PDA

View Full Version : Looking for ASI



Pages : [1] 2 3

Chad Rohlfsen
05-02-2015, 08:13 PM
Okay, the K10, with about 2400 pops, found an Onge component, but it's less than half of what Zach's was. I am not sure if the SC Asian component is masking some, or not. I'll leave it up to you guys, as to what you want to see. Would you prefer a supervised run, with just the tribals? I can use Onge, Papuan, Atayal, EEF, Bedouin, ANE, WHG, as the components. What do you folks think?

The Barnacle
05-02-2015, 08:30 PM
Go for it my man Chad!

Chad Rohlfsen
05-02-2015, 08:45 PM
Now, here's the kicker. I know that Kostenki shows South Asian, and we see this WHG/UHG stuff in South Asia. Do I dare include him? Should I use a countermeasure, like Yamnaya, to make sure that he isn't soaking up ANI or Yamnaya? Or, I can use some other SC Asian pop as a reference (Kalash), although I really cringe at doing that. Perhaps, WHG will cover it just fine. Input is greatly appreciated!

jesus
05-02-2015, 08:47 PM
Try whatever comes to your mind and post the results that make the most sense. Having Onge and Papuan in the same tool is gonna show very interesting results. Try including Samaritans instead of Bedouins.

Dr_McNinja
05-02-2015, 10:26 PM
Try everything you can think of, time allowing. :) How about including distinct Southeast Asian and Northeast Asian populations if possible? There's probably a distinct modern Southeast Asian signal in India due to later migrations, especially among some tribal groups. In any case, if we have all three, then Southeast Asian in South Central Asia (Afghanistan, Pakistan, North/west India) would probably indicate something Siberian. I don't know why those signals sometimes overlap. But it's better than it going into Onge somehow (assuming it doesn't go into Onge in the first place).

Also, try to run the entire Gujarati (A, B, etc) set if you can, there's a lot of diversity in there.

I think Kalash will give you a lot of trouble, but it's worth a shot seeing how they come out and if they don't work and make their own component that no one else is a member of, discard them until another time when you have the time to mess with the dataset to get them to behave.

Anabasis
05-02-2015, 10:32 PM
While most of the genetician and researcher interested in South Asian or Steppe people why not there is an study on ancient Anatolians? We know that agriculture spread over the Europe and South Asia via Anatolia-Mesopotamia. (Am i wrong?) Archaeologist found the oldest tamples which is dated 10-11K ybp in Göbeklitepe. Many thing has begun in Fertile crescent. But in calculators most of the components divides near eastern into several pieces like East med/ West Asian or ENF/Caucasian/Gedrosia. I realy couldnt understand why Caucasia is so different then ENF? What we see is Caucasian component is dominant in ancient ruins where first agricaltural zones take place. Arent they which they brought agricalture to Europe? Then how can ENF can be different then Caucasia? And how can Cauciasia became diferent then Gedrosa in opposite direction.

Chad Rohlfsen
05-02-2015, 10:37 PM
Try everything you can think of, time allowing. :) How about including distinct Southeast Asian and Northeast Asian populations if possible? There's probably a distinct modern Southeast Asian signal in India due to later migrations, especially among some tribal groups. In any case, if we have all three, then Southeast Asian in South Central Asia (Afghanistan, Pakistan, North/west India) would probably indicate something Siberian. I don't know why those signals sometimes overlap. But it's better than it going into Onge somehow (assuming it doesn't go into Onge in the first place).

Also, try to run the entire Gujarati (A, B, etc) set if you can, there's a lot of diversity in there.

I think Kalash will give you a lot of trouble, but it's worth a shot seeing how they come out and if they don't work and make their own component that no one else is a member of, discard them until another time when you have the time to mess with the dataset to get them to behave.

It seems possible to separate the Onge from other groups. Using SE Asians may not be great, as they probably have Onge ancestry, as well. I can get a little creative after a few unsupervised runs. I can always add pops, bit by bit, to look for something. I think I'll start with WHG, ANE, EEF, Bedouin, Yoruba, Atayal, and maybe EHG, with tribals, all unsupervised for a few k's. Although, substituting ANE with Karitiana would show us any potential "ANE" flow, incase MA1 won't make a cluster before it get's too noisy. I'll get started on this as soon as my K11 is finished, which should be in about 1-2 hours.

I think I'll include the following in unsupervised, to start.

Maybe, a K 6 or 7.
WHG's
EEF's
Basque
BedouinA and B
Yoruba
Atayal
Ami
Han
MA1/EHG/ or Karitiana
Papuan
Onge
maybe Yamnaya
and all tribal pops I have.

Another interesting thing to test later is Papuans as a mix of Onge, Austronesian, and Denisovan.


Does that sound alright?

Chad Rohlfsen
05-02-2015, 10:38 PM
While most of the genetician and researcher interested in South Asian or Steppe people why not there is an study on ancient Anatolians? We know that agriculture spread over the Europe and South Asia via Anatolia-Mesopotamia. (Am i wrong?) Archaeologist found the oldest tamples which is dated 10-11K ybp in Göbeklitepe. Many thing has begun in Fertile crescent. But in calculators most of the components divides near eastern into several pieces like East med/ West Asian or ENF/Caucasian/Gedrosia. I realy couldnt understand why Caucasia is so different then ENF? What we see is Caucasian component is dominant in ancient ruins where first agricaltural zones take place. Arent they which they brought agricalture to Europe? Then how can ENF can be different then Caucasia? And how can Cauciasia became diferent then Gedrosa in opposite direction.

I've already got a good idea what they're like. I'm not concerned there. Caucasus, Gedrosia, etc, all have EHG/ANE ancestry. They won't be like the first farmers.

Dr_McNinja
05-02-2015, 10:47 PM
It seems possible to separate the Onge from other groups. Using SE Asians may not be great, as they probably have Onge ancestry, as well. I can get a little creative after a few unsupervised runs. I can always add pops, bit by bit, to look for something. I think I'll start with WHG, ANE, EEF, Bedouin, Yoruba, Atayal, and maybe EHG, with tribals, all unsupervised for a few k's. Although, substituting ANE with Karitiana would show us any potential "ANE" flow, incase MA1 won't make a cluster before it get's too noisy. I'll get started on this as soon as my K11 is finished, which should be in about 1-2 hours.

I think I'll include the following in unsupervised, to start.

Maybe, a K 6 or 7.
WHG's
EEF's
Basque
BedouinA and B
Yoruba
Atayal
Ami
Han
MA1/EHG/ or Karitiana
Papuan
Onge
and all tribal pops I have.

Another interesting thing to test later is Papuans as a mix of Onge, Austronesian, and Denisovan.


Does that sound alright?

Sounds good. Very curious to see what you find out about the relationship between Onge, Denisovan, Papuans, etc.

Dr_McNinja
05-02-2015, 10:51 PM
While most of the genetician and researcher interested in South Asian or Steppe people why not there is an study on ancient Anatolians? We know that agriculture spread over the Europe and South Asia via Anatolia-Mesopotamia. (Am i wrong?) Archaeologist found the oldest tamples which is dated 10-11K ybp in Göbeklitepe. Many thing has begun in Fertile crescent. But in calculators most of the components divides near eastern into several pieces like East med/ West Asian or ENF/Caucasian/Gedrosia. I realy couldnt understand why Caucasia is so different then ENF? What we see is Caucasian component is dominant in ancient ruins where first agricaltural zones take place. Arent they which they brought agricalture to Europe? Then how can ENF can be different then Caucasia? And how can Cauciasia became diferent then Gedrosa in opposite direction.In addition to what Chad said, I think Anatolia is far more important for its haplogroup lineages. Admixture wise it's close to Near East, Caucasus, Europe, etc but a lot of haplogroups are related to Caucasus and some might have originated near there. I'm J2b2-M241 for example, from the South Asian branch, but my closest in STRs are a few J2b1 people from Turkey. If J2b1 is there, and J2b is also highly frequent near there, perhaps J2b2-M241 had its genesis near there. Same goes for some other farming-associated haplogroups. Another Punjabi here (paulgill, same clan as me) is J1 and his closest in STRs is also from Turkey.

jesus
05-02-2015, 10:54 PM
It seems possible to separate the Onge from other groups. Using SE Asians may not be great, as they probably have Onge ancestry, as well. I can get a little creative after a few unsupervised runs. I can always add pops, bit by bit, to look for something. I think I'll start with WHG, ANE, EEF, Bedouin, Yoruba, Atayal, and maybe EHG, with tribals, all unsupervised for a few k's. Although, substituting ANE with Karitiana would show us any potential "ANE" flow, incase MA1 won't make a cluster before it get's too noisy. I'll get started on this as soon as my K11 is finished, which should be in about 1-2 hours.

I think I'll include the following in unsupervised, to start.

Maybe, a K 6 or 7.
WHG's
EEF's
Basque
BedouinA and B
Yoruba
Atayal
Ami
Han
MA1/EHG/ or Karitiana
Papuan
Onge
maybe Yamnaya
and all tribal pops I have.

Another interesting thing to test later is Papuans as a mix of Onge, Austronesian, and Denisovan.


Does that sound alright?
BedouinA and B have significant SSA, I think BedouinA are over 8% SSA on Average. BedouinB have significantly higher SSA, in this case using BedouinA by itself would be better.

Dr_McNinja
05-02-2015, 10:59 PM
Another thing to experiment with later if you get time is Ust-Ishim and Onge (and other South Asians).

Chad Rohlfsen
05-02-2015, 11:26 PM
BedouinB are about 5% SSA, on average, BedouinA hit double digits. In order to separate EEF from Bedouin, I'll need a few pops, like Sardinians and BedouinA. It's just to make sure components come out. It may also bring out one single cluster that peaks in EEF. We shall see. I'll get to that, this evening. Starting at K4 or 5.

parasar
05-02-2015, 11:46 PM
Now, here's the kicker. I know that Kostenki shows South Asian, and we see this WHG/UHG stuff in South Asia. Do I dare include him? Should I use a countermeasure, like Yamnaya, to make sure that he isn't soaking up ANI or Yamnaya? Or, I can use some other SC Asian pop as a reference (Kalash), although I really cringe at doing that. Perhaps, WHG will cover it just fine. Input is greatly appreciated!

Do we have enough coverage on K14? I would avoid using something with low coverage. Plus K14 to me is looks admixed - Ust-Ishim type + Basal.

Chad Rohlfsen
05-03-2015, 12:00 AM
Do we have enough coverage on K14? I would avoid using something with low coverage. Plus K14 to me is looks admixed - Ust-Ishim type + Basal.

I use fewer SNPs, abt 100k, to keep it even with the South Asian samples, which have lower coverage. K14 looks almost like a basal West Eurasian. He does show South Asian, so, it's hard to say. I can just avoid him, for now. I'll just work with the smaller list, to start.

Everyone can thank David, for getting me started on this adventure. I'll be setting up the first run in a few minutes.

Sein
05-03-2015, 04:01 AM
Chad,

Although you've already started some analyses (I was late to the party), I was hoping you could eventually try something along these lines, when you find the time. The supervised ADMIXTURE run would have these components:

Onge

Nganasan (they have ANE, but they are predominately ENA. If you use Han or She, South Asian ASI will be a combination of both Han and Onge, which will be confusing for northern South Asians, who often have some actual ENA of Siberian origin. So the Nganasan are best)

BedouinB

Yamnaya

MA1 (ANE)

EHG

WHG

Mbuti

The populations to be tested would be:

All Indian tribal populations in your data-set

UP Brahmins (if you have them)

UP Chamar (if you have them)

Sindhi

Brahui

Baloch

"Pathans" (Pashtuns from HGDP data-set)

"Pashtuns" (Pashtuns from Di Cristofaro data-set, you might have these. If you don't, that's okay)

Pomiri Tajiks

Yaghnobi Tajiks

Iranians

In theory, this should be a good supervised ADMIXTURE run.

I would greatly appreciate this. Thanks in advance.

Edit: Added Yamnaya as an extra supervised component.

Chad Rohlfsen
05-03-2015, 04:28 AM
Alright, I'll post some stuff eventually here. So far, at K6, an Onge component pops up, but there is still a South Asian cluster. At the moment, the Paniya are at about 12% Onge. I'll keep working up in K value to break up that South Asian cluster. I may have to add in a few more pops if nothing more shows up by k12.

Dr_McNinja
05-03-2015, 04:34 AM
Alright, I'll post some stuff eventually here. So far, at K6, an Onge component pops up, but there is still a South Asian cluster. At the moment, the Paniya are at about 12% Onge. I'll keep working up in K value to break up that South Asian cluster. I may have to add in a few more pops if nothing more shows up by k12.

How does that relate to K<6 runs? Did a South Asian component in K<6 break into South Asian+Onge in K=6?

Chad Rohlfsen
05-03-2015, 05:15 AM
Yeah, the South Asian peaked in Mala, at about 81%, with Paniya peaking in Onge. K7, the SA cluster got larger and Onge dropped quite a bit. There is certainly a good amount of Onge there. I will keep going up, until South Asian is totally broken up. I may have to throw in a couple other pops, but I think I have good coverage up until 12 or 13. I'll see how far I can push it. I may have to complete this tomorrow. I'm going to crash pretty soon. K8 is running.

