2016, 11:33 PM
14.48% NE-Euro
1.10% SE-Asian
3.59% Siberian
6.47% NE-Asian
3.71% Papuan
2.31% American
0.98% Beringian
0.30% Mediterranean
2.59% SW-Asian
0.00% San
0.57% E-African
0.30% Pygmy
0.28% W-African
63.31% CHG

whats ur ethinicity

tamilgangster
01-25-2016, 11:44 PM
Taking just the West Eurasian components (Baloch, Caucasian, NE-Euro, SW-Asian, Med, E-African), I get this:

[1,] "iranian_harappa_9" "12.5617"
[2,] "kurd_xing_24" "12.671"
[3,] "kurd_yunusbayev_6" "12.7512"
[4,] "kurd_harappa_11" "12.9147"
[5,] "iranian_behar_19" "13.4729"
[6,] "kumyk_yunusbayev_14" "13.8638"
[7,] "azeri_harappa_2" "14.0015"
[8,] "stalskoe_xing_5" "14.0373"
[9,] "lezgin_behar_18" "14.3815"
[10,] "urkarah_xing_18" "15.6192"
[11,] "turkish_harappa_4" "15.7879"
[12,] "uzbekistan-jew_behar_2" "16.4334"
[13,] "chechen_yunusbayev_20" "16.4771"
[14,] "turk-istanbul_hodoglugil_20" "16.9109"
[15,] "turk_behar_19" "17.0388"
[16,] "turk-kayseri_hodoglugil_23" "17.6351"
[17,] "armenian_harappa_2" "18.0836"
[18,] "turkmen_yunusbayev_11" "18.1456"
[19,] "iraqi-arab_harappa_4" "19.0502"
[20,] "balkar_yunusbayev_19" "19.9781"
[21,] "assyrian_harappa_6" "20.3139"
[22,] "adygei_hgdp_17" "20.3713"
[23,] "north-ossetian_yunusbayev_15" "20.4765"
[24,] "iraqi-mandaean_harappa_2" "20.5406"
[25,] "turk-aydin_hodoglugil_20" "20.8852"


[1,] "17.4% ethiopian-jew_behar_12 + 82.6% lezgin_behar_18" "5.9047"
[2,] "17.3% ethiopian_behar_19 + 82.7% lezgin_behar_18" "6.0329"
[3,] "81.9% lezgin_behar_18 + 18.1% tygray_pagani_21" "6.0753"
[4,] "17.7% amhara_pagani_25 + 82.3% lezgin_behar_18" "6.1977"
[5,] "17.2% afar_pagani_12 + 82.8% lezgin_behar_18" "6.4483"
[6,] "72.7% lezgin_behar_18 + 27.3% yemenese_behar_8" "6.7237"
[7,] "84.6% lezgin_behar_18 + 15.4% somali_harappa_2" "6.8115"
[8,] "83.7% lezgin_behar_18 + 16.3% oromo_pagani_21" "6.9443"
[9,] "79.3% lezgin_behar_18 + 20.7% qatari_henn2012_156" "6.9874"
[10,] "20.1% bedouin_hgdp_46 + 79.9% lezgin_behar_18" "7.1162"
[11,] "26.2% egyptian_behar_12 + 73.8% lezgin_behar_18" "7.2886"
[12,] "85.1% lezgin_behar_18 + 14.9% somali_pagani_23" "7.2912"
[13,] "14.9% esomali_pagani_17 + 85.1% lezgin_behar_18" "7.349"
[14,] "80.1% lezgin_behar_18 + 19.9% saudi_behar_20" "7.3626"
[15,] "85% lezgin_behar_18 + 15% somali_reich_6" "7.4261"
[16,] "71.2% urkarah_xing_18 + 28.8% yemenese_behar_8" "7.6485"
[17,] "43.8% iraqi-arab_harappa_4 + 56.2% urkarah_xing_18" "7.6579"
[18,] "85% lezgin_behar_18 + 15% wolayta_pagani_8" "7.7283"
[19,] "39.5% palestinian_harappa_2 + 60.5% urkarah_xing_18" "7.8199"
[20,] "27.9% egyptian_behar_12 + 72.1% urkarah_xing_18" "7.8264"
[21,] "78.7% lezgin_behar_18 + 21.3% yemen-jew_behar_15" "7.8653"
[22,] "24.8% egypt_henn2012_19 + 75.2% lezgin_behar_18" "7.8882"
[23,] "40.8% iraqi-arab_harappa_4 + 59.2% lezgin_behar_18" "8.0943"
[24,] "33.9% jordanian_behar_20 + 66.1% urkarah_xing_18" "8.1252"
[25,] "76.9% urkarah_xing_18 + 23.1% yemen-jew_behar_15" "8.173"


And,

[1,] "iranian_harappa_9" "14.4873"
[2,] "kurd_harappa_11" "14.6192"
[3,] "iranian_behar_19" "14.8024"
[4,] "kurd_xing_24" "15.805"
[5,] "stalskoe_xing_5" "15.9342"
[6,] "kurd_yunusbayev_6" "15.9534"
[7,] "azeri_harappa_2" "16.4366"
[8,] "turkish_harappa_4" "16.6373"
[9,] "turkmen_yunusbayev_11" "17.0223"
[10,] "kumyk_yunusbayev_14" "17.681"
[11,] "turk-istanbul_hodoglugil_20" "17.7344"
[12,] "turk-kayseri_hodoglugil_23" "18.4263"
[13,] "turk_behar_19" "18.4386"
[14,] "uzbekistan-jew_behar_2" "18.6355"
[15,] "iraqi-arab_harappa_4" "18.6444"
[16,] "lezgin_behar_18" "18.7055"
[17,] "urkarah_xing_18" "19.0265"
[18,] "turk-aydin_hodoglugil_20" "19.3282"
[19,] "armenian_harappa_2" "19.3483"
[20,] "chechen_yunusbayev_20" "20.9949"
[21,] "palestinian_harappa_2" "21.5653"
[22,] "tajik_yunusbayev_15" "22.1685"
[23,] "nogai_yunusbayev_16" "22.6389"
[24,] "iraqi-mandaean_harappa_2" "22.7308"
[25,] "syrian_behar_16" "22.9792"

My S-Indian had 44.67% of that and Sapporo had 47.42%. So it's close to half.