Kurd
05-03-2015, 05:19 AM
Now, here's the kicker. I know that Kostenki shows South Asian, and we see this WHG/UHG stuff in South Asia. Do I dare include him? Should I use a countermeasure, like Yamnaya, to make sure that he isn't soaking up ANI or Yamnaya? Or, I can use some other SC Asian pop as a reference (Kalash), although I really cringe at doing that. Perhaps, WHG will cover it just fine. Input is greatly appreciated!

Actually, Kalash would not be a bad choice, since I am sure your goal is to identify populations that are least admixed (closest to ancestral) and they certainly are in my belief the least admixed in that region due to their isolation for geographic (mountains) and cultural/ religious (they are surrounded by muslim groups).

With regards to other choices, some of my PCA under various threads may help you identify populations that are least overlaping to reduce redundancy

Chad Rohlfsen
05-03-2015, 01:45 PM
Okay, the South Asian is starting to break up a bit, as the Malay head over to a component with Cambodians, at K9. I'll keep climbing.

I already have Bedouins, EEF, Yamnaya, and Gujarati. Should I put in a few Caucasus pops, or do you guys think it will confuse it?

Kurd
05-03-2015, 02:05 PM
Okay, the South Asian is starting to break up a bit, as the Malay head over to a component with Cambodians, at K9. I'll keep climbing.

I already have Bedouins, EEF, Yamnaya, and Gujarati. Should I put in a few Caucasus pops, or do you guys think it will confuse it?

Chad,

I believe you may be ok with Armenian with regards to overlap and redundancy with EEF. I would like also to see a SC Asian pop, such as one mentioned by Sein. Kalash is also a great choice, because they are closest to ancestral from the pops in that area. At higher k they have their own distinct indo european component which is not really shared significantly by any of their neighbors (see my MDLP world 22 thread). I assume you will have an E and SE pop or two

Chad Rohlfsen
05-03-2015, 02:11 PM
Chad,

I believe you may be ok with Armenian with regards to overlap and redundancy with EEF. I would like also to see a SC Asian pop, such as one mentioned by Sein. Kalash is also a great choice, because they are closest to ancestral from the pops in that area. At higher k they have their own distinct indo european component which is not really shared significantly by any of their neighbors (see my MDLP world 22 thread). I assume you will have an E and SE pop or two

I have all tribals(including some like Vishwa Brahmin), plus Cambodian, Papuan, Yoruba, BedouinA and B, Yamnaya, EEF, WHG, EHG, MA1, Sardinian, Basque, Ami, Atayal, Han, Yakut, Gujaratis, and Onge. I am very tempted to drop the Bedouins, Yoruba, and throw in Denisovans, with Armenians. Kalash is another idea. If I don't see anything by K13, I'm going all supervised.

Kurd
05-03-2015, 02:22 PM
I have all tribals(including some like Vishwa Brahmin), plus Cambodian, Papuan, Yoruba, BedouinA and B, Yamnaya, EEF, WHG, EHG, MA1, Sardinian, Basque, Ami, Atayal, Han, Yakut, Gujaratis, and Onge. I am very tempted to drop the Bedouins, Yoruba, and throw in Denisovans, with Armenians. Kalash is another idea. If I don't see anything by K13, I'm going all supervised.

Including the SSA shifted bedouin should not be a problem as they should form their own cluster at low k. If excluded then members with substantial SSA ancestry would get non informative results. If you have to drop them to conserve the ks, it should not be a big deal either since. Your focus is on S Asia

I like your choices, but try to find a SC Asian pop that does not interfere if possible since there are quite a few members from The surrounding area

Chad Rohlfsen
05-03-2015, 02:32 PM
Including the SSA shifted bedouin should not be a problem as they should form their own cluster at low k. If excluded then members with substantial SSA ancestry would get non informative results. If you have to drop them to conserve the ks, it should not be a big deal either since. Your focus is on S Asia

I like your choices, but try to find a SC Asian pop that does not interfere if possible since there are quite a few members from The surrounding area

Take your pick. I have about everything, except Pashtuns. I've got Sindhis, Pathan, Kalash, Brahui, Balochi, Hazara, Punjabi, etc.

Kurd
05-03-2015, 02:37 PM
Take your pick. I have about everything, except Pashtuns. I've got Sindhis, Pathan, Kalash, Brahui, Balochi, Hazara, Punjabi, etc.

Kalash for the aforementioned reasons. The others have alot of redundancy amongst each other, but if you dont have Pashtun, I would go with either Balochi or Brahui.

Edit: if you find that Armenian has more overlap with EEF than you like, you can look at a caucasus group N or NE of them

Chad Rohlfsen
05-03-2015, 02:54 PM
Okay, sounds good.

Chad Rohlfsen
05-03-2015, 02:56 PM
How about the Burusho, Balochi, Kalash, Lithuanians, Armenians, Georgians, Iranians, Iraqi Jews, and Nganasan?

Chad Rohlfsen
05-03-2015, 03:03 PM
Okay, I've added Thais, Lezgins and Abkhasians, to the list. Running K10. I should have coverage to 15-16, now.

Kurd
05-03-2015, 03:11 PM
How about the Burusho, Balochi, Kalash, Lithuanians, Armenians, Georgians, Iranians, Iraqi Jews, and Nganasan?

Either Armenian or Georgian, since there is too much overlap between the two. Of the two, Georgian would have less EEF overlap. Dont know enough about Iraq Jews or Nganasan. Burusho may be a distraction to Kalash if Kalash is included

Chad Rohlfsen
05-03-2015, 03:51 PM
This is what I have, outside of usual tribals/austro-asiatics

Ami
Atayal
Basque
BedouinB
Bengali
Cambodian
Dai
GujaratiA,B,C,D
Han
Han N China
Karitiana
MA1
Onge
Papuan
Sardinian
Vishwabrahmin
Yakut
Yamnaya
Yoruba
Georgian
Iranian
Iraqi Jew
Kalash
Lezgin
Balochi
Nganasan
Lithuanian
Thai
WHG
EHG
EEF

Chad Rohlfsen
05-03-2015, 03:53 PM
I may trim off some East Asians. Let me see how a couple runs look.

Kurd
05-03-2015, 04:02 PM
Chad,

I am a little curious as to which distinct population macro clusters formed at k =5 and which sub groups they peaked in

Chad Rohlfsen
05-03-2015, 04:06 PM
I can pull that up. It didn't have the new pops though. Would you like a k5, with the new ones, after the K10 is done?

Chad Rohlfsen
05-03-2015, 04:07 PM
I do remember the Onge were pretty split between East Asian, South Asian, and Papuan.

Kurd
05-03-2015, 04:11 PM
I can pull that up. It didn't have the new pops though. Would you like a k5, with the new ones, after the K10 is done?


That would be interesting

Chad Rohlfsen
05-04-2015, 02:06 AM
Okay, I think that the ancients are messing up the data. I'm going to remove them and start at K5. I won't give up, until I figure this out, so, please bear with me.

Dr_McNinja
05-04-2015, 02:09 AM
Okay, I think that the ancients are messing up the data. I'm going to remove them and start at K5. I won't give up, until I figure this out, so, please bear with me.

What was happening?

Chad Rohlfsen
05-04-2015, 03:38 AM
K5
East Asian- Ami/Atayal 100%
Sub-Saharan-Yoruba 100%
Papuan-Papuan 100%
South Asian?-Onge 84%
West Eurasian?-Basque 100%

Chad Rohlfsen
05-04-2015, 03:39 AM
I think the ancients were messing with the clusters. I couldn't get South Asian to break up, through K14.

Dr_McNinja
05-04-2015, 03:57 AM
K5
East Asian- Ami/Atayal 100%
Sub-Saharan-Yoruba 100%
Papuan-Papuan 100%
South Asian?-Onge 84%
West Eurasian?-Basque 100%

This looks promising. What were the results like in this K5 run for some South Asian populations?

Chad Rohlfsen
05-04-2015, 04:04 AM
K6

Siberian-Nganasan 100%
Papuan-Papuan 100%
Yoruba-Yoruba 100%
South Asian?-Onge 81%
East Asian-Ami/Atayal 100%
West Eurasian/European- Basque 98.5%

Chad Rohlfsen
05-04-2015, 04:16 AM
K5

https://docs.google.com/spreadsheets/d/1pNDJTgAdPat8swmm7xRwRDZmqaOb5TJ4xlxCcJWtc4M/edit#gid=1205141266

Dr_McNinja
05-04-2015, 04:32 AM
That's really interesting. I wonder what that South Asian component reflects. It's everywhere!

Chad Rohlfsen
05-04-2015, 04:34 AM
I'm running K7-11, overnight. I'll have a peek around lunchtime. If the Onge component is strong, I'll go ahead and throw in all 2500+ samples, minus ancients.

Chad Rohlfsen
05-04-2015, 04:37 AM
After all this, I may try and make a few ASI zombies. No promises on results. I'm still learning.

tamilgangster
05-04-2015, 12:14 PM
Its redunted to have a seperate south asian component. THe problem with the eurogenes k7 is the the ASE isn't seperated into a "mongoloid" and "australoid" component. It would be great to see a test run which contains, a ANE, EEF WHG etc like the one on eurogenes but also have an onge component.

ANother test I would like to see done is one using, the gedrosian component, Yamnya and onge an east eurasian and if possible a component to represent the SW asian type ENF

Chad Rohlfsen
05-04-2015, 02:39 PM
I'll do what I can. It could take days, or weeks.

Chad Rohlfsen
05-05-2015, 12:36 AM
I didn't have the same luck on K11. It won't break up. I will have to try another route.

Dr_McNinja
05-05-2015, 09:37 PM
What components remained in K11?

Chad Rohlfsen
05-05-2015, 10:48 PM
What components remained in K11?

I don't recall, sorry. I deleted it all, and am trying another angle. I've thrown the Paniya, Onge, and Papuans, in a K run for Europeans. So far, it's working. Paniyas are up to 28% Onge, without Papuans, on K6. I'm running it now, with Papuans. I should have something to post as this progresses. It may take a few days to get it right. I will see which K looks best, and slowly add some tribal pops, along with some SC and Central Asians.

Chad Rohlfsen
05-05-2015, 11:38 PM
K6 has Paniya at..

28% Siberian
14% Papuan
20% Onge
5% Yoruba
25% EEF
8% WHG/EHG

Chad Rohlfsen
05-05-2015, 11:49 PM
I added a few pops and I'm running it again.

Chad Rohlfsen
05-05-2015, 11:51 PM
Ill see how this looks, with Armenians, and add Bedouins to the next run. I've also got two Denisovans. I may have to go up in Ks for that to pop. It'll be a slow grind from here. The SSA is probably compensating for extra BE in their farmer ancestry.

Dr_McNinja
05-06-2015, 12:12 AM
K6 has Paniya at..

28% Siberian
14% Papuan
20% Onge
5% Yoruba
25% EEF
8% WHG/EHG

What sort of numbers do other non-South Asian populations get in terms of Onge/Papuan? Any?

Chad Rohlfsen
05-06-2015, 12:32 AM
What sort of numbers do other non-South Asian populations get in terms of Onge/Papuan? Any?

Onge

Mongola-9%
Yakut-4%
Palestinian-2-3%
Balkans-up to 2%
S Europe-up to 2%
E Europe-up to about 1%

Chad Rohlfsen
05-06-2015, 12:42 AM
Uzbek, Balochi, Uygur, all at 4-6% Onge. I'll add Bedouins and Australians, now. If all looks good, I'll go to K7.

MonkeyDLuffy
05-06-2015, 12:56 AM
Can you check Punjabis and Sindhis for Onge too? I remember seeing a chart where Onge admixture was around 18-20% in North west South asia.

Chad Rohlfsen
05-06-2015, 01:16 AM
I'll get to them. I have to try a few things to maximize the Onge.

Dr_McNinja
05-06-2015, 01:26 AM
Onge

Mongola-9%
Yakut-4%
Palestinian-2-3%
Balkans-up to 2%
S Europe-up to 2%
E Europe-up to about 1%

Interesting, those are higher than K8 ASE for Palestinian, Mongola, and Yakut.

parasar
05-06-2015, 01:28 AM
Can you check Punjabis and Sindhis for Onge too? I remember seeing a chart where Onge admixture was around 18-20% in North west South asia.
Do you mean ASI?

Ongee admixture at 18% would be impossible. There is very little Ongee admixture in South Asia. Overall the Ongee separated from ASI 1700 generations back and from the ANI branch about 2000 generations back. So the Ongee are only slightly more related to ASI than to the ANI branch. And for ANI itself, the difference would be even less, as a portion of ASI ancestors mixed with ANI ancestors to form ANI.

Chad Rohlfsen
05-06-2015, 01:32 AM
Interesting, those are higher than K8 ASE for Palestinian, Mongola, and Yakut.

I'm going to keep rounding out the populations, and see where it goes.

MonkeyDLuffy
05-06-2015, 01:45 AM
Do you mean ASI?

Ongee admixture at 18% would be impossible. There is very little Ongee admixture in South Asia. Overall the Ongee separated from ASI 1700 generations back and from the ANI branch about 2000 generations back. So the Ongee are only slightly more related to ASI than to the ANI branch. And for ANI itself, the difference would be even less, as a portion of ASI ancestors mixed with ANI ancestors to form ANI.

No there was a chart user jesus posted somewhere which showed high percentage of onge in south asia. I forgot to save it

parasar
05-06-2015, 01:57 AM
I'm going to keep rounding out the populations, and see where it goes.

Chad,

Is it possible to test the Paniya's relationship to the Ongee? The Paniya look to be the type of population that the Ongee once separated from way-way back. Of course the Paniya should then also be nearly equally close to others as they are to the Ongee, including Oceanians, ENA, WHG, and perhaps even to Basal.

What do you think about this - one of the reasons the South Asian is not breaking up is that it has certain alleles in common that are absent in others.

jesus
05-06-2015, 02:01 AM
No there was a chart user jesus posted somewhere which showed high percentage of onge in south asia. I forgot to save it
http://www.harappadna.org/2011/04/reference-3-admixture-k11/

paulgill
05-06-2015, 02:01 AM
No there was a chart user jesus posted somewhere which showed high percentage of onge in south asia. I forgot to save itYou may have noticed it at Harrapa Project.

parasar
05-06-2015, 02:02 AM
No there was a chart user jesus posted somewhere which showed high percentage of onge in south asia. I forgot to save it

It must be one of those problematic ADMIXTURE graphs where populations such the Ongee and the Kalash tend to form components even though they are just isolated drifted branches. While the Kalash are at least in the few thousands, the last I recall there were 50 or so Ongee.

MonkeyDLuffy
05-06-2015, 02:20 AM
It must be one of those problematic ADMIXTURE graphs where populations such the Ongee and the Kalash tend to form components even though they are just isolated drifted branches. While the Kalash are at least in the few thousands, the last I recall there were 50 or so Ongee.

You mean that chart isn't accurate?

parasar
05-06-2015, 02:29 AM
You mean that chart isn't accurate?

The chart is accurate, it is just that these ADMIXTURE generated components are not necessarily representing real ancestral populations.
That's the reason Zack would always put this qualifier:

Do note that the admixture components do not necessarily represent real ancestral populations. Also, the names I have chosen for the components should be thought of as mnemonics to ease discussion. I chose them based on which populations in my data these components peaked in. They do not tell anything directly about ancestral populations.

Chad Rohlfsen
05-06-2015, 03:11 AM
Looking okay, so far. Adding a few pops and going K8.

Chad Rohlfsen
05-06-2015, 03:25 AM
Chad,

Is it possible to test the Paniya's relationship to the Ongee? The Paniya look to be the type of population that the Ongee once separated from way-way back. Of course the Paniya should then also be nearly equally close to others as they are to the Ongee, including Oceanians, ENA, WHG, and perhaps even to Basal.

What do you think about this - one of the reasons the South Asian is not breaking up is that it has certain alleles in common that are absent in others.

That could be possible. They look like a bit of a mix. Although, this could be that 60-80% of the ancestry in Dravidians and such remained unchanged for 4-6k years and drifted to its own component. I don't think it's just about the Paniya. You see very similar results in the Hakkipikki, Mala, and others. I've added a few pops to this, to show how similar they are. If the K8 is looking okay, I'll post it and take a few requests for more pops in the K9. Oh, I've added a few Gujaratis to this run. If I get a South Asian cluster, I may trim off a few people. Unless, you guys would like me to add components until the Paniya drop to 64% of what would possibly be an ASI cluster.

Kurd
05-06-2015, 03:39 AM
The chart is accurate, it is just that these ADMIXTURE generated components are not necessarily representing real ancestral populations.
That's the reason Zack would always put this qualifier:

In unsupervised runs, it is probably difficult to get isolated (perhaps drifted) populations like Kalash to break up into ancestral components because they tend to form their own ancestral components at lower ks. However, Chad can probably figure out their admixture components in supervised runs, by running them against zombies comprised of components considered "ancestral" to the Kalash. I recall in such supervised runs, they have been shown to be comprised of only a couple of "ancestral' components, I think with a 60/40 W Asian/ S Asian split. I believe that it was their relatively high W Asian as compared to other S Asian populations, that caused some to believe that Indo Aryans originated in W Asia.

Also, whether classifying a group as the Kalash as drift vs ancestral, we would have to define how old we define "ancestral'. Do you agree?

parasar
05-06-2015, 04:24 AM
In unsupervised runs, it is probably difficult to get isolated (perhaps drifted) populations like Kalash to break up into ancestral components because they tend to form their own ancestral components at lower ks. However, Chad can probably figure out their admixture components in supervised runs, by running them against zombies comprised of components considered "ancestral" to the Kalash. I recall in such supervised runs, they have been shown to be comprised of only a couple of "ancestral' components, I think with a 60/40 W Asian/ S Asian split. I believe that it was their relatively high W Asian as compared to other S Asian populations, that caused some to believe that Indo Aryans originated in W Asia.

Also, whether classifying a group as the Kalash as drift vs ancestral, we would have to define how old we define "ancestral'. Do you agree?

To determine ancestral populations one would need to look at allele diversity. Haplotype diversity of alleles in descendant or admixed recipient populations is lower than in source populations. I can check this when the locus is limited say with Y or mtDNA. But I'm not familiar with the methods that would need to be applied to autosomal datasets.

Taking the example of the Ongee. They have mtDNA M31. Their M31 sub-branch diverged from a mainland South Asian branch perhaps 30000 years back. So while on their branch they will have high diversity, the parallel branches are in South Asia. Furthermore their divergent line(s) (now with many additional sub-branches and mutations) is missing in South Asia. So we could reasonably conclude the Ongee are not ancestral to South Asians but that a South Asian population 30000ybp was ancestral to the Ongee.

Kurd
05-06-2015, 04:44 AM
To determine ancestral populations one would need look at allele diversity. Haplotype diversity of alleles in descendant or admixed recipient populations is lower than in source populations. I can check this when the locus is limited say with Y or mtDNA. But I'm not familiar with the methods that would need to be applied to autosomal datasets.

Taking the example of the Ongee. They have mtDNA M31. Their M31 sub-branch diverged from a South Asian branch perhaps 30000 years back. So while on their branch they will have high diversity, the parallel branches are in South Asia. Furthermore their divergent line(s) (now with many additional sub-branches and mutations) is missing in South Asia. So we could conclude the Ongee are not ancestral to South Asians but that a South Asian population 30000ybp was ancestral to the Ongee.

Autosomal methods involve generating allele frequencies for the K ancestral components by performing an unsupervised admixture analysis. This is followed by creating "zombies" using these allele frequencies. For example, if we are talking about S Asian being one of the components, then the S Asian zombie's admixture would be 100% S Asian. The zombie is then used in supervised runs to infer the admixture proportions of your sample subjects.

I will rephrase my previously posted comment. Is there not a possibility that the Kalash are a relatively intact remnant of an a population ancestral to some other S Asian or W Asian groups.

parasar
05-06-2015, 05:00 AM
Autosomal methods involve generating allele frequencies for the K ancestral components by performing an unsupervised admixture analysis. This is followed by creating "zombies" using these allele frequencies. For example, if we are talking about S Asian being one of the components, then the S Asian zombie's admixture would be 100% S Asian. The zombie is then used in supervised runs to infer the admixture proportions of your sample subjects.

I will rephrase my previously posted comment. Is there not a possibility that the Kalash are a relatively intact remnant of an a population ancestral to some other S Asian or W Asian groups.

I'm not sure, at least from the Y DNA side. Actually on the Y, the Paniya would probably qualify as "a population ancestral to some other S Asian or W Asian groups." The reason I say that is that the Paniya (75%) are on a F line (probably some type of H, but not certain) which is divergent to most other South Asian lines. Furthermore, they have 15% C-M130. They look like a basal South Asian (if not W Asian) population that has received only limited later input (P, Q, R, R1a1, R2) but nothing like the Kalash have.
http://journals.plos.org/plosone/article/figure/image?size=large&id=info:doi/10.1371/journal.pone.0050269.t002

Kurd
05-06-2015, 05:09 AM
I'm not sure, at least from the Y DNA side. Actually on the Y, the Paniya would probably qualify as "a population ancestral to some other S Asian or W Asian groups." The reason I say that is that the Paniya (75%) are on a F line (probably some type of H, but not certain) which is divergent to most other South Asian lines. Furthermore, they have 15% C-M130. They look like a basal South Asian (if not W Asian) population that has received only limited later input (P, Q, R, R1a1, R2) but nothing like the Kalash have.
http://journals.plos.org/plosone/article/figure/image?size=large&id=info:doi/10.1371/journal.pone.0050269.t002

Thanks Parasar, I appreciate the info. Let it never be said that Parasar is not an expert in Y-DNA :)

Sein
05-06-2015, 05:14 AM
All of this is really good stuff.

But I still think one can make life easier, just by having the Onge as a supervised component, along with a bunch of ancient West Eurasian supervised components (ANE, EHG, WHG, EEF, Yamnaya), by also including BedouinB as a supervised component, and finally having Nganasan as a Siberian supervised component.

In theory, this should work fine. But we won't know, unless it's tried.

Then again, unsupervised ADMIXTURE runs are thought to be more "objective" (the post-modernist side of me put in the quotation marks, :biggrin1:), so I can see the value in trying to tease out an unsupervised Onge component that acts like ASI.

tamilgangster
05-06-2015, 06:14 AM
I think that the siberian mostly accounts ANE, possibly also might also include some paleomongoloid. Piramalai kallars also scored around 5% east african, so that explains the yoruba score. I believe that the yoruba is from the Basal eurasian/SW asian type ENF. Can you run the Paniya sample on the original eurogenes k7. On harrapa DNA they score 47% onge. On the other harappa DNA test they score 84% south indian but lack baloch component, but also score high levels of Caucasian papuan and SE Asian

tamilgangster
05-06-2015, 06:17 AM
Onge

Mongola-9%
Yakut-4%
Palestinian-2-3%
Balkans-up to 2%
S Europe-up to 2%
E Europe-up to about 1%

Mongolians having 9% onge, is most likely due to lack of proxies from southern mongoloid populations(non siberian). If thats the case Han must score around 15-20% onge. I think you should add a Han component or an Aytal.

tamilgangster
05-06-2015, 06:20 AM
Do you mean ASI?

Ongee admixture at 18% would be impossible. There is very little Ongee admixture in South Asia. Overall the Ongee separated from ASI 1700 generations back and from the ANI branch about 2000 generations back. So the Ongee are only slightly more related to ASI than to the ANI branch. And for ANI itself, the difference would be even less, as a portion of ASI ancestors mixed with ANI ancestors to form ANI.

On the Harappa DNA they had seperated an onge component, and NW indian populations scored around 15% onge. Similar results also were found on dodecad and eurogenes. By ASI im assuming u mean reich's ASI then they would score any where from 30-40% ASI, which is closer in value to harappa DNA's South Indian component.

tamilgangster
05-06-2015, 06:22 AM
The chart is accurate, it is just that these ADMIXTURE generated components are not necessarily representing real ancestral populations.
That's the reason Zack would always put this qualifier:

Onge percentages are only slightly higher than ASE values on k8

Chad Rohlfsen
05-07-2015, 08:17 PM
K7 might be best. I'll check my K12-16, in a few minutes.

Chad Rohlfsen
05-08-2015, 12:28 AM
Higher Ks are duds. I'm going supervised with Europeans, on a couple runs. Then, I'll add some tribals, bit by bit.

Chad Rohlfsen
05-08-2015, 01:08 AM
I think the K7 was best. Paniya were about 18% Onge, 14% Papuan/Aborigine, 9% Yamnaya/EHG, 31% Bedouin, 26% Siberian, 2% SSA. I'm doing a K11, with Europeans. I'll add a Paniya or two after this run if it's successful.

Sapporo
05-08-2015, 01:15 AM
I think the K7 was best. Paniya were about 18% Onge, 14% Papuan/Aborigine, 9% Yamnaya/EHG, 31% Bedouin, 26% Siberian, 2% SSA. I'm doing a K11, with Europeans. I'll add a Paniya or two after this run if it's successful.

That sounds pretty close to how West Eurasian I'd guess the Paniya to be. Close to 40%. A rough estimate using HAP is 45% or so if we assume the South Indian component is roughly 50% West Eurasian and 50% ASI + East Eurasian like.

https://docs.google.com/spreadsheet/ccc?key=0AuW3R0Ys-P4HdDhib1M5OE1wWENNb2haUFFWZzNBMEE#gid=0

parasar
05-08-2015, 05:00 PM
That sounds pretty close to how West Eurasian I'd guess the Paniya to be. Close to 40%. A rough estimate using HAP is 45% or so if we assume the South Indian component is roughly 50% West Eurasian and 50% ASI + East Eurasian like.

https://docs.google.com/spreadsheet/ccc?key=0AuW3R0Ys-P4HdDhib1M5OE1wWENNb2haUFFWZzNBMEE#gid=0

The 18% Onge, 14% Papuan/Aborigine, 9% Yamnaya/EHG, 26% Siberian all fall on the so called crown Eurasian side. http://www.anthrogenica.com/showthread.php?4416-A-3rd-branch-of-Crown-Eurasians&highlight=crown+eurasian

The 31% Bedouin, 2% SSA falling on the so called basal Eurasian side.