This finally explains where the east african "noise" comes from, the west eurasian portion of the South Indian component, includes some CHG but also must be basal eruasian rich, due to it having a bedouin like component

tamilgangster
01-25-2016, 11:45 PM
The East Eurasian part comes out as,

[1,] "44.1% mongolian_rasmussen_9 + 55.9% tongan_xing_13" "12.5352"
[2,] "41.6% daur_hgdp_9 + 58.4% tongan_xing_13" "12.5721"
[3,] "41.7% hezhen_hgdp_9 + 58.3% tongan_xing_13" "12.8817"
[4,] "46.1% kyrgyz_hodoglugil_16 + 53.9% tongan_xing_13" "13.0945"
[5,] "42.6% mongola_hgdp_10 + 57.4% tongan_xing_13" "13.5292"
[6,] "46.5% kyrgyz_xing_25 + 53.5% tongan_xing_13" "13.5817"
[7,] "58.5% tongan_xing_13 + 41.5% xibo_hgdp_9" "14.046"
[8,] "41.5% oroqen_hgdp_9 + 58.5% tongan_xing_13" "14.2564"
[9,] "36.2% aus_reich_3 + 63.8% lahu_hgdp_8" "14.3636"
[10,] "45.8% kazakh_harappa_5 + 54.2% tongan_xing_13" "14.4782"
[11,] "41.6% buryat_xing_25 + 58.4% tongan_xing_13" "15.1645"
[12,] "45% mongolian_rasmussen_9 + 55% samoan_xing_13" "15.468"
[13,] "44% altaian_rasmussen_13 + 56% tongan_xing_13" "15.5047"
[14,] "37% tibet_simonson_31 + 63% tongan_xing_13" "15.5322"
[15,] "42.3% daur_hgdp_9 + 57.7% samoan_xing_13" "15.6937"
[16,] "72.6% lahu_hgdp_8 + 27.4% papuan_hgdp_17" "15.827"
[17,] "47.1% kyrgyz_hodoglugil_16 + 52.9% samoan_xing_13" "15.8831"
[18,] "42.5% hezhen_hgdp_9 + 57.5% samoan_xing_13" "15.9843"
[19,] "37.4% aus_reich_3 + 62.6% vietnamese_xing_7" "16.0785"
[20,] "40.7% buryat_rasmussen_18 + 59.3% tongan_xing_13" "16.0936"
[21,] "53% tongan_xing_13 + 47% uyghur_hgdp_10" "16.1524"
[22,] "47.6% kyrgyz_xing_25 + 52.4% samoan_xing_13" "16.1988"
[23,] "42.4% nepalese-b_xing_6 + 57.6% tongan_xing_13" "16.4547"
[24,] "46.3% hazara_hgdp_22 + 53.7% tongan_xing_13" "16.4919"
[25,] "60.3% tongan_xing_13 + 39.7% tu_hgdp_10" "16.5315"


And,

[1,] "46.2% mongolian_rasmussen_9 + 53.8% tongan_xing_13" "10.5446"
[2,] "48.1% kyrgyz_hodoglugil_16 + 51.9% tongan_xing_13" "11.8045"
[3,] "44.1% buryat_xing_25 + 55.9% tongan_xing_13" "12.1699"
[4,] "48.6% kyrgyz_xing_25 + 51.4% tongan_xing_13" "12.2922"
[5,] "43.5% oroqen_hgdp_9 + 56.5% tongan_xing_13" "12.4484"
[6,] "46.5% altaian_rasmussen_13 + 53.5% tongan_xing_13" "12.8536"
[7,] "43.4% buryat_rasmussen_18 + 56.6% tongan_xing_13" "12.8912"
[8,] "47.9% kazakh_harappa_5 + 52.1% tongan_xing_13" "13.1217"
[9,] "46.9% mongolian_rasmussen_9 + 53.1% samoan_xing_13" "13.4126"
[10,] "42.6% daur_hgdp_9 + 57.4% tongan_xing_13" "13.4632"
[11,] "55.9% tongan_xing_13 + 44.1% tuvinian_rasmussen_14" "13.5322"
[12,] "42.7% hezhen_hgdp_9 + 57.3% tongan_xing_13" "13.9666"
[13,] "48.9% kyrgyz_hodoglugil_16 + 51.1% samoan_xing_13" "14.3961"
[14,] "49.5% kyrgyz_xing_25 + 50.5% samoan_xing_13" "14.7021"
[15,] "44.8% buryat_xing_25 + 55.2% samoan_xing_13" "14.881"
[16,] "44.2% oroqen_hgdp_9 + 55.8% samoan_xing_13" "15.2499"
[17,] "47.3% altaian_rasmussen_13 + 52.7% samoan_xing_13" "15.3252"
[18,] "44.1% buryat_rasmussen_18 + 55.9% samoan_xing_13" "15.4612"
[19,] "48.8% kazakh_harappa_5 + 51.2% samoan_xing_13" "15.5121"
[20,] "43.2% mongola_hgdp_10 + 56.8% tongan_xing_13" "15.6582"
[21,] "55.1% samoan_xing_13 + 44.9% tuvinian_rasmussen_14" "15.9935"
[22,] "43.3% daur_hgdp_9 + 56.7% samoan_xing_13" "16.0257"
[23,] "43.4% hezhen_hgdp_9 + 56.6% samoan_xing_13" "16.4947"
[24,] "58.2% tongan_xing_13 + 41.8% xibo_hgdp_9" "16.6247"
[25,] "47.8% hazara_hgdp_22 + 52.2% tongan_xing_13" "16.6801"

The issue is the ANE in the component, which is presumably why there's a shift towards the north.

Or that came in with the Caucasian-like component and it's 70-75% Karasuk-like and 25-30% Oceania/Polynesia.

ALot of ANE in the south Indian component is getting mixed with the east eurasian present creating a pseudo siberian like vibe

DMXX
01-27-2016, 03:27 PM
So you and Sap are ‘iranic’ like without the SI? This is remarkable.

Specifically, West Iranic (or Iranian-speaking West Asian).

This reminds me of the ADMIXTURE component rebalancing one may observe with the Uyghurs. Once their East Eurasian admixture is removed and the West/South Eurasian components are maximised, they essentially resemble modern Pamiris.

I've suspected for a while that, by the time of Christ, the migration trajectories common to both populations (as well as ensuing bidirectional movements prior to and after the Silk Road's founding) had reached their zenith.

The ramifications of this are quite interesting if true. It would mean that the swathe of territory from the Uzbek highlands through to the Tarim was a distinct biogeographical zone.

jesus
01-27-2016, 05:43 PM
Specifically, West Iranic (or Iranian-speaking West Asian).

This reminds me of the ADMIXTURE component rebalancing one may observe with the Uyghurs. Once their East Eurasian admixture is removed and the West/South Eurasian components are maximised, they essentially resemble modern Pamiris.

I've suspected for a while that, by the time of Christ, the migration trajectories common to both populations (as well as ensuing bidirectional movements prior to and after the Silk Road's founding) had reached their zenith.

The ramifications of this are quite interesting if true. It would mean that the swathe of territory from the Uzbek highlands through to the Tarim was a distinct biogeographical zone.

Also many Hazaras can be modeled as 1/2 Mongol/Kazakh + 1/2 Afghan Pashtun, suggesting that the pre Mongol local population was similar to Afghan Pashtuns. Those hazaras didn't have close Pashtun matches on Gedmatch, so the liklihood of them having recent Pashtun or Pashtun type ancestry (as in the last 100-200 years) is pretty low. The pre Mongol population probably spoke some east Iranic dialects or were of east iranic origins, but Farsi was spoken as a main language as well.