So it looks like the Paniya have been reasonably isolated* from before the basal Eurasian split about 50000ybp, i.e., both ANI ancestors as well ASI ancestors (Ust-Ishim) derive from a Paniya like population.

*Paniya. It is very interesting to see that they have C and F (this covers the OoA bottleneck period), then they lack the Neolithic types, but then again have a variety of low proportion K-M9 types. This could mean: 1. The Neolithics were the last to enter the subcontinent and/or 2. The K-M9 actually originated within a Paniya like population, and/or 3. the K-M9 is also intrusive and later but pushed far deeper in to the pre-existing populations.

C - 15.28%
F - 75%

H - 0%
G - 0%
J - 0%

K - 1.39%
L - 1.39%
Q - 1.39%
R - 1.39%
R1a1 - 2.78%
R2a - 1.39%

Chad Rohlfsen
05-09-2015, 04:50 AM
Are you saying they're some kind of isolate? Certainly not Ust_Ishim like, as he was not derived in any one direction, where as the Paniya obviously have East and West Eurasian mixed ancestry, with Basal Eurasian. South Asia has been a crossing zone for a long time. It's more like a kitchen sink, as Europe is. You have an Onge people, mixed with a Papuan/Aborigine type, or a single in-between pop, that is mixed with Central Asian HG, Near Eastern farmers, and Asian farmers, with possibly a pinch of Yamnaya. The K7 showed an obvious once super-pop which ties in SE Asians with S Asians. It was the SE Asians which were more affected by incoming rice farmers, obviously. I'd bet they're more like 20% Basal Eurasian, 25% West Eurasian, 53% East Eurasian, with about half of that being some type of Basal East Eurasian. That 2% SSA might be real, or it could be extra archaic stuff.

tamilgangster
05-09-2015, 06:58 AM
The 18% Onge, 14% Papuan/Aborigine, 9% Yamnaya/EHG, 26% Siberian all fall on the so called crown Eurasian side. http://www.anthrogenica.com/showthread.php?4416-A-3rd-branch-of-Crown-Eurasians&highlight=crown+eurasian

The 31% Bedouin, 2% SSA falling on the so called basal Eurasian side.

So it looks like the Paniya have been reasonably isolated* from before the basal Eurasian split about 50000ybp, i.e., both ANI ancestors as well ASI ancestors (Ust-Ishim) derive from a Paniya like population.

*Paniya. It is very interesting to see that they have C and F (this covers the OoA bottleneck period), then they lack the Neolithic types, but then again have a variety of low proportion K-M9 types. This could mean: 1. The Neolithics were the last to enter the subcontinent and/or 2. The K-M9 actually originated within a Paniya like population, and/or 3. the K-M9 is also intrusive and later but pushed far deeper in to the pre-existing populations.

C - 15.28%
F - 75%

H - 0%
G - 0%
J - 0%

K - 1.39%
L - 1.39%
Q - 1.39%
R - 1.39%
R1a1 - 2.78%
R2a - 1.39%

Paniyars have not been isolated that long but they have been isolated long enough that they did not recieve geneflow from gedrosian like population. Which can be explained by lack of baloch component

tamilgangster
05-09-2015, 07:10 AM
That sounds pretty close to how West Eurasian I'd guess the Paniya to be. Close to 40%. A rough estimate using HAP is 45% or so if we assume the South Indian component is roughly 50% West Eurasian and 50% ASI + East Eurasian like.

https://docs.google.com/spreadsheet/ccc?key=0AuW3R0Ys-P4HdDhib1M5OE1wWENNb2haUFFWZzNBMEE#gid=0

On Harappa DNA they score 45% onge, and the rest is "south Asian" component. In that test south asian comopent refers to all west eurasian elements in south asia, which includes all the ANE, basal eurasian and gedrosian. For example the Hakkipikki that scores 70% on harappa DNA, scores 10% East asian and 7% east african. And those samples have 20% baloch. So id estimate that Paniya would have over 15% east asian and more than 10% East african.

tamilgangster
05-09-2015, 07:44 AM
That could be possible. They look like a bit of a mix. Although, this could be that 60-80% of the ancestry in Dravidians and such remained unchanged for 4-6k years and drifted to its own component. I don't think it's just about the Paniya. You see very similar results in the Hakkipikki, Mala, and others. I've added a few pops to this, to show how similar they are. If the K8 is looking okay, I'll post it and take a few requests for more pops in the K9. Oh, I've added a few Gujaratis to this run. If I get a South Asian cluster, I may trim off a few people. Unless, you guys would like me to add components until the Paniya drop to 64% of what would possibly be an ASI cluster.

In many k runs Irulas show up as a distinct component because those tests only go far so back.

Arbogan
05-09-2015, 12:04 PM
How can yamnayas be 13% Onge.

tamilgangster
05-09-2015, 01:03 PM
How can yamnayas be 13% Onge.

Thats due to lack of proper proxies, this test has many flaws in it, its a prototype

Chad Rohlfsen
05-09-2015, 01:40 PM
Are you talking about the K5? It's about the reference pops selected by the program.

parasar
05-09-2015, 03:19 PM
Paniyars have not been isolated that long but they have been isolated long enough that they did not recieve geneflow from gedrosian like population. Which can be explained by lack of baloch component

Do have any numbers-timeframes?
Long enough means more than 500 generations in genetic terms (per Metspalu's resolution limits) or more than 12500 years. That does look correct as far as the Neolithic time-frame is concerned, J and G both are missing.
But H is missing too, or at least the common South Asian varieties of H since Paniya F* could be an early H split from the rest of H.

parasar
05-09-2015, 03:27 PM
Are you saying they're some kind of isolate? Certainly not Ust_Ishim like, as he was not derived in any one direction, where as the Paniya obviously have East and West Eurasian mixed ancestry, with Basal Eurasian. South Asia has been a crossing zone for a long time. It's more like a kitchen sink, as Europe is. You have an Onge people, mixed with a Papuan/Aborigine type, or a single in-between pop, that is mixed with Central Asian HG, Near Eastern farmers, and Asian farmers, with possibly a pinch of Yamnaya. The K7 showed an obvious once super-pop which ties in SE Asians with S Asians. It was the SE Asians which were more affected by incoming rice farmers, obviously. I'd bet they're more like 20% Basal Eurasian, 25% West Eurasian, 53% East Eurasian, with about half of that being some type of Basal East Eurasian. That 2% SSA might be real, or it could be extra archaic stuff.

The Paniya - a type of isolate, and as I mentioned from "before the basal Eurasian split." This would make the Paniya only party Ust-Ishim like who, unlike the Paniya, had West 0 Eurasian - the so called "basal Eurasian."

Chad Rohlfsen
05-09-2015, 03:47 PM
Not really Basal Eurasian, as far as the farmer stuff, just undifferentiated Crown Eurasian. The Onge and/or Ust-Ishim like stuff is actually on a slight East Eurasian shift. Very small though. Remember, even Han were slightly closer to Ust-Ishim, than WHG's. There very well could be some Kostenki "Basal" stuff in WHG, which caused this difference. I would bet in the mid-to-low single digits range.

parasar
05-09-2015, 05:50 PM
Not really Basal Eurasian, as far as the farmer stuff, just undifferentiated Crown Eurasian. The Onge and/or Ust-Ishim like stuff is actually on a slight East Eurasian shift. Very small though. Remember, even Han were slightly closer to Ust-Ishim, than WHG's. There very well could be some Kostenki "Basal" stuff in WHG, which caused this difference. I would bet in the mid-to-low single digits range.

Exactly, and the Ongee even slightly closer than the Han.

J Man
05-09-2015, 09:18 PM
Do have any numbers-timeframes?
Long enough means more than 500 generations in genetic terms (per Metspalu's resolution limits) or more than 12500 years. That does look correct as far as the Neolithic time-frame is concerned, J and G both are missing.
But H is missing too, or at least the common South Asian varieties of H since Paniya F* could be an early H split from the rest of H.

J is present among other "Hill tribe forager" groups from India though. It is all of the South Asias J2b variety as well.

Kale
05-09-2015, 09:57 PM
There very well could be some Kostenki "Basal" stuff in WHG, which caused this difference. I would bet in the mid-to-low single digits range.

MbutiPygmy Ust'-Ishim : Kostenki14 Loschbour 0.0004 0.055
MbutiPygmy Ust'-Ishim : Kostenki14 Dai 0.0006 0.096

That's as dead even a result as you can get. Fair enough to say Kostenki has nothing basal to Ust-Ishim.

Dr_McNinja
05-11-2015, 12:04 PM
I was messing around with the HarappaWorld calculator and removed the S-Indian component. Most of the remaining admixture went into Baloch, Caucasian, SE-Asian, NE-Asian, Papuan, and SW-Asian.

When I removed Papuan, about an equivalent amount dispersed into all the other components, including the African ones. Also the Papuan component's allele frequencies were like 100% for over 40,000 of the SNPs. Even more if you consider anything over, say, 90%.

tamilgangster
05-11-2015, 12:08 PM
The Paniya - a type of isolate, and as I mentioned from "before the basal Eurasian split." This would make the Paniya only party Ust-Ishim like who, unlike the Paniya, had West 0 Eurasian - the so called "basal Eurasian."

Paniya aren't part Ust-Ishim, or atleast there is no proven links. Its still unclear who the Ust Ishim are, so one can't really make such conclusions. Paniyars, are 45-50% South eurasian and the rest is archaic west eurasian which most likely has basal eurasian affinities. Whether they have ANE is unclear. The paniyars haven't been tested for ANE ENF etc so one can't determine until the test has been done

parasar
05-11-2015, 03:19 PM
Paniya aren't part Ust-Ishim, or atleast there is no proven links. Its still unclear who the Ust Ishim are, so one can't really make such conclusions. Paniyars, are 45-50% South eurasian and the rest is archaic west eurasian which most likely has basal eurasian affinities. Whether they have ANE is unclear. The paniyars haven't been tested for ANE ENF etc so one can't determine until the test has been done

I said party Ust-Ishim like rather than part Ust-Ishim. By the former I mean shared OoA ancestry. They do look have some derived ancestry (about 10%) from a population related to Ust-Ishim which from their Y-dna looks to be of the ANE type.

Dr_McNinja
05-11-2015, 09:49 PM
I made two Harappa-derived calculators, with 11 components. I took out Siberian/Beringian/American and S-Indian. In one there is no Papuan and in the other (K11b), there is.

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=72 (no Papuan)

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=73 (Papuan)

I just ran the South Asian kits through it because it's otherwise the same as Harappa.

A little bit of a pattern does emerge. Baloch obviously seems to point in the direction of Balochistan, but Caucasian seems to be raised generally (above 16-17%) in the northern/northwestern areas (i.e, Pahari areas of Pakistan/India and in Afghanistan towards the Caucasus).

The difference between SE-Asian and NE-Asian clarifies a bit on geographic ancestry. Although there is some leak/overlap between the two components (like within my family) since they are acting as proxies for unknown or missing ancestral components, but for the most part we see ashwinb (Gujarat/Maharashtra) and parasar (Bihar) who otherwise have high S-Indian revealed to have lower SE-Asian and more NE-Asian and Papuan. Papuan of course is high everywhere, including in the northwest throughout the extent of Pakistan. So a low SE-Asian to high NE-Asian ratio can safely be assumed to represent ancestry from closer to the north (Himalayas). This has interesting implications for ashwinb. I recall that other Gujarati groups like Lohanas had noticeable NE-Asian too, and their history is supposed to be from the north.

So there are all these clues pointing to a more complex admixture history in North India.

In the same way that we can see all these components (African, Papuan, NE-Asian, SE-Asian, SW-Asian) and know they represent a "missing" component (S-Indian), we could probably presume that the normal results with Siberian, European, Caucasian, Amerindian, East Asian, etc could possibly represent a missing component. And since Yamnaya didn't account for South Asian WHG totally, it's possible it's an as-yet undiscovered eastern Central Asian group (perhaps where R1a-Z93 came from).

Chad Rohlfsen
05-12-2015, 01:44 AM
I'm seeing the same thing. The Paniya prefer Turks over Atayal, in a big way. So, that SE Asian stuff could be part of the ASI. It could be a mix of both East Asian and West Asian Neolithic mixing with a Papuan/Onge mixed folks. There's really no way to tell if that EHG/ANE stuff was already there. Han Chinese come out 75% Turkic-like, and 25% Atayal like. That 3-1 ratio goes to about 6-1 in the Paniya. This Atayal stuff must be part of a later expansion, post-ASI. ASI may be more from Chinese than Indo-Chinese Neolithic. It kind of looks like ASI stuff once covered from India to Indo-China. I'm tinkering with the K10. I've added a Mongol and Amerindian component, to see if they're better fits. Possibly an MA1 component next. If all looks good, I'll add several more pops. I'll eventually work my way to including all pops.

tamilgangster
05-12-2015, 10:01 AM
I said party Ust-Ishim like rather than part Ust-Ishim. By the former I mean shared OoA ancestry. They do look have some derived ancestry (about 10%) from a population related to Ust-Ishim which from their Y-dna looks to be of the ANE type.

Its possible that they do contain some DNA from the Ust Ishim lineage. Its still unknown weather they have ANE, and how related ANE is to Ust Ishim

tamilgangster
05-12-2015, 10:13 AM
I made two Harappa-derived calculators, with 11 components. I took out Siberian/Beringian/American and S-Indian. In one there is no Papuan and in the other (K11b), there is.

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=72 (no Papuan)

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=73 (Papuan)

I just ran the South Asian kits through it because it's otherwise the same as Harappa.

A little bit of a pattern does emerge. Baloch obviously seems to point in the direction of Balochistan, but Caucasian seems to be raised generally (above 16-17%) in the northern/northwestern areas (i.e, Pahari areas of Pakistan/India and in Afghanistan towards the Caucasus).

The difference between SE-Asian and NE-Asian clarifies a bit on geographic ancestry. Although there is some leak/overlap between the two components (like within my family) since they are acting as proxies for unknown or missing ancestral components, but for the most part we see ashwinb (Gujarat/Maharashtra) and parasar (Bihar) who otherwise have high S-Indian revealed to have lower SE-Asian and more NE-Asian and Papuan. Papuan of course is high everywhere, including in the northwest throughout the extent of Pakistan. So a low SE-Asian to high NE-Asian ratio can safely be assumed to represent ancestry from closer to the north (Himalayas). This has interesting implications for ashwinb. I recall that other Gujarati groups like Lohanas had noticeable NE-Asian too, and their history is supposed to be from the north.

So there are all these clues pointing to a more complex admixture history in North India.

In the same way that we can see all these components (African, Papuan, NE-Asian, SE-Asian, SW-Asian) and know they represent a "missing" component (S-Indian), we could probably presume that the normal results with Siberian, European, Caucasian, Amerindian, East Asian, etc could possibly represent a missing component. And since Yamnaya didn't account for South Asian WHG totally, it's possible it's an as-yet undiscovered eastern Central Asian group (perhaps where R1a-Z93 came from).

The ANE values for south Asian populations on this run are only around 20% range, vs on the eurogenes run its averages around 30%. A;so another thing I notice, is the the NE asian is fairly high on both tests. I believe this represents either ANE, or Austroasiatic migration. You should run these tests again but also include a siberian, american, and Bering component to determine how much of the NE asian is ANE and how much is remnants from archaic mongoloid strains. The combined Papuan and SE asian values are similar to ASE values on eurogenes K7 and the onge component on harrappa DNA. The SW asian and East african most likely represent Basal Eursian from SW asian type ENF.

Chad Rohlfsen
05-12-2015, 02:24 PM
I'm getting interesting results with a Mongol/Manchuria and Amerindian component. I may try to sneak in MA1 now. I think it's getting close to adding all South Asians.

Dr_McNinja
05-12-2015, 03:40 PM
The ANE values for south Asian populations on this run are only around 20% range, vs on the eurogenes run its averages around 30%. A;so another thing I notice, is the the NE asian is fairly high on both tests. I believe this represents either ANE, or Austroasiatic migration. You should run these tests again but also include a siberian, american, and Bering component to determine how much of the NE asian is ANE and how much is remnants from archaic mongoloid strains. The combined Papuan and SE asian values are similar to ASE values on eurogenes K7 and the onge component on harrappa DNA. The SW asian and East african most likely represent Basal Eursian from SW asian type ENF.Removing the Siberian/Amerindian/Beringian results in the ANE from those components going into other ANE-containing components like Baloch/Caucasian/NE-Euro and the East Eurasian going mostly into NE-Asian, some into SE-Asian.

Papuan is separate from ASE. In K7 to K8 where an Oceanian component was introduced, ASE remained the same mostly, but everything else dumped a little into Oceanian.

So Harappa S-Indian is comprised of most of the SE-Asian/ASE, a good chunk of the NE-Asian, all the Papuan, and a bunch of the ENF. It's a component of modern South Indians which probably represents ancestral (pre-2000-3000ybp) India, both North and South, before most of the modern Indo-Europeanization happened. But it's present highest in South India today who had the least influence from Central/West Asia in the post-2k-3kybp timeframe.

Chad Rohlfsen
05-12-2015, 04:29 PM
K12 looks great. Ill try SC Asians with it now. If it has issues, I'll throw in MA1 as a component. Central Asians are almost as Mongol like, as they are Turkic. Does that sound legit?

Dr_McNinja
05-12-2015, 08:13 PM
K12 looks great. Ill try SC Asians with it now. If it has issues, I'll throw in MA1 as a component. Central Asians are almost as Mongol like, as they are Turkic. Does that sound legit?

What components are in the K12?

Chad Rohlfsen
05-12-2015, 09:22 PM
WHG, EEF, Yamnaya, Bedouin, Onge, Papuan, Atayal, Yakut, Karitiana, Daur, Yoruba, and Nganasan. I've added MA1, with SC Asians. I'll know in 3 hours if I've worked around the Kalash. I may need to add a 14, from some South Chinese groups to find the real Mongol input vs earlier stuff. It depends if they lack ASI related stuff.

Chad Rohlfsen
05-12-2015, 10:42 PM
Okay, Hmong and She folks look the best. They totally lack Turkic and ASI stuff. Any objections to using them as an East Asian Neolithic pop, or is the preference with southern Han?

Dr_McNinja
05-13-2015, 02:12 AM
Okay, Hmong and She folks look the best. They totally lack Turkic and ASI stuff. Any objections to using them as an East Asian Neolithic pop, or is the preference with southern Han?

As long as the component is giving the expected proportions, if it works it works.

Chad Rohlfsen
05-13-2015, 03:27 AM
Okay. I had to scrap the Karitiana, since using ANE and other Asians gave the program issues with finding a solution. So, I've removed the natives and gone with ANE and Han, as my 12th and 13th components. All looks well, but I will probably be asleep before it finishes. I'll post an update as soon as I have some news.

Dr_McNinja
05-16-2015, 10:40 PM
Any news?

Chad Rohlfsen
05-17-2015, 12:01 AM
Still doing some runs. I may try it without Yamnaya, and just go straight derived Eurasians(WHG, EHG, Han, Mongol, Turk, Onge, Papuan, Atayal, and possibly Karitiana, plus EEF, maybe Druze, and Yoruba.

Chad Rohlfsen
05-17-2015, 12:17 AM
What's really the preference here? I'll tailor it to everyone's request.

Kale
05-17-2015, 01:54 AM
I've got a few thoughts...
- What happens with Karitiana if you don't run them in the initial component forming stage?
- I think Samaritans (if you have them) might be a better choice then Druze. Less isolated/endogomous/whatever and just about as Near Eastern as can get on most tests.
- Something like Ju_Hoan or Mbuti might be a good choice in addition to Yoruba.

Also on the topic of ASI, I've one D-stat in mind that I haven't seen run yet that could perhaps be interesting or informative.

Chimp:Ust-Ishim, Onge:[Eurasians without basal, african, or denisovan admixture...so WHG, EHG, MA1, Kostenki, Dai, Han, Japanese, etc.]

Dr_McNinja
05-17-2015, 02:08 AM
What's really the preference here? I'll tailor it to everyone's request.
Something that can distinguish an Onge-like component from ANE, other types of East Eurasian (like Oceanian), etc.

Kind of like West Eurasia K8 but perhaps allowing ENF/WHG/ANE to form components (like Gedrosia, Steppe, EHG, European, etc).

Even a West Eurasia K8-type would be cool since I've only seen a few results from that one (from those who paid).

Chad Rohlfsen
05-17-2015, 02:11 AM
Something that can distinguish an Onge-like component from ANE, other types of East Eurasian (like Oceanian), etc.

Kind of like West Eurasia K8 but perhaps allowing ENF/WHG/ANE to form components (like Gedrosia, Steppe, EHG, European, etc).

Even a West Eurasia K8-type would be cool since I've only seen a few results from that one (from those who paid).

Okay. Should I scrap Yamnaya?

Chad Rohlfsen
05-17-2015, 02:18 AM
I've got a few thoughts...
- What happens with Karitiana if you don't run them in the initial component forming stage?
- I think Samaritans (if you have them) might be a better choice then Druze. Less isolated/endogomous/whatever and just about as Near Eastern as can get on most tests.
- Something like Ju_Hoan or Mbuti might be a good choice in addition to Yoruba.

Also on the topic of ASI, I've one D-stat in mind that I haven't seen run yet that could perhaps be interesting or informative.

Chimp:Ust-Ishim, Onge:[Eurasians without basal, african, or denisovan admixture...so WHG, EHG, MA1, Kostenki, Dai, Han, Japanese, etc.]

I'm trying to avoid Near Easterners with more than a couple percent of SSA.

Kurd
05-17-2015, 03:07 AM
Okay. Should I scrap Yamnaya?

Scrap it, and go derived

Kale
05-17-2015, 03:11 AM
Is it even possible to find Near Easterners more Near Eastern like than EEF that don't have a more than a couple percent SSA? I figured Samaritans would be about the top candidate if there was one.

EDIT: I agree with scrapping Yamnaya too. Too recent, too conflating.

Chad Rohlfsen
05-17-2015, 03:39 AM
Okay, starting at K10

Atayal
BedouinB
Chukchi
Karitiana
EHG
WHG
EEF
Onge
Papuan
Yoruba

Chad Rohlfsen
05-17-2015, 05:02 PM
K10 was a little goofy. I need to break up the ENA a little more. Maybe, add Motala. I'll have a few runs done in the next couple of hours.

tamilgangster
05-18-2015, 11:50 AM
Still doing some runs. I may try it without Yamnaya, and just go straight derived Eurasians(WHG, EHG, Han, Mongol, Turk, Onge, Papuan, Atayal, and possibly Karitiana, plus EEF, maybe Druze, and Yoruba.

THese components are fairly representitive and accurate, I suggest getting rid of the Turk Component though, since its already mixed, and get rid of the druze component since neither of them represent Archaic populations. Having Aytal, Han and Mongol are good for representing east eurasian populations

tamilgangster
05-18-2015, 11:52 AM
Okay, starting at K10

Atayal
BedouinB
Chukchi
Karitiana
EHG
WHG
EEF
Onge
Papuan
Yoruba

Add Han sample also because Aytal only accounts for Austronesian mongoloids but not for other east eurasian populations. Get rid of the bedouin component, because having it along with EEF is redundent

Chad Rohlfsen
05-18-2015, 10:28 PM
I need the Bedouin for North Africans and such, plus another West Asian, like Palestinians or Druze, or Yamnaya and Corded become full EHG. I'm playing around with it. It may take a few days. I have a Southern Han group. The Yakuts don't have much Near Eastern, at all. Just a couple of percent, same as the Daur. They do differ in the amount of Han and Chukchi like stuff. It does work.

Chad Rohlfsen
05-19-2015, 03:35 AM
Getting close, everyone. I should have something before the weekend is up.

Dr_McNinja
05-19-2015, 10:52 PM
When I remove Oceanian from Eurogenes K36, 50-60% of it goes into Malayan (Southeast Asian) and the rest just splits up into all the other components. But not West/North Caucasus. It goes into European, African, Siberian/Amerindian, Armenian, and Near East.

In HarappaWorld calculator, only 20-25% of Papuan went into SE-Asian. 30-35% went into NE-Asian. The rest split up into the rest of the categories.

That's obviously something archaic. Could it be Denisovan? Or perhaps more recent like archaic South Asian-specific ANE? I figure Denisovan would show up as African?

Kale
05-20-2015, 12:50 PM
I've seen many other calculators where if you put a Neanderthal/Denisovan genome in, it comes out 99% African, 1% Eurasian.