DMXX
01-27-2016, 05:52 PM
Also many Hazaras can be modeled as 1/2 Mongol/Kazakh + 1/2 Afghan Pashtun, suggesting that the pre Mongol local population was similar to Afghan Pashtuns. Those hazaras didn't have close Pashtun matches on Gedmatch, so the liklihood of them having recent Pashtun or Pashtun type ancestry (as in the last 100-200 years) is pretty low. The pre Mongol population probably spoke some east Iranic dialects or were of east iranic origins, but Farsi was spoken as a main language as well.

Completely agree. As a side note, I've considered this fairly consistent circumstantial evidence that the early Turkish steppe groups were predominantly East Eurasian autosomally.

Shaikorth
01-27-2016, 08:35 PM
Also many Hazaras can be modeled as 1/2 Mongol/Kazakh + 1/2 Afghan Pashtun, suggesting that the pre Mongol local population was similar to Afghan Pashtuns. Those hazaras didn't have close Pashtun matches on Gedmatch, so the liklihood of them having recent Pashtun or Pashtun type ancestry (as in the last 100-200 years) is pretty low. The pre Mongol population probably spoke some east Iranic dialects or were of east iranic origins, but Farsi was spoken as a main language as well.


Completely agree. As a side note, I've considered this fairly consistent circumstantial evidence that the early Turkish steppe groups were predominantly East Eurasian autosomally.

Razib's Treemix runs pretty consistently modeled Uygurs and Hazaras as North Euro + basal East Eurasian. This is probably catching something dating back to Bronze Age at least, unless there were Andronovo and Sintashta-like groups alive during the Turkic expansions which would be very surprising. Tajiks do sound like a better fit for the actual pre-Turkic population there.

http://www.unzcloud.com/wp-content/uploads/2015/08/KalashOut.71.jpg

http://www.unzcloud.com/wp-content/uploads/2015/08/KalashOut.8.jpg

surbakhunWeesste
01-27-2016, 08:51 PM
Also many Hazaras can be modeled as 1/2 Mongol/Kazakh + 1/2 Afghan Pashtun, suggesting that the pre Mongol local population was similar to Afghan Pashtuns. Those hazaras didn't have close Pashtun matches on Gedmatch, so the liklihood of them having recent Pashtun or Pashtun type ancestry (as in the last 100-200 years) is pretty low. The pre Mongol population probably spoke some east Iranic dialects or were of east iranic origins, but Farsi was spoken as a main language as well.


https://www.youtube.com/watch?v=2isXwDyVE9s

Dr_McNinja
01-27-2016, 10:32 PM
Ah cool and thanks, can you do it for Rukha? or just any other Afghan Pashtun whose data you have hands on?

Did Rukha's and it appears near identical in overall admixture but slightly more shows up as Mediterranean than Caucasian/NE-Euro/SW-Asian in the West Eurasian part and more shows up as NE-Asian than SE-Asian in the East Eurasian part. The other components are spread the same though.

Dr_McNinja
01-27-2016, 10:51 PM
Did Rukha's and it appears near identical in overall admixture but slightly more shows up as Mediterranean than Caucasian/NE-Euro/SW-Asian in the West Eurasian part and more shows up as NE-Asian than SE-Asian in the East Eurasian part. The other components are spread the same though.

This just looks very Iranian/Kurdish. It could be the African stuff is archaic and related to ASI, and not from the Dravidian era. Because ignoring the African gives you this:

McNinja: 27.78% Baloch, 44.13% Caucasian, 15.63% NE-Euro, 12.45% SW-Asian
Sapporo: 27.54% Baloch, 39.52% Caucasian, 15.74% NE-Euro, 5% Med, 12.20% SW-Asian
Rukha: 31.23% Baloch, 39.64% Caucasian, 13.45% NE-Euro, 17.3% Med, 8.72% SW-Asian

(This is roughly 30-40% of the S-Indian component)

I don't know if the NE-Euro/Med represents any significant WHG/EHG affinity, could just be a weak signal in those geographic directions (since they're mostly Caucasian-like anyhow). Or it could be legit and the West Iranian population which originated the Dravidians had some Steppe ancestry.

Combined with the strong Siberia/Mongolia signal, it's no surprise I got a minimum ~15-20% Karasuk-like layer across all South Asia in the absence of a South Asian component.

Dr_McNinja
01-27-2016, 11:23 PM
This just looks very Iranian/Kurdish. It could be the African stuff is archaic and related to ASI, and not from the Dravidian era. Because ignoring the African gives you this:

McNinja: 27.78% Baloch, 44.13% Caucasian, 15.63% NE-Euro, 12.45% SW-Asian
Sapporo: 27.54% Baloch, 39.52% Caucasian, 15.74% NE-Euro, 5% Med, 12.20% SW-Asian
Rukha: 31.23% Baloch, 39.64% Caucasian, 13.45% NE-Euro, 17.3% Med, 8.72% SW-Asian

(This is roughly 30-40% of the S-Indian component)

I don't know if the NE-Euro/Med represents any significant WHG/EHG affinity, could just be a weak signal in those geographic directions (since they're mostly Caucasian-like anyhow). Or it could be legit and the West Iranian population which originated the Dravidians had some Steppe ancestry.

Combined with the strong Siberia/Mongolia signal, it's no surprise I got a minimum ~15-20% Karasuk-like layer across all South Asia in the absence of a South Asian component.

soulblighter (South Indian Brahmin) - 25.78% Baloch, 41.9% Caucasian, 19.47% NE-Euro, 12.82% SW-Asian
Reza (Bengali) - 30.36% Baloch, 40.96% Caucasian, 18.74% NE-Euro, 9.94% SW-Asian

East Eurasian is generally similar, exactly 24% Papuan (from the non-West Eurasian part) for almost everyone. SE-Asian around 30%. NE-Asian/Siberian/etc varied a bit.

Dr_McNinja
01-28-2016, 02:58 AM
soulblighter (South Indian Brahmin) - 25.78% Baloch, 41.9% Caucasian, 19.47% NE-Euro, 12.82% SW-Asian
Reza (Bengali) - 30.36% Baloch, 40.96% Caucasian, 18.74% NE-Euro, 9.94% SW-Asian

East Eurasian is generally similar, exactly 24% Papuan (from the non-West Eurasian part) for almost everyone. SE-Asian around 30%. NE-Asian/Siberian/etc varied a bit.

So if you remove NE-Euro from that, it's Kurdish in the Oracle with a GD of single digits.