Dr_McNinja
05-21-2015, 07:54 PM
MDLP K27 is interesting. Here's the original:


0.00% Nilotic-Omotic
38.96% Ancestral-South-Indi
7.04% North-European-Balti
1.79% Uralic
0.50% Australo-Melanesian
1.72% East-Siberean
0.00% Ancestral-Yayoi
5.75% Caucasian-Near-Easte
0.00% Tibeto-Burman
0.63% Austronesian
0.00% Central-African-Pygm
0.63% Central-African-Hunt
0.01% Nilo-Saharian
0.00% North-African
27.32% Gedrosia-Caucasian
0.00% Cushitic
0.04% Congo-Pygmean
0.00% Bushmen
1.35% South-Meso-Amerindia
1.68% South-West-European
0.24% North-Amerindian
2.61% Arabic
1.51% North-Circumpolar
4.29% Kalash
0.76% Papuan-Australian
2.62% Baltic-Finnic
0.55% Bantu



[1,] "79.2% Brahmin_Uttar-Pradesh + 20.8% Turkmenian" "2.922"
[2,] "82.7% Brahmin_Uttar-Pradesh + 17.3% Kurd" "3.0369"
[3,] "81.2% Brahmin_Uttar-Pradesh + 18.8% Iranian" "3.0809"
[4,] "76% Brahmin_Uttar-Pradesh + 24% Parsi" "3.0931"
[5,] "77.8% Meghawal + 22.2% Nogay" "3.1912"
[6,] "80.8% Hindi + 19.2% Lak" "3.2427"
[7,] "74.5% Hindi + 25.5% Tadijk" "3.2494"
[8,] "17% Azeri + 83% Brahmin_Uttar-Pradesh" "3.2729"
[9,] "80.4% Hindi + 19.6% Urkarah" "3.276"
[10,] "80.9% Hindi + 19.1% Lezgin" "3.2881"

Removing South Indian:



0.29% Nilotic-Omotic
9.36% North-European-Balti
3.08% Uralic
2.24% Australo-Melanesian
3.14% East-Siberean
1.68% Ancestral-Yayoi
10.56% Caucasian-Near-Easte
2.00% Tibeto-Burman
6.40% Austronesian
0.00% Central-African-Pygm
0.95% Central-African-Hunt
1.14% Nilo-Saharian
0.00% North-African
28.96% Gedrosia-Caucasian
0.00% Cushitic
0.00% Congo-Pygmean
0.00% Bushmen
1.01% South-Meso-Amerindia
0.06% South-West-European
1.45% North-Amerindian
4.32% Arabic
1.77% North-Circumpolar
11.94% Kalash
3.98% Papuan-Australian
3.44% Baltic-Finnic
2.22% Bantu

Removing South Indian and Gedrosian:


HRP0349 FTDNA

0.62% Nilotic-Omotic
17.44% North-European-Balti
5.36% Uralic
2.19% Australo-Melanesian
2.58% East-Siberean
1.71% Ancestral-Yayoi
14.21% Caucasian-Near-Easte
1.45% Tibeto-Burman
6.91% Austronesian
0.02% Central-African-Pygm
0.97% Central-African-Hunt
1.55% Nilo-Saharian
0.02% North-African
0.18% Cushitic
0.00% Congo-Pygmean
0.00% Bushmen
1.27% South-Meso-Amerindia
0.00% South-West-European
2.21% North-Amerindian
8.72% Arabic
1.96% North-Circumpolar
18.51% Kalash
4.17% Papuan-Australian
5.29% Baltic-Finnic
2.67% Bantu


---

HRP0393

1.13% Nilotic-Omotic
23.53% North-European-Balti
9.44% Uralic
2.05% Australo-Melanesian
0.26% East-Siberean
0.19% Ancestral-Yayoi
13.33% Caucasian-Near-Easte
2.66% Tibeto-Burman
5.51% Austronesian
0.00% Central-African-Pygm
0.67% Central-African-Hunt
1.84% Nilo-Saharian
0.81% North-African
1.77% Cushitic
0.00% Congo-Pygmean
0.00% Bushmen
1.12% South-Meso-Amerindia
0.04% South-West-European
2.95% North-Amerindian
6.47% Arabic
0.00% North-Circumpolar
17.54% Kalash
2.84% Papuan-Australian
4.93% Baltic-Finnic
0.95% Bantu
Removing South Indian, Gedrosian, Kalash, and North-European-Baltic:


HRP0349 FTDNA

0.69% Nilotic-Omotic
9.89% Uralic
2.59% Australo-Melanesian
2.15% East-Siberean
1.67% Ancestral-Yayoi
22.30% Caucasian-Near-Easte
1.13% Tibeto-Burman
7.92% Austronesian
0.07% Central-African-Pygm
1.11% Central-African-Hunt
1.88% Nilo-Saharian
0.49% North-African
0.40% Cushitic
0.02% Congo-Pygmean
0.00% Bushmen
2.10% South-Meso-Amerindia
3.86% South-West-European
3.21% North-Amerindian
12.04% Arabic
2.53% North-Circumpolar
4.89% Papuan-Australian
16.03% Baltic-Finnic
3.02% Bantu


---

HRP0393

1.14% Nilotic-Omotic
14.31% Uralic
2.39% Australo-Melanesian
0.05% East-Siberean
0.10% Ancestral-Yayoi
21.18% Caucasian-Near-Easte
2.34% Tibeto-Burman
6.28% Austronesian
0.19% Central-African-Pygm
0.82% Central-African-Hunt
2.00% Nilo-Saharian
1.87% North-African
1.54% Cushitic
0.00% Congo-Pygmean
0.00% Bushmen
2.35% South-Meso-Amerindia
7.79% South-West-European
4.26% North-Amerindian
9.56% Arabic
0.00% North-Circumpolar
3.50% Papuan-Australian
17.01% Baltic-Finnic
1.32% Bantu

---

HRP0341

0.80% Nilotic-Omotic
12.77% Uralic
2.44% Australo-Melanesian
1.11% East-Siberean
0.52% Ancestral-Yayoi
22.28% Caucasian-Near-Easte
0.40% Tibeto-Burman
6.97% Austronesian
0.25% Central-African-Pygm
0.14% Central-African-Hunt
1.58% Nilo-Saharian
2.25% North-African
2.64% Cushitic
0.00% Congo-Pygmean
0.00% Bushmen
1.85% South-Meso-Amerindia
2.69% South-West-European
4.28% North-Amerindian
12.64% Arabic
1.99% North-Circumpolar
3.45% Papuan-Australian
17.26% Baltic-Finnic
1.69% Bantu

---

Duffy

0.58% Nilotic-Omotic
13.92% Uralic
2.49% Australo-Melanesian
0.17% East-Siberean
3.09% Ancestral-Yayoi
21.44% Caucasian-Near-Easte
1.87% Tibeto-Burman
6.04% Austronesian
0.04% Central-African-Pygm
0.66% Central-African-Hunt
1.94% Nilo-Saharian
2.85% North-African
1.26% Cushitic
0.00% Congo-Pygmean
0.00% Bushmen
3.75% South-Meso-Amerindia
2.23% South-West-European
1.48% North-Amerindian
11.20% Arabic
1.93% North-Circumpolar
4.24% Papuan-Australian
16.65% Baltic-Finnic
2.19% Bantu
I'm gonna run more kits through this last K23 rendition and put it into a spreadsheet.

There are several Southeast Asian components, including multiple Oceanian components.

Population averages of the original K27: https://docs.google.com/spreadsheets/d/1k31ReQv2QWhfzynBAoAXh052O9kegFCyU7a64xTFexc/edit#gid=0

North-European-Baltic peaked in Latvia, Sweden-Mesolithic, then Lithuania, then Russia/Belarus.

Baltic-Finnic peaked in Finland. And "BRA", whatever that is.

Southwest European peaked in Basque and Sardinian.

Uralic peaks in Selkup and Ket, then Udmurd/Mari/Saami.

I could've probably left the original North European component in there, I might put it back and run everything again, but the remaining components (Baltic-Finnic, SW-European, Uralic) do capture it all just as well.

Australo-Melanesian peaks in Tongan, Mamanwa, and Samoan. Then NAN_Melanesian.

Tibeto-Burman peaks in Aonaga, Naga, and Nysha. Then East Indian. Han have 54% of it.

Yayoi is in Japan.

Austronesia peaks in Mlabri, Iban, Htin, and Javanese. 71% in Dai.

Papuan is the usual Papuan.

Onge are in the spreadsheet. They are 8.3% Papuan, 14.83% Austronesia, 8.76% Yayoi, 2.63% East Siberian, 8.16% Australo-Melanesian, 48% South Indian, 4.37% Nilotic-Omotic, with the rest scattered in different components.

Dr_McNinja
05-21-2015, 10:27 PM
Here are some of the results so far (mostly Indians/Pakistanis, I'll do Afghans next):

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=76

Scroll all the way to the right for the components grouped into categories.

It looks like Austronesian is definitely the closest component to ASI, but it's still too East Eurasian but that's always going to be the case unless we get actual ancient ASI DNA.

Australo-Melanesian, like Papuan, is high everywhere, but doesn't really have a peak. It's even more "basal" in appearance than Papuan. It seems to also be connected to ASI.

Tibeto-Burman looks like a Himalayan signal, but is evidence that "South Indian"/"South Asian" components are capturing a bunch of East Eurasian admixture that is a northernly or northeasternly signal. Without it, the "ASI" totals look low and almost even for everyone, not what we are used to or would expect. And that's probably closer to reality but there's probably some ASI in this too, since it's over a quarter in Southeast Asian peoples.

I'm surprised at how high Yayoi is. It could partially contain the real East Eurasian component of ASI here (similar to ASE in Eurogenes K7/K8) because it's an ancient isolate from the indigenous peoples of the Ryukyu islands which lie between Japan and Taiwan. Chinese are 17.77% Yayoi, 51.01% Tibeto-Burman, 30.19% Austronesian while Japanese are 89.26% Yayoi, 7.50% Tibeto-Burman, 1.23% Austronesian.

EDIT: There's a lot of African and there's some ASI in there without a doubt as well and it probably explains any more discrepancies in ASI (why some are lower or higher than Eurogenes ASE). Also some ENF. Cushitic looks like an ENF signal that is peaking in Southwest India (Sindh/Gujarat/South-Southwest India).

Dr_McNinja
05-21-2015, 10:56 PM
I added both of my parents and it looks clearly like some of my Uralic went into East-Siberian, which would explain the discrepancy in East European scores between us in here and in other calculators, and also the large Arctic percentage observed for myself in other calculators.

Also, Cushitic, since it's from Ethiopia, could be a part of ASI like Nilo-Omotic almost definitely is (latter is high in Onge). It could be like a basal Eurasian part of ASI or something.

EDIT: Also despite representing probably distinct signals, Tibeto-Burman and Austronesian are getting mixed back and forth even though "Total ASI" isn't fluctuating as much (my grandmother and mother are an example, they should otherwise be identical... her parents were first cousins, and her father is related to my dad and her cousin).

EDIT #2: A significant proportion of Uralic is also from ASI. This chart shows the relationship too:

http://i.imgur.com/HHG76CV.png

EDIT #3: The Uralic-EastSiberian thing happened to HRP0370 too.

Varun R
05-21-2015, 10:59 PM
Thanks McNinja! My Austronesian looked pretty high, which was expected. I do tend to get relatively high African scores in these runs...

Edit: Thanks for explaining what some of the African components represent.

Chad Rohlfsen
05-21-2015, 11:35 PM
I'm getting closer and closer. Just fine tuning with the right samples.

J Man
05-22-2015, 12:01 AM
MDLP K27 is interesting. Here's the original:



Removing South Indian:



Removing South Indian and Gedrosian:

Removing South Indian, Gedrosian, Kalash, and North-European-Baltic:

I'm gonna run more kits through this last K23 rendition and put it into a spreadsheet.

There are several Southeast Asian components, including multiple Oceanian components.

Population averages of the original K27: https://docs.google.com/spreadsheets/d/1k31ReQv2QWhfzynBAoAXh052O9kegFCyU7a64xTFexc/edit#gid=0

North-European-Baltic peaked in Latvia, Sweden-Mesolithic, then Lithuania, then Russia/Belarus.

Baltic-Finnic peaked in Finland. And "BRA", whatever that is.

Southwest European peaked in Basque and Sardinian.

Uralic peaks in Selkup and Ket, then Udmurd/Mari/Saami.

I could've probably left the original North European component in there, I might put it back and run everything again, but the remaining components (Baltic-Finnic, SW-European, Uralic) do capture it all just as well.

Australo-Melanesian peaks in Tongan, Mamanwa, and Samoan. Then NAN_Melanesian.

Tibeto-Burman peaks in Aonaga, Naga, and Nysha. Then East Indian. Han have 54% of it.

Yayoi is in Japan.

Austronesia peaks in Mlabri, Iban, Htin, and Javanese. 71% in Dai.

Papuan is the usual Papuan.

Onge are in the spreadsheet. They are 8.3% Papuan, 14.83% Austronesia, 8.76% Yayoi, 2.63% East Siberian, 8.16% Australo-Melanesian, 48% South Indian, 4.37% Nilotic-Omotic, with the rest scattered in different components.

"BRA" is probably La Brana.

Dr_McNinja
05-22-2015, 10:23 PM
I tweaked the MDLP K27 calculator again and uploaded new results for the kits I have here:

https://docs.google.com/spreadsheet/ccc?key=0AuXBmvmgdkfVdFMtRHVlZDBuQ3lMcjhxMDE4V3JoY lE#gid=76

Scroll all the way to the right. The non-colored columns tally up the stuff in the colored columns which are original components.

I reintroduced Gedrosian to soak up the ANE so it wasn't all going into European/Arctic components. I tried to isolate the non-Uralic vs. Uralic part of European. The cumulative European score pretty well approximated other calculators no matter what. So I kept Southwest Europe which would use up WHG and some kind of Mediterranean/EEF and Uralic. The latter being quite East Eurasian-like and comprised of the Samoyedic people from the Uralic and East Siberian components (Selkup, Ket, Nganassan). However, it technically probably represents something usually showing up (or which should show up) in West Eurasian components since if you want to get a total West Eurasian score that is similar to what we see in everything else, Uralic has to count as West Eurasian here.

But on top of that, Uralic is definitely containing a little ASI/ASE for some people.

The spread between users is decreased a little as there's a bit of a Mediterranean boost at the low end boosting some South Indians, the Iranian, the Afghans, etc. The upper end seems within range of the normal European components we've seen in Gedmatch calculators.

Technically it should be combined with Native American to represent the full Central Asian/Steppe contribution.

ASI is still a bit volatile. Papuan is ironically quite stable across calculators for individuals. Even adding/removing components barely changed proportions of that component. Whatever it is, it's something real that isn't recognized as anything else for now. Austronesian, Tibeto-Burman, and East Eurasian (Yayoi) do mix back and forth a little. So the difference between ASI here and ASE in K7/K8 is probably in East Eurasian, maybe African. Also, there's a good chunk (1-2% probably) of ASI in Gedrosian for people with high Gedrosian. Likewise, you see higher ASI for lower Gedrosian.

The Caucasus, Gedrosia, and Arab components seem to work well enough. This Caucasus component is 53% in Georgians, 48% in Abkhasians, so it's light on ANE, so it picked up some of the ENF left from the disintegrated South Indian component.

In the geographical sort, I ignored Gedrosian for anyone west of Haryana since it was just an ANE sink.

EDIT: I don't know what that high Arab score for my mom's cousin represents. It's higher than noise, and it's higher than the South/West Indian ENF boost. It would look like a mix with Baloch were it not for the higher ASI (even discounting the Papuan, both indicate a direction towards India, not Balochistan).

Chad Rohlfsen
05-23-2015, 02:28 AM
Here's a K10. As you see, EHG doesn't really work for Yamnaya. I need to use something else to make it click. Unfortunately, we don't have that pop, so I'd have to use a surrogate. Some type of West Asian, without known Yamnaya ancestry, that doesn't overlap in Europe much. Syrians worked alright. I'm wondering about Iranian Jews.

https://docs.google.com/spreadsheets/d/1RrgONw8NnjckKtn_0oYWmDyuvykYZdtrhW7jNoD4kmE/edit#gid=861672766

Chad Rohlfsen
05-23-2015, 03:06 AM
Any suggestions?

Krefter
05-23-2015, 04:01 AM
Here's a K10. As you see, EHG doesn't really work for Yamnaya. I need to use something else to make it click. Unfortunately, we don't have that pop, so I'd have to use a surrogate. Some type of West Asian, without known Yamnaya ancestry, that doesn't overlap in Europe much. Syrians worked alright. I'm wondering about Iranian Jews.

https://docs.google.com/spreadsheets/d/1RrgONw8NnjckKtn_0oYWmDyuvykYZdtrhW7jNoD4kmE/edit#gid=861672766

Do you mean SW Asian. Because Yamnaya's scoring 80% in EHG.

I have two quick question. Is this K10 supervised? And is it easy to separate SW Asian from EEF? I see in Davidski's and your tests they're separated.

Chad Rohlfsen
05-23-2015, 04:07 AM
I mean that I need extra, eastern stuff, outside of EHG, so, in a way, yes. The problem is finding a good proxy that won't overtake European EEF. Syrians, Georgians, Armenians, all failed to pull Yamnaya out of full EHG, so I'd have to look towards someone like the Kalash, or whatever.

Chad Rohlfsen
05-23-2015, 04:21 AM
I may give the Balochi a shot. They lack the EEF, so they shouldn't mess with Europeans too much.

Krefter
05-23-2015, 05:26 AM
I mean that I need extra, eastern stuff, outside of EHG, so, in a way, yes. The problem is finding a good proxy that won't overtake European EEF. Syrians, Georgians, Armenians, all failed to pull Yamnaya out of full EHG, so I'd have to look towards someone like the Kalash, or whatever.

In ANE K8 Yamnaya doesn't do well with any modern West Asians. They all score too much ENF. The non-EHG ancestors of Yamnya would probably score almost as much ANE in ANE K8 as ENF. I doubt there's a perfect proxy living today for Yamnya's West Asian ancestors.

Chad Rohlfsen
05-23-2015, 05:51 AM
I may just remove Yamnaya from the spreadsheet and see what happens. I'm trying it with the Balochi, first. Also, a run with an Ust_Ishim pop, just to see how it plays out.

tamilgangster
05-23-2015, 08:44 AM
I added both of my parents and it looks clearly like some of my Uralic went into East-Siberian, which would explain the discrepancy in East European scores between us in here and in other calculators, and also the large Arctic percentage observed for myself in other calculators.

Also, Cushitic, since it's from Ethiopia, could be a part of ASI like Nilo-Omotic almost definitely is (latter is high in Onge). It could be like a basal Eurasian part of ASI or something.

EDIT: Also despite representing probably distinct signals, Tibeto-Burman and Austronesian are getting mixed back and forth even though "Total ASI" isn't fluctuating as much (my grandmother and mother are an example, they should otherwise be identical... her parents were first cousins, and her father is related to my dad and her cousin).

EDIT #2: A significant proportion of Uralic is also from ASI. This chart shows the relationship too:

http://i.imgur.com/HHG76CV.png

EDIT #3: The Uralic-EastSiberian thing happened to HRP0370 too.

I believe uralic score is excess ANE and the Cushitic score is from SW asian type ENF not ASE

tamilgangster
05-23-2015, 08:48 AM
Thanks McNinja! My Austronesian looked pretty high, which was expected. I do tend to get relatively high African scores in these runs...

Edit: Thanks for explaining what some of the African components represent.

The austronesian component also includes Austroasiatic, there was a migration of austroasiatics 30000 BP from the northeast of India which predated dravidians. Its similar to the SE asian component on harappa DNA

tamilgangster
05-23-2015, 09:14 AM
Do you mean SW Asian. Because Yamnaya's scoring 80% in EHG.

I have two quick question. Is this K10 supervised? And is it easy to separate SW Asian from EEF? I see in Davidski's and your tests they're separated.

The name EEF is misleadning it should be called Mediteranian or sardinian. THe EHG component is even found among Paniyars who are tribals. I believe that the EHG found among them is related to an archaic migration from a population rich in ANE or from some Ust Ishim related group. This archaic EHG migration is probably what account for Haplgroup r1A1

Krefter
05-23-2015, 10:33 AM
The name EEF is misleadning it should be called Mediteranian or sardinian.

The terminology is hard to keep track of, I agree. But EEF makes sense because it's an abbreviation for: Early European Farmer.

No one except maybe Sardinians are like 100% of this type nowadays. So, Mediterranean would be misleading, because Spanish and Italians aren't 100% EEF, they have been influenced by East Europe and the Middle East since the Neolithic.

Dr_McNinja
05-23-2015, 01:56 PM
I used an Ust-Ishim proxy pop based on allele frequencies in K7 and it pretty much works like Oceanian in K8 in the few I tried (South Indian Brahmin, Haryana Jatt, Punjabi Jatt Sikh, and me) in that everyone got Ust-Ishim %s equivalent (a little less than) to their K8 Oceanian numbers except for me where I got 8-9%. It was also elevated in Afghans (HRP0370) at 6%. It's interesting. Chad, can you post the results with Ust-Ishim if it works?

Also, a Gedrosian component is probably what you need for Yamnaya. Something high in ANE and ENF with a touch of ASE.

EDIT: It was only at 2.x% for Parasar, so it isn't tracking Oceanian everywhere I suppose or Parasar's Oceanian is boosted by East Eurasian. It's at 8.x% for me, 7.x% in HRP0350 Punjabi Jatt, 6.8x% in Haryana Jatt, 6.8x% in Afghan Pashtun, 2-3% in HRP0341, and 4.x% in Soulblighter.

Dr_McNinja
05-23-2015, 02:01 PM
Do you suppose just adding 5 to 10% East Eurasian to Yamnaya would make a good proxy for Indo-European Eastern steppe? (Central Asia)

Kale
05-23-2015, 02:08 PM
I think just removing a bit of EHG would probably do the trick. I mean the other half of Yamnaya had to come from somewhere.

Chad Rohlfsen
05-23-2015, 05:00 PM
South Asians become about 80% Ust Ishim, this is expected with the mix of Near East, West and East Eurasian. It's about 5-10% in West Asia and 1-5% in Europe. I took out Bedouins and just left an EEF component. It is making very interesting results. I'll have some more to post, this evening.

Edit: I'll bet that removing the EEF component, and making it all about Ust_Ishim, plus derived E and W Eurasians would be interesting. I'd bet on 10-20% SSA in SW Asians.

Kale
05-23-2015, 05:55 PM
Hey Chad, an Abstract just posted on Eurogenes might be of interest to you.

450 diverse high coverage whole genome sequences reveal ancient population admixture in modern human populations

Luca Pagani

"Finally, ChromoPainter (Lawson et al. 2012) and MSMC (Schiffels and Durbin 2014) have cemented genetic evidence of an early African origin for the people currently inhabiting Papua New Guinea. Our results are compatible with a first migration out of Africa of these Oceanian populations, which subsequently experienced 80% of gene flow from populations coming from the second, main Eurasian out of Africa."

Chad Rohlfsen
05-24-2015, 01:46 AM
http://www.ncbi.nlm.nih.gov/m/pubmed/25994871/

Dr_McNinja
05-24-2015, 03:36 PM
South Asians become about 80% Ust Ishim, this is expected with the mix of Near East, West and East Eurasian. It's about 5-10% in West Asia and 1-5% in Europe. I took out Bedouins and just left an EEF component. It is making very interesting results. I'll have some more to post, this evening.

Edit: I'll bet that removing the EEF component, and making it all about Ust_Ishim, plus derived E and W Eurasians would be interesting. I'd bet on 10-20% SSA in SW Asians.

How about having ASE/ASI/Onge/etc and Ust-Ishim? Basically Eurogenes K7/K8 with Ust-Ishim, with and without Oceanian?

Chad Rohlfsen
05-24-2015, 03:43 PM
I can make the Paniya a component, but that will likely remove any Onge from everyone, as they're the highest. That could remove any Ust_Ishim, as well. I can't really create an ASI, with where it's at. I'll post more results as they come, and we can all decide what looks best.

Dr_McNinja
05-24-2015, 04:06 PM
I can make the Paniya a component, but that will likely remove any Onge from everyone, as they're the highest. That could remove any Ust_Ishim, as well. I can't really create an ASI, with where it's at. I'll post more results as they come, and we can all decide what looks best.

I think trying to get an Onge proxy for ASI would be a better priority than trying to implement Ust-Ishim for now, if possible.

Chad Rohlfsen
05-24-2015, 04:08 PM
Okay. I'll have that soon. Are you interested in a sort of ANI proxy, like Syrians, or just go straight Bedouin, with EHG, MA1? Something interesting that I see is that when using Syrians, South Asians, even Kalash, are under 10% EHG. I wonder if the ancient and more basal like appearance of MA1 caused inflated northern scores. People like Brahui and Balochi go 80% Syrian, and Kalash are like 70%, with 10% EHG, the rest goes to Austroasiatic, Siberian, Onge, and Papuan.

Dr_McNinja
05-24-2015, 04:21 PM
Okay. I'll have that soon. Are you interested in a sort of ANI proxy, like Syrians, or just go straight Bedouin, with EHG, MA1? Something interesting that I see is that when using Syrians, South Asians, even Kalash, are under 10% EHG. I wonder if the ancient and more basal like appearance of MA1 caused inflated northern scores. People like Brahui and Balochi go 80% Syrian, and Kalash are like 70%, with 10% EHG, the rest goes to Austroasiatic, Siberian, Onge, and Papuan.I think something's up with that since I don't see any way South Asians could be 70-80% Near Eastern. The EHG can make sense though. But that means all the ANE is going into Syrian.

I think Bedouin, EHG (or just leave it out if needed for now), WHG, MA1, East Eurasian, African, Onge, Oceanian would be a good way to start, so we can compare the Onge to Eurogenes' ASE and then try to make some educated guesses on where to go next.

Chad Rohlfsen
05-24-2015, 05:14 PM
I think something's up with that since I don't see any way South Asians could be 70-80% Near Eastern. The EHG can make sense though. But that means all the ANE is going into Syrian.

I think Bedouin, EHG (or just leave it out if needed for now), WHG, MA1, East Eurasian, African, Onge, Oceanian would be a good way to start, so we can compare the Onge to Eurogenes' ASE and then try to make some educated guesses on where to go next.

Sounds good

Sein
05-24-2015, 05:27 PM
Just as a controlled experiment, it would be nice if you could still try this supervised run.

The supervised ADMIXTURE run would have these components:

Onge

Nganasan

BedouinB

Yamnaya

MA1 (ANE)

EHG

WHG

Mbuti

The populations to be tested would be:

UP Brahmins (if you have them)

UP Chamar (if you have them)

Sindhi

Brahui

Baloch

Pathans

Kalash

Pomiri Tajiks

Yaghnobi Tajiks

Iranians

Dr_McNinja
05-24-2015, 06:16 PM
^ GujaratiA and GujaratiD are pretty good too if you don't have Brahmins/Chamar.

Chad Rohlfsen
05-24-2015, 07:02 PM
^ GujaratiA and GujaratiD are pretty good too if you don't have Brahmins/Chamar.