When I put in a component for the Xing Kurdish averages, I get this:

31.38% Baloch
5.37% Caucasian
10.33% NE-Euro
6.41% SE-Asian
3.73% Siberian
4.02% NE-Asian
5.52% Papuan
1.79% American
2.33% Beringian
0.00% Mediterranean
0.05% SW-Asian
0.00% San
1.29% E-African
0.67% Pygmy
0.06% W-African
27.04% W-Iranian

Sapporo:

36.96% Baloch
4.71% Caucasian
12.55% NE-Euro
4.95% SE-Asian
3.71% Siberian
2.51% NE-Asian
3.40% Papuan
1.66% American
1.90% Beringian
0.00% Mediterranean
0.29% SW-Asian
0.00% San
1.67% E-African
0.00% Pygmy
0.00% W-African
25.71% W-Iranian

HRP0393:

36.46% Baloch
10.01% Caucasian
19.19% NE-Euro
5.30% SE-Asian
3.28% Siberian
3.27% NE-Asian
3.86% Papuan
2.12% American
0.55% Beringian
2.06% Mediterranean
0.70% SW-Asian
0.63% San
0.77% E-African
0.14% Pygmy
0.68% W-African
10.97% W-Iranian

Zeph (Assyrian):

12.09% Baloch
48.58% Caucasian
0.75% NE-Euro
0.02% SE-Asian
0.08% Siberian
0.02% NE-Asian
0.61% Papuan
0.00% American
0.08% Beringian
7.32% Mediterranean
15.18% SW-Asian
0.00% San
0.01% E-African
0.00% Pygmy
0.00% W-African
15.26% W-Iranian

South Italian:

5.63% Baloch
25.05% Caucasian
18.43% NE-Euro
0.10% SE-Asian
0.04% Siberian
0.06% NE-Asian
0.21% Papuan
0.04% American
0.05% Beringian
26.00% Mediterranean
11.89% SW-Asian
0.13% San
0.44% E-African
0.08% Pygmy
0.18% W-African
11.68% W-Iranian

Sweden:

7.22% Baloch
3.79% Caucasian
57.83% NE-Euro
0.46% SE-Asian
0.76% Siberian
0.61% NE-Asian
0.24% Papuan
0.13% American
0.45% Beringian
22.81% Mediterranean
1.29% SW-Asian
0.04% San
0.01% E-African
0.00% Pygmy
0.00% W-African
4.35% W-Iranian

Iranian (DMXX):

21.05% Baloch
32.11% Caucasian
6.04% NE-Euro
1.07% SE-Asian
1.46% Siberian
1.98% NE-Asian
1.26% Papuan
0.35% American
1.15% Beringian
5.84% Mediterranean
9.44% SW-Asian
0.01% San
0.26% E-African
0.07% Pygmy
0.07% W-African
17.85% W-Iranian

abbas (Iranian?)

25.40% Baloch
36.19% Caucasian
2.95% NE-Euro
0.83% SE-Asian
1.01% Siberian
0.66% NE-Asian
0.41% Papuan
0.33% American
0.33% Beringian
2.14% Mediterranean
10.79% SW-Asian
0.04% San
0.54% E-African
0.04% Pygmy
0.20% W-African
18.16% W-Iranian

Parasar (Bihar Brahmin)

34.74% Baloch
2.05% Caucasian
12.08% NE-Euro
5.38% SE-Asian
1.22% Siberian
8.64% NE-Asian
6.54% Papuan
2.36% American
0.49% Beringian
0.11% Mediterranean
0.08% SW-Asian
0.00% San
1.67% E-African
0.00% Pygmy
0.27% W-African
24.36% W-Iranian

ashwin (Gujarat/Maharashtra Brahmin)

33.18% Baloch
2.43% Caucasian
7.68% NE-Euro
5.55% SE-Asian
4.00% Siberian
9.41% NE-Asian
7.20% Papuan
1.16% American
0.33% Beringian
0.00% Mediterranean
0.08% SW-Asian
0.00% San
2.08% E-African
0.00% Pygmy
0.96% W-African
25.94% W-Iranian

Turk (dogukan):

15.35% Baloch
33.42% Caucasian
4.51% NE-Euro
2.23% SE-Asian
4.41% Siberian
3.64% NE-Asian
0.54% Papuan
1.05% American
0.47% Beringian
7.70% Mediterranean
10.21% SW-Asian
0.01% San
0.02% E-African
0.01% Pygmy
0.00% W-African
16.43% W-Iranian

Turk (Hanna):

12.19% Baloch
46.57% Caucasian
5.44% NE-Euro
0.42% SE-Asian
2.63% Siberian
1.56% NE-Asian
0.41% Papuan
0.29% American
0.42% Beringian
8.13% Mediterranean
5.63% SW-Asian
0.01% San
0.11% E-African
0.01% Pygmy
0.01% W-African
16.17% W-Iranian

DLuffy (Punjabi Tarkhan):

34.38% Baloch
6.98% Caucasian
13.71% NE-Euro
4.04% SE-Asian
3.58% Siberian
6.36% NE-Asian
4.75% Papuan
1.55% American
1.29% Beringian
0.00% Mediterranean
0.11% SW-Asian
0.00% San
1.52% E-African
0.00% Pygmy
0.16% W-African
21.58% W-Iranian

My mom's cousin:

32.35% Baloch
1.92% Caucasian
10.10% NE-Euro
6.35% SE-Asian
1.76% Siberian
3.88% NE-Asian
5.66% Papuan
2.16% American
1.86% Beringian
0.00% Mediterranean
0.33% SW-Asian
0.03% San
0.92% E-African
0.90% Pygmy
0.00% W-African
31.76% W-Iranian

Khetran Baloch

36.37% Baloch
7.08% Caucasian
7.02% NE-Euro
3.12% SE-Asian
2.92% Siberian
1.34% NE-Asian
3.27% Papuan
1.69% American
0.78% Beringian
0.00% Mediterranean
1.05% SW-Asian
0.07% San
0.84% E-African
0.83% Pygmy
0.30% W-African
33.33% W-Iranian

This isn't happening for Pashtun (or South Asians like Bored or Everest), just most South Asians.

EDIT: My dad:

29.81% Baloch
2.85% Caucasian
10.87% NE-Euro
5.38% SE-Asian
2.72% Siberian
6.02% NE-Asian
4.85% Papuan
1.53% American
1.93% Beringian
0.00% Mediterranean
0.13% SW-Asian
0.58% San
1.03% E-African
0.36% Pygmy
0.00% W-African
31.94% W-Iranian

Dr_McNinja
01-28-2016, 05:08 AM
It worked on HRP0402, HRP0350 (both Punjabi Jatt - Pahari area) but not HRP0353 (Punjabi Gujjar).

It also worked on HRP0370 (Pashtun):

25.45% Baloch
7.49% Caucasian
12.87% NE-Euro
1.44% SE-Asian
2.83% Siberian
3.59% NE-Asian
1.22% Papuan
1.09% American
0.55% Beringian
0.78% Mediterranean
0.30% SW-Asian
0.17% San
0.72% E-African
0.00% Pygmy
0.00% W-African
41.48% W-Iranian

tamilgangster
01-28-2016, 10:01 AM
When removing S-Indian from Harappa, this is where the admixture went:

Me (37.18% S-Indian):

Baloch - 11.51%
Caucasian - 18.29%
NE-Euro - 6.48%
SE-Asian - 17.43%
Siberian - 5.92%
NE-Asian - 11.43%
Papuan - 12.93%
Amerindian - 2.64%
Med - 0.0%
Beringian - 1.99%
SW-Asian - 5.16%
San - 0.38%
E-African - 3.23%
Pygmy - 1.48%
W-African - 1.10%

Sapporo (30.15% S-Indian)

Baloch - 11.67%
Caucasian - 16.75%
NE-Euro - 6.67%
SE-Asian - 16.29%
Siberian - 8.13%
NE-Asian - 9.62%
Papuan - 11.70%
Amerindian - 2.95%
Beringian - 2.82%
Med - 2.12%
SW-Asian - 5.17%
San - 0.0%
E-African - 5.04%
Pygmy - 0.50%
W-African - 0.53%

For kicks, I put that into Harappa Oracle.

Mine:

[1,] "31.5% tongan_xing_13 + 68.5% turkmen_yunusbayev_11" "10.9646"
[2,] "31.3% tongan_xing_13 + 68.7% uzbek_behar_15" "12.1198"
[3,] "30.9% samoan_xing_13 + 69.1% turkmen_yunusbayev_11" "12.1334"
[4,] "66.4% nogai_yunusbayev_16 + 33.6% tongan_xing_13" "12.399"
[5,] "30.7% samoan_xing_13 + 69.3% uzbek_behar_15" "13.1783"
[6,] "67% nogai_yunusbayev_16 + 33% samoan_xing_13" "13.5296"
[7,] "34% tongan_xing_13 + 66% turk-aydin_hodoglugil_20" "14.7725"
[8,] "66.6% tajik_yunusbayev_15 + 33.4% tongan_xing_13" "15.0488"
[9,] "32.1% tongan_xing_13 + 67.9% uyghur_hgdp_10" "15.2025"
[10,] "67.1% hazara_hgdp_22 + 32.9% tongan_xing_13" "15.4729"
[11,] "33.4% samoan_xing_13 + 66.6% turk-aydin_hodoglugil_20" "15.766"
[12,] "32.8% samoan_xing_13 + 67.2% tajik_yunusbayev_15" "15.9418"
[13,] "31.4% samoan_xing_13 + 68.6% uyghur_hgdp_10" "16.0797"
[14,] "67.7% hazara_hgdp_22 + 32.3% samoan_xing_13" "16.3061"
[15,] "36.2% tongan_xing_13 + 63.8% turk-istanbul_hodoglugil_20" "16.3094"
[16,] "63.9% iranian_harappa_9 + 36.1% tongan_xing_13" "16.407"
[17,] "35.8% tongan_xing_13 + 64.2% turkish_harappa_4" "16.4204"
[18,] "63.2% azeri_harappa_2 + 36.8% tongan_xing_13" "16.4356"
[19,] "36.4% tongan_xing_13 + 63.6% turk-kayseri_hodoglugil_23" "16.8103"
[20,] "36.9% tongan_xing_13 + 63.1% turk_behar_19" "16.819"
[21,] "63.6% kurd_harappa_11 + 36.4% tongan_xing_13" "16.9057"
[22,] "63.6% iranian_behar_19 + 36.4% tongan_xing_13" "16.9844"
[23,] "62.2% kumyk_yunusbayev_14 + 37.8% tongan_xing_13" "17.0419"
[24,] "22.4% melanesian_hgdp_10 + 77.6% turkmen_yunusbayev_11" "17.105"
[25,] "63.9% stalskoe_xing_5 + 36.1% tongan_xing_13" "17.2112"

Sapporo:

[1,] "29.6% tongan_xing_13 + 70.4% uzbek_behar_15" "11.6242"
[2,] "30.1% tongan_xing_13 + 69.9% turkmen_yunusbayev_11" "11.658"
[3,] "29.1% samoan_xing_13 + 70.9% uzbek_behar_15" "12.5095"
[4,] "29.6% samoan_xing_13 + 70.4% turkmen_yunusbayev_11" "12.5572"
[5,] "67.7% nogai_yunusbayev_16 + 32.3% tongan_xing_13" "13.319"
[6,] "68.2% nogai_yunusbayev_16 + 31.8% samoan_xing_13" "14.187"
[7,] "68.2% tajik_yunusbayev_15 + 31.8% tongan_xing_13" "15.1353"
[8,] "68.6% hazara_hgdp_22 + 31.4% tongan_xing_13" "15.6222"
[9,] "30.6% tongan_xing_13 + 69.4% uyghur_hgdp_10" "15.7134"
[10,] "32.8% tongan_xing_13 + 67.2% turk-aydin_hodoglugil_20" "15.7475"
[11,] "31.3% samoan_xing_13 + 68.7% tajik_yunusbayev_15" "15.8493"
[12,] "69.1% hazara_hgdp_22 + 30.9% samoan_xing_13" "16.2806"
[13,] "30.1% samoan_xing_13 + 69.9% uyghur_hgdp_10" "16.4021"
[14,] "32.3% samoan_xing_13 + 67.7% turk-aydin_hodoglugil_20" "16.5181"
[15,] "21.1% melanesian_hgdp_10 + 78.9% turkmen_yunusbayev_11" "17.4238"
[16,] "20.5% melanesian_hgdp_10 + 79.5% uzbek_behar_15" "17.4987"
[17,] "27.3% singapore-malay_sgvp_89 + 72.7% turkmen_yunusbayev_11" "17.571"
[18,] "29.7% khmer-cambodian_xing_3 + 70.3% turkmen_yunusbayev_11" "17.669"
[19,] "65% iranian_harappa_9 + 35% tongan_xing_13" "17.67"
[20,] "64.2% azeri_harappa_2 + 35.8% tongan_xing_13" "17.7618"
[21,] "26.4% dai_hgdp_10 + 73.6% turkmen_yunusbayev_11" "17.7758"
[22,] "26.7% dai-chinese_1000genomes_91 + 73.3% turkmen_yunusbayev_11" "17.8136"
[23,] "35.2% tongan_xing_13 + 64.8% turk-istanbul_hodoglugil_20" "17.8183"
[24,] "29.9% lahu_hgdp_8 + 70.1% turkmen_yunusbayev_11" "17.8439"
[25,] "34.7% tongan_xing_13 + 65.3% turkish_harappa_4" "17.8445"

FWIW, Tongan/Samoan are ~70% SE-Asian, ~30% Papuan. So I'm guessing that's something similar to Onge on a PCA.

Also interesting, the lots of East Eurasian is making it look like an East Asian-admixed Central Asian population. Karasuk did well in such a thing (admixture and IBS), but qpAdm doesn't seem to find an affinity. Or it could've been 50% of a population that's a cross between something from Oceania and further north.

can you do this for paniya and piramalai kallar by that i mean run a test with removed south indian component. Also the results dont add up to 100%

Dr_McNinja
01-28-2016, 06:30 PM
Using Baloch/Gedrosian, CHG, and Kurd/W-Iranian seems to work better.

HRP0393 Haryana Jatt:

26.44% Baloch
19.28% NE-Euro
5.38% SE-Asian
3.38% Siberian
3.25% NE-Asian
3.86% Papuan
2.14% American
0.33% Beringian
1.72% Mediterranean
0.36% SW-Asian
0.69% San
0.77% E-African
0.07% Pygmy
0.72% W-African
19.52% CHG
12.09% W-Iranian

I believe that 12.09% W-Iranian is the entire West Eurasian part of their former 30% S-Indian.