I have included the Gujarati, Balochi, Brahui, Kalash, Burusho, Paniya, Iranians, Tajiks, and I believe Pathans and a couple others. Of course, I have in a ton of Siberians, Central Asians, East Asians, Europeans, and West Asians. Right now, the components are

WHG
Motala
EHG
MA1
Yamnaya
EEF
Bedouin
Yoruba
Ulchi
Atayal
Nganasan
Onge
Papuan

Chad Rohlfsen
05-24-2015, 07:12 PM
I'd still love to be able to play around with that idea in that Siberian paper of a third Eurasian group, maybe something Ust_Ishim like. I did one with Kostenki, and it wasn't anything like one might expect. Baltic Europeans come out as 80% Kostenki. He made up a majority of the EHG samples too. I wonder what combining Kostenki and Ust Ishim would look like. I'm tempted to try it. Kostenki plus MA1 and Nganasan might make EHG. Loschbour might not hurt either.

Dr_McNinja
05-24-2015, 08:54 PM
I made a modified version of Eurogenes ANE K7 with Ust-Ishim and EHG, I'm posting results here:

https://docs.google.com/spreadsheets/d/1BdR1BbYAijPNtfY02sbQNXZXqqwVOtuwrVjAcd4ljd0/edit#gid=0

I didn't use any actual reference samples, I just modified the allele frequencies from the original components (which keeps these derived components behaving). So I don't know if this is worth anything..

EHG and WHG look close to what they should be for South Asians but it's hard to say what EHG "should be" for South Asians to begin with. They're affected by the instability of the base K7 calculator with WHG/ENF. But these WHG numbers are in the ballpark of K8 WHG. So it could mean K7 WHG was being boosted by EHG and something archaic (which was then alleviated with K8 Oceanian).

I have to do another run with EHG and no Ust-Ishim.

I don't know what Ust-Ishim represents here, except that it looks like it belongs both with ANE and with the HG. Moreso with ANE. It peaks in Nepal Brahmin and myself, and is high across the Himalayan/Hindu Kush area as far as Iran (NK19191's ANE+Ust-Ishim here is 18.92%, his K7 ANE is 18.95%). Lower ANE+Ust-Ishim correlates to geographical distance from the Himalayan/Hindu-Kush area. For example, the Afghans and Sapporo have more ENF which indicates a southwesternly direction (towards Balochistan or the Mideast). Parasar's is low because there's genuine East Eurasian and ENF admixture (Everest doesn't have the ENF).

Ideally I'd run this alongside Oceanian but the K8 calculator files are not public. If someone can add Oceanian to K7 and try to tweak it the same way David did, I could run this in there. One possibility is this would make Oceanian peak where we'd expect it to, along with ASE, and remove the peak in the north/northwest, but that's just a guess.

Chad Rohlfsen
05-24-2015, 10:10 PM
Another thing that I've noticed with using Ust_Ishim, is that the Sub-Saharan in the Paniya disappears. I'm not keen on Dravidian history. Do they have real SSA in the 2-3% range?

Kale
05-24-2015, 11:17 PM
I'd still love to be able to play around with that idea in that Siberian paper of a third Eurasian group, maybe something Ust_Ishim like. I did one with Kostenki, and it wasn't anything like one might expect. Baltic Europeans come out as 80% Kostenki. He made up a majority of the EHG samples too. I wonder what combining Kostenki and Ust Ishim would look like. I'm tempted to try it. Kostenki plus MA1 and Nganasan might make EHG. Loschbour might not hurt either.

The other 20% predominantly East Eurasian with a tiny slice of East African?

Chad Rohlfsen
05-24-2015, 11:26 PM
McNinja,
You may have something there, brother. Look at Ust_Ishim looking to peak in Central Asians. Take out that gene flow from Turks and Mongols, the Tajiks would be 15%+ Ust Ishim. Do you have anyone from the Caucasus? I wonder if it is Ust Ishim like folks making this teal component, and that is why we can't get a good proxy. I bet if I'd mix something farmer, with EHG and Ust_Ishim, I could get that teal stuff. The only thing that will need changing is that ENF. My Near Eastern is probably double that 18%, and WHG should be about half of that.

Chad Rohlfsen
05-24-2015, 11:38 PM
Let me see how this run looks, and I may add Ust_Ishim.

tamilgangster
05-25-2015, 12:24 AM
The terminology is hard to keep track of, I agree. But EEF makes sense because it's an abbreviation for: Early European Farmer.

No one except maybe Sardinians are like 100% of this type nowadays. So, Mediterranean would be misleading, because Spanish and Italians aren't 100% EEF, they have been influenced by East Europe and the Middle East since the Neolithic.

Even sardianians are around 70% EEF and french spanish and mainland italians are around 50% this type. THis component is also found in high levels among Maghrebis and Levantines. There are two types of ENF from this, A bedouin type and a Mediteranian type. THe Bedouin type possibly might have basal eurasian affinities.

tamilgangster
05-25-2015, 12:49 AM
Do you suppose just adding 5 to 10% East Eurasian to Yamnaya would make a good proxy for Indo-European Eastern steppe? (Central Asia)

There is no evidence of indoeuropeans carrying Any east eurasian genes, no indoeuropean populations have east eurasian affininites(unless its from admixture with local population eg bengalis)

Chad Rohlfsen
05-25-2015, 01:10 AM
There is no evidence of indoeuropeans carrying Any east eurasian genes, no indoeuropean populations have east eurasian affininites(unless its from admixture with local population eg bengalis)

EHG's look like they carry about 10% Amerindian like admixture. The Karelian looks like it has a little more ENA than the Samaran. MtDNA C is floating around from something. It's not an early addition to Europe, prior to the E and W Eurasian split, as it isn't found in Western or Northern Europeans. It's restricted to Karelia and the Steppes. I think it's more likely that some Native American like groups migrated west, than all moving to the Americas. It's quite possible that people from the Yeniseian Valley to the tip of South America were about identical until about 6-8kya.

Remember IR1 shows up as 8-9% Nganasan.

tamilgangster
05-25-2015, 01:26 AM
South Asians become about 80% Ust Ishim, this is expected with the mix of Near East, West and East Eurasian. It's about 5-10% in West Asia and 1-5% in Europe. I took out Bedouins and just left an EEF component. It is making very interesting results. I'll have some more to post, this evening.

Edit: I'll bet that removing the EEF component, and making it all about Ust_Ishim, plus derived E and W Eurasians would be interesting. I'd bet on 10-20% SSA in SW Asians.

Thats very misleading, it just appears that way because both groups appear to be very mixed genetically.You should create a test simply using Ust Ishim Basal Eurasian and Onge along with SSA and possibly an east eurasian sample. You cant have a test using both archaic and modern components, its best to have a test soley using archaic populations

tamilgangster
05-25-2015, 01:29 AM
I can make the Paniya a component, but that will likely remove any Onge from everyone, as they're the highest. That could remove any Ust_Ishim, as well. I can't really create an ASI, with where it's at. I'll post more results as they come, and we can all decide what looks best.

Making the Paniya a component would be redundent

tamilgangster
05-25-2015, 01:36 AM
I think something's up with that since I don't see any way South Asians could be 70-80% Near Eastern. The EHG can make sense though. But that means all the ANE is going into Syrian.

I think Bedouin, EHG (or just leave it out if needed for now), WHG, MA1, East Eurasian, African, Onge, Oceanian would be a good way to start, so we can compare the Onge to Eurogenes' ASE and then try to make some educated guesses on where to go next.

THis is perfect just get rid of EHG

tamilgangster
05-25-2015, 01:57 AM
EHG's look like they carry about 10% Amerindian like admixture. The Karelian looks like it has a little more ENA than the Samaran. MtDNA C is floating around from something. It's not an early addition to Europe, prior to the E and W Eurasian split, as it isn't found in Western or Northern Europeans. It's restricted to Karelia and the Steppes. I think it's more likely that some Native American like groups migrated west, than all moving to the Americas. It's quite possible that people from the Yeniseian Valley to the tip of South America were about identical until about 6-8kya.

Remember IR1 shows up as 8-9% Nganasan.

EHG contain Amerindian like admixture because its most likely related to the ANE branch the mixed with east eurasians to produce amerindians. Also it cant be east eurasian related if its not found among east asian populations.

Chad Rohlfsen
05-25-2015, 02:11 AM
EHG contain Amerindian like admixture because its most likely related to the ANE branch the mixed with east eurasians to produce amerindians. Also it cant be east eurasian related if its not found among east asian populations.

If what's not found among East Eurasians? Amerindian admixture is in Siberia. Not every individual crossed Beringia and some back-migrated.

J Man
05-25-2015, 02:14 AM
If what's not found among East Eurasians? Amerindian admixture is in Siberia. Not every individual crossed Beringia and some back-migrated.

Yes...The Koryak people have some evidence of back migration from the Americas in their genes I think.

Chad Rohlfsen
05-25-2015, 02:43 AM
Yes...The Koryak people have some evidence of back migration from the Americas in their genes I think.

Chukchis, for sure.

Dr_McNinja
05-25-2015, 02:53 AM
I added a Swedish individual (highest yet) and Caucasus individual.

One of the issues here is that WHG in K7 and in other Admixture runs becomes inflated/drifted relative to ENF. So even Yamnaya becomes overwhelmingly similar to Karelia/Samara HG. You'd need to do whatever it is David did to make his West Eurasia K8. If I had that calculator, I could make the proportions better.

I tried it with only Ust-Ishim or only EHG and the results are mostly the same, off by 1 or 2%. For example, with only Ust-Ishim I had 8.93% of it and with only EHG I had 6.02%, but together I had 8.11% and 5.27%. The components are very stable.

Ust-Ishim doesn't look Central Asia, but seems to peak in South Central Asia and extends across Central Asia and South Asia and near Iran. It doesn't seem to extend into the North Caucasus but is there in the Swedish individual. I don't know what it could represent other than some kind of South Asian-specific archaic ANE signal (besides Ust-Ishim being literally an archaic sort of ANE I guess...). It could indicate that some of the earliest ANE populations moved south?

J Man
05-25-2015, 03:19 AM
Chukchis, for sure.

Makes sense.

Krefter
05-25-2015, 04:59 AM
I added a Swedish individual (highest yet) and Caucasus individual.

One of the issues here is that WHG in K7 and in other Admixture runs becomes inflated/drifted relative to ENF. So even Yamnaya becomes overwhelmingly similar to Karelia/Samara HG. You'd need to do whatever it is David did to make his West Eurasia K8. If I had that calculator, I could make the proportions better.

I tried it with only Ust-Ishim or only EHG and the results are mostly the same, off by 1 or 2%. For example, with only Ust-Ishim I had 8.93% of it and with only EHG I had 6.02%, but together I had 8.11% and 5.27%. The components are very stable.

Ust-Ishim doesn't look Central Asia, but seems to peak in South Central Asia and extends across Central Asia and South Asia and near Iran. It doesn't seem to extend into the North Caucasus but is there in the Swedish individual. I don't know what it could represent other than some kind of South Asian-specific archaic ANE signal (besides Ust-Ishim being literally an archaic sort of ANE I guess...). It could indicate that some of the earliest ANE populations moved south?

Isn't Ust-Ishim too old to have specific affinities to any Eurasians today? I would think Ust-Ishim can only reveal details about the early branching of Eurasians(like the two independent Basal branchs in the ancient Near East and Kostenki) before his time. But I shouldn't assume he doesn't have any specific relation to certain modern Eurasians, we have to let him speak for himself.

MA1 shares ancestry with K14 and WHG that Ust-Ishim lacks. So, unless MA1 descends from two(or more) Eurasian branches Ust-Ishim can't be anymore connected to ANE than to East Asians.

Chad Rohlfsen
05-25-2015, 06:02 AM
We need Central Asians/South Siberians from 8-10kya. There could be something there.

tamilgangster
05-25-2015, 06:47 AM
What would be great if is there was a DNA sample from balangoda man

Kale
05-25-2015, 01:22 PM
MA1 shares ancestry with K14 and WHG that Ust-Ishim lacks. So, unless MA1 descends from two(or more) Eurasian branches Ust-Ishim can't be anymore connected to ANE than to East Asians.

I'm not seeing any particular problem with that. If Ust-Ishim's group branches out into ANE and East Asians, and then West Eurasians mix with ANE to make MA1 and the like.

I don't think Ust-Ishim is equidistant to East and West Eurasians because he is the ancestor of both, but rather because he is so minimally derived from the root of East and West Eurasians (probably in the East direction) as to not be statistically significant.

Kurd
05-25-2015, 01:24 PM
I added a Swedish individual (highest yet) and Caucasus individual.

One of the issues here is that WHG in K7 and in other Admixture runs becomes inflated/drifted relative to ENF. So even Yamnaya becomes overwhelmingly similar to Karelia/Samara HG. You'd need to do whatever it is David did to make his West Eurasia K8. If I had that calculator, I could make the proportions better.

I tried it with only Ust-Ishim or only EHG and the results are mostly the same, off by 1 or 2%. For example, with only Ust-Ishim I had 8.93% of it and with only EHG I had 6.02%, but together I had 8.11% and 5.27%. The components are very stable.

Ust-Ishim doesn't look Central Asia, but seems