But for most other South Asians, that W-Iranian/Kurd component just sucks up a lot of their admixture (because of the affinity for SW-Asian/Mediterranean admixture):

Me:

25.39% Baloch
10.41% NE-Euro
6.42% SE-Asian
3.74% Siberian
4.03% NE-Asian
5.53% Papuan
1.80% American
2.34% Beringian
0.00% Mediterranean
0.08% SW-Asian
0.00% San
1.30% E-African
0.68% Pygmy
0.07% W-African
12.89% CHG
25.31% W-Iranian

Sapporo:

31.39% Baloch
12.74% NE-Euro
4.97% SE-Asian
3.73% Siberian
2.52% NE-Asian
3.42% Papuan
1.68% American
1.91% Beringian
0.00% Mediterranean
0.58% SW-Asian
0.00% San
1.71% E-African
0.00% Pygmy
0.00% W-African
12.94% CHG
22.40% W-Iranian

But with CHG, it works in non-South Asians:

Zeph/Assyrian:

0.00% Baloch
0.00% NE-Euro
0.00% SE-Asian
0.00% Siberian
0.00% NE-Asian
0.00% Papuan
0.00% American
0.00% Beringian
7.32% Mediterranean
5.92% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
1.49% CHG
85.28% W-Iranian

South Italian:

0.00% Baloch
16.67% NE-Euro
0.00% SE-Asian
0.00% Siberian
0.00% NE-Asian
0.00% Papuan
0.00% American
0.00% Beringian
26.59% Mediterranean
7.46% SW-Asian
0.02% San
0.42% E-African
0.00% Pygmy
0.02% W-African
1.89% CHG
46.93% W-Iranian

Sweden:

4.04% Baloch
57.87% NE-Euro
0.45% SE-Asian
0.70% Siberian
0.55% NE-Asian
0.25% Papuan
0.18% American
0.45% Beringian
22.80% Mediterranean
1.53% SW-Asian
0.04% San
0.02% E-African
0.00% Pygmy
0.00% W-African
5.77% CHG
5.35% W-Iranian

Caucasus/Karachay:

0.00% Baloch
15.62% NE-Euro
0.01% SE-Asian
2.98% Siberian
4.00% NE-Asian
0.00% Papuan
0.23% American
0.44% Beringian
7.31% Mediterranean
0.00% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
29.91% CHG
39.50% W-Iranian

DMXX (Iranian):

0.00% Baloch
3.19% NE-Euro
0.11% SE-Asian
0.97% Siberian
2.65% NE-Asian
0.94% Papuan
0.00% American
0.69% Beringian
3.19% Mediterranean
1.82% SW-Asian
0.00% San
0.00% E-African
0.00% Pygmy
0.00% W-African
12.35% CHG
74.08% W-Iranian

Dr_McNinja
01-28-2016, 06:35 PM
can you do this for paniya and piramalai kallar by that i mean run a test with removed south indian component. Also the results dont add up to 100%

No, I don't have my Linux install anymore with all the genome data and tools to mess around with that. I'm just running these on my PC.

VOX
01-28-2016, 07:34 PM
What works is if you put the Linux distro in virtual environment like VMWare Player or Virtual Box. Using a light desk top like LXDE/XFCE/LXQT would help speed things up as well in the virtual machine which tends to be slightly slower.

Dr_McNinja
01-28-2016, 08:10 PM
Also worked for Bored:

25.20% Baloch
11.74% NE-Euro
6.28% SE-Asian
2.18% Siberian
6.91% NE-Asian
3.73% Papuan
2.15% American
2.47% Beringian
0.24% Mediterranean
0.43% SW-Asian
0.30% San
0.59% E-African
0.33% Pygmy
0.62% W-African
22.78% CHG
14.06% W-Iranian

That 14.06% W-Iranian/Kurd is the ~40% West Eurasian part of his ~35% S-Indian.

Dr_McNinja
01-31-2016, 06:46 PM
Looking for ASI, but I think I got a decent ANI component:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=805280337

ANI and CHG are very close, but split accordingly. ANI is like a cross between HAP_Baloch and Kurds while CHG is a cross between HAP_Baloch and HAP_Caucasian. I tried various versions, including some E-African in ANI, but it's very hard to fully separate the two. ANI gets boosted by CHG in a bunch of places.

Except for me. I had ~35% HAP Baloch and ~37% S-Indian which is approx ~47% West Eurasian from those two components. The little bits of SW-Asian (even less than 1%) should be significant and indicative of Baloch or other ancestry west of the Indus from several generations back (since ANI has a strong SW-Asian affinity and should have pulled it all in).

Some of the extra CHG is undoubtedly Steppe-related, since not all the NE-Euro/Siberian/Beringian/Amerindian went into the two Steppe components. But I can't lump it all into Steppe since some of it is also undoubtedly just from west of the Indus in recent history and not ancestrally Steppe. You can sort of get an idea though. All of the Haryana Jatt's CHG is probably Steppe-related. The Afghans should have a lot of Steppe ancestry in their CHG too, since the better Steppe fit for them in Sintashta which has more of it than the two used here (Andronovo for West and Karasuk for East).

I keep getting some Afghan affinity combined with extra South Indian affinity (like far south... it's this extreme pull towards ASI) and less of whatever's in the middle, so it's that phenomena I've always mentioned, getting pulled towards two opposite extremes for some reason.

This one also shows it:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1084480849

Might be something to do with mixed ancestry within South Asia. Combining that Afghan with S-India and putting it in N-India_Br would establish a continuum between other North Indians (Haryana, Pahari) and Punjabis, where I'd expect us to be. Another possible explanation is not enough ANE. Low ANE but high ASE = South India whereas most people from the Pakistan area have higher ANE and higher ASE which pulls towards Gujarat area (picked up as Baloch or North India). Our ASE in the K7/K8 is normal, but our ANE is under the average by just a bit (~1-1.5%).

DMXX
01-31-2016, 07:57 PM
What are you modelling your steppe components on, Doc?

We know from various calculators (confirmed by Mathieson et al.) that Sintashta and Andronovo* can both be modeled as a mixture of Yamnaya and EEF:

http://4.bp.blogspot.com/-6IxdvII-LQs/ViGboGhpIDI/AAAAAAAADiw/E47emLBllX8/s1600/Neolithic.png

That would make the Sintashta average something resembling 25% EHG, 25% CHG, 20% WHG and 20% EEF.

* Based on location and elevated Yamnaya_Samara, looks like there was some low-scale Afanasievo admixture among them

Coldmountains
01-31-2016, 08:38 PM
Looking for ASI, but I think I got a decent ANI component:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=805280337

ANI and CHG are very close, but split accordingly. ANI is like a cross between HAP_Baloch and Kurds while CHG is a cross between HAP_Baloch and HAP_Caucasian. I tried various versions, including some E-African in ANI, but it's very hard to fully separate the two. ANI gets boosted by CHG in a bunch of places.

Except for me. I had ~35% HAP Baloch and ~37% S-Indian which is approx ~47% West Eurasian from those two components. The little bits of SW-Asian (even less than 1%) should be significant and indicative of Baloch or other ancestry west of the Indus from several generations back (since ANI has a strong SW-Asian affinity and should have pulled it all in).

Some of the extra CHG is undoubtedly Steppe-related, since not all the NE-Euro/Siberian/Beringian/Amerindian went into the two Steppe components. But I can't lump it all into Steppe since some of it is also undoubtedly just from west of the Indus in recent history and not ancestrally Steppe. You can sort of get an idea though. All of the Haryana Jatt's CHG is probably Steppe-related. The Afghans should have a lot of Steppe ancestry in their CHG too, since the better Steppe fit for them in Sintashta which has more of it than the two used here (Andronovo for West and Karasuk for East).

I keep getting some Afghan affinity combined with extra South Indian affinity (like far south... it's this extreme pull towards ASI) and less of whatever's in the middle, so it's that phenomena I've always mentioned, getting pulled towards two opposite extremes for some reason.

This one also shows it:

https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1084480849

Might be something to do with mixed ancestry within South Asia. Combining that Afghan with S-India and putting it in N-India_Br would establish a continuum between other North Indians (Haryana, Pahari) and Punjabis, where I'd expect us to be. Another possible explanation is not enough ANE. Low ANE but high ASE = South India whereas most people from the Pakistan area have higher ANE and higher ASE which pulls towards Gujarat area (picked up as Baloch or North India). Our ASE in the K7/K8 is normal, but our ANE is under the average by just a bit (~1-1.5%).

Not more than 1/3 of CHG among Pashtuns, Jatts and Tajiks can be directly from the steppe in my opinion. Andronovo and Sintashta were lower in it than any modern Central Asians and CHG is highest among Baluchs and Burusho who have lowest steppe affinity. But it is possible that mixed BMAC/steppe Indo-Iranian tribes were very high in CHG and that they contributed very much to modern Indo-Iranian ancestry. CHG looks very old in this region and rather decreased after Indo-Iranians arrived there . We have no ancient DNA from this region and I would not be surprised if EHG or Siberian-admixture was already present in small amounts there so much of what looks "steppe" can be older.

Dr_McNinja
01-31-2016, 10:37 PM
What are you modelling your steppe components on, Doc?

We know from various calculators (confirmed by Mathieson et al.) that Sintashta and Andronovo* can both be modeled as a mixture of Yamnaya and EEF:

http://4.bp.blogspot.com/-6IxdvII-LQs/ViGboGhpIDI/AAAAAAAADiw/E47emLBllX8/s1600/Neolithic.png

That would make the Sintashta average something resembling 25% EHG, 25% CHG, 20% WHG and 20% EEF.

* Based on location and elevated Yamnaya_Samara, looks like there was some low-scale Afanasievo admixture among them

Andronovo's basically Northern European with some extra Baloch/Gedrosian. Karasuk is that with a bunch of Siberian.

This older run: https://docs.google.com/spreadsheets/d/1XGY-UIBC2GcUWGl1FizNvu9ofle4ZPHvfNmbrVYxGMo/edit#gid=1040594487 Shows the two Steppe components without being limited by the ANI component.

The Eastern Steppe layer is in all of South Asia, it's also in the HAP S-Indian component. Considering R1a penetrates really far into the subcontinent, and there's some Andronovo affinity being picked up by Chad in his experiments, I'm not sure what this means. Could be the original R1a-L657 migration with a more pronounced "westernly" Andronovo-like settlement in North India. There's definitely a bunch of extra ANE-East Eurasian in India though.

Interestingly, the affinity to the Karasuk-like component is there only in South Asians, but not Tajiks or other Central Asian groups. These are simulated components assembled, frankenstein-like, from the modern components. So I have no idea why the pull isn't universal. If anything, modern day Central Asians should have more affinity than South Asians. Could be the Gedrosian/Baloch content, but Karasuk doesn't have that much of it anyway (10-15% I believe), so I don't know.

Dr_McNinja
06-26-2016, 05:22 AM
Tried making a "Dravidian" component out of that West Eurasian part of HAP S-Indian, this is what it showed up in:

(Baloch/Caucasian components were still there from original calculator, so they went up a little too)

Me: 10.69%
HRP0393 Haryana: 9.38%
Sapporo: 11.12%
South Italian: 10.79%
Assyrian (Zeph): 4.28%
Sweden: 3.34%
Caucasus (karachay): 3.49%
Iranian (DMXX): 11.04%
longbowman: 10.98%
parasar: 9.46%
abbas (Iranian of some kind?): 13.07%
ashwin: 9.99%
Turk (dogukan): 3.73%
Turk (Hanna): 7.08%
Anabasis: 2.98%
DLuffy: 10.71%
Chad: 3.99%
Iranian (NK19191): 12.10%
Bored: 9.73%
Pashtun/Tajik (WN): 11.48%
Sein's friend: 10.36%
N-African Berber: 6.37%
Khetran Baloch: 12.38%
Everest: 8.91%
Rukha: 12.00%
Sein: 12.10%
HRP0370 Pashtun: 12.89%
Soulblighter (South Indian Brahmin): 9.88%
Reza (Bengali): 7.34%

Doesn't work if I put in a simulated CHG component instead of Baloch/Caucasian, everything gets taken in by CHG.
I put Eurogenes K7's ANE and ASE components in with HarappaWorld and removed S-Indian. SNPs down to 70k:

https://docs.google.com/spreadsheets/d/15ffQQ5Ggv24AduJF8cA-tClMiLJsgjNRU4gmfbGYc6I/edit?usp=sharing

Note, the South American/European individual (Brazilian or Hispanic is my guess) has 35% Amerindian and 2.7% ANE. The Han Chinese person did not get any ASE.

The South Asians look like Iranians with a bunch of ANE/ASE/Papuan instead of Caucasian. Steppe (NE-Euro/Med) is similar except among some (Haryana Jatt most noticeably), unless we consider Caucasian in South Asians to be from the Steppe.

Dr_McNinja
06-27-2016, 02:06 AM
I didn't finish putting in results beyond me, HRP0393 and Sapporo, but I'll do that later tonight:

https://docs.google.com/spreadsheets/d/15ffQQ5Ggv24AduJF8cA-tClMiLJsgjNRU4gmfbGYc6I/edit#gid=1979418746

ANI is West Eurasian part of S-Indian, ASI is ENA part. It generally adds up well, but as you can see some of the West Eurasian still slips into Caucasian, SW-Asian, etc. There is also some ANE in ASI still, which was tied up with East Eurasian and I didn't remove it in this run.

In the next one where I filled everyone out:

https://docs.google.com/spreadsheets/d/15ffQQ5Ggv24AduJF8cA-tClMiLJsgjNRU4gmfbGYc6I/edit#gid=1740070869

Steppe_Sintashta is consistent throughout since it's based off ANE (assembled via Eurogenes K7 components) and MA1/ANE is extremely consistent in South Asia. I figure that should correlate to Steppe since most Indo-European cultures and lineages dominate the subcontinent, even amongst Dravidian-speaking peoples. It also pulls away whatever Steppe was in S-Indian before.

I used a pseudo-Iran_Chl component along with the original Baloch/Gedrosia component because Iran_N would just take all the admixture away from the other two. Baloch/Gedrosia does a well enough job attracting what Iran_N would have. It's not all smooth sailing since Iran_Chl steals some admixture from Caucasus/Baloch in Afghans and Iranians (wow @ NK19191), but DMXX and abbas look good so I highlighted those two. Ideally, Iran_Chl would have tapered off in Afghans as Caucasian went up. Iran_Chl takes over for the non-Steppe part of ANI (so, Neolithic India) in South Asia. In fact, adding it to ASI gives almost exactly HAP S-Indian for a few people.

I tweaked ASI to remove what little ANE was still in there to free it up for other components. It's 53% East Eurasian, 19% ASE, 25% Papuan (this is based off HAP S-Indian's ASI... there are other ways to get a decent component, but I was more focused on ANI here).

Edit: Like in the first one with ANE/ASE added to Harappa, South Asians look like Iranians (using DMXX and abbas) shifted towards Steppe instead of Caucasian and ASI instead of SW-Asian.

Dr_McNinja
06-27-2016, 05:22 AM
https://docs.google.com/spreadsheets/d/15ffQQ5Ggv24AduJF8cA-tClMiLJsgjNRU4gmfbGYc6I/edit#gid=1979418746

I filled it out, follows a similar pattern and shows the ANI correlates most closely with Caucasian (which is still here in this one).

Hanna
06-27-2016, 08:04 AM
https://docs.google.com/spreadsheets/d/15ffQQ5Ggv24AduJF8cA-tClMiLJsgjNRU4gmfbGYc6I/edit#gid=1740070869


This spreadsheet shows Anabasis scoring 49% Northeast European, I think that's a mistake?

Dr_McNinja
06-27-2016, 08:55 AM
Thanks, fixed that

Dr_McNinja
06-30-2016, 01:09 AM
A probable Haryanvi/Punjabi Sikh (presumably Jatt, their name was 'Singh') turned up in mine and my father's cousin matches on Gedmatch (with 8.1 cM to dad, 7.1 cM to me... Edit: And 5.5 cM to my mother, so he's probably related to us through the Pansota side which both my mom/dad share... Edit #2: He's also got 5.5 cM with a Grewal, a relative of HRP0418 and 5.8 cM to HRP0393 Beniwal Haryana Jatt and 5.3 cM to another Gill possibly related to paulgill).

Highest European I've seen in one yet:

Population
S-Indian 24.87
Baloch 35.54
Caucasian 10.76
NE-Euro 20.38
SE-Asian -
Siberian 0.72
NE-Asian -
Papuan 1.09
American 1.79
Beringian -
Mediterranean 4.21
SW-Asian 0.59
San -
E-African -
Pygmy -
W-African -

Single Population Sharing:

# Population (source) Distance
1 haryana-jatt (harappa) 3.86
2 punjabi-jatt-sikh (harappa) 9.24
3 punjabi-khatri (harappa) 11.13
4 pathan (hgdp) 11.59
5 kashmiri (harappa) 12.75
6 sindhi (harappa) 12.82
7 nepalese-a (xing) 12.96
8 bhatia (harappa) 13.14
9 punjabi-brahmin (harappa) 13.31
10 punjabi-jatt-muslim (harappa) 13.36
11 kashmiri-pandit (reich) 13.46
12 burusho (hgdp) 13.59
13 singapore-indian-c (sgvp) 13.8
14 pashtun (harappa) 14.01
15 punjabi (harappa) 14.05
16 kalash (hgdp) 14.31
17 up-muslim (harappa) 14.72
18 punjabi-ramgarhia (harappa) 15.05
19 kashmiri-pahari (harappa) 15.36
20 punjabi-arain (xing) 15.55

Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 93.6% haryana-jatt (harappa) + 6.4% ukranian (yunusbayev) @ 1.72
2 93.2% haryana-jatt (harappa) + 6.8% slovenian (xing) @ 1.73
3 93.2% haryana-jatt (harappa) + 6.8% hungarian (behar) @ 1.77
4 94.1% haryana-jatt (harappa) + 5.9% russian (behar) @ 1.8
5 94.1% haryana-jatt (harappa) + 5.9% belorussian (behar) @ 1.81
6 77.8% punjabi-arain (xing) + 22.2% russian (behar) @ 1.84
7 92.9% haryana-jatt (harappa) + 7.1% romanian-a (behar) @ 1.93
8 93.1% haryana-jatt (harappa) + 6.9% bulgarian (yunusbayev) @ 1.96
9 93.8% haryana-jatt (harappa) + 6.2% mordovian (yunusbayev) @ 1.96
10 94.7% haryana-jatt (harappa) + 5.3% lithuanian (behar) @ 2.03
11 93.9% haryana-jatt (harappa) + 6.1% n-european (xing) @ 2.11
12 94% haryana-jatt (harappa) + 6% utahn-white (hapmap) @ 2.16
13 77.9% punjabi-arain (xing) + 22.1% belorussian (behar) @ 2.16
14 94.1% haryana-jatt (harappa) + 5.9% utahn-white (1000genomes) @ 2.21
15 85.8% punjabi-jatt-sikh (harappa) + 14.2% belorussian (behar) @ 2.24
16 94.2% haryana-jatt (harappa) + 5.8% british (1000genomes) @ 2.28
17 95.6% haryana-jatt (harappa) + 4.4% finnish (1000genomes) @ 2.31
18 94.1% haryana-jatt (harappa) + 5.9% french (hgdp) @ 2.32
19 94.3% haryana-jatt (harappa) + 5.7% orcadian (hgdp) @ 2.32
20 85.8% punjabi-jatt-sikh (harappa) + 14.2% russian (behar) @ 2.35

Population
ANE 29.62
ASE 12.98
WHG-UHG 17.29
East_Eurasian 2.25
West_African 0.98
East_African 1.54
ENF 35.34

I emailed him asking for which part of Punjab or Haryana he's from...

Dr_McNinja
06-30-2016, 01:35 AM
Andronovo:

S-Indian 0.54
Baloch 21.23
Caucasian 2.40
NE-Euro 56.39
SE-Asian -
Siberian 1.93
NE-Asian -
Papuan -
American 1.05
Beringian 1.22
Mediterranean 14.37
SW-Asian -
San -
E-African -
Pygmy 0.06
W-African 0.81

Sintashta: 30.99 ANE, 8.25 ASE, 43.05 WHG, 0.05 E-Eu, 1.84 W-African, 15.84 ENF (keeping in mind WHG to ENF is overestimated in ANE K7)

3000 years later and it's like this guy is only one generation removed from the Steppe. That's insane.

But he's got basically 23-24% Iranian-like Caucasian admixture (low ANE) in addition to 35% Baloch (high ANE, like Neolithic Iranian).

And Afanasievo genomes were getting 11% ASE. It's like they mixed Steppe with just Iranian. That would be enough to give his results.

Edit: In terms of the modern populations, it'd be like mixing the Steppe HarappWorld profile (Baloch+NE-Euro) with something that has S-Indian+Baloch. Because S-Indian is inversely correlating to NE-Euro more than any other component. But because the Baloch stays constant it strongly suggests NE-Euro came from a population high in Baloch and NE-Euro which obviously has only been Steppe populations put through the calculator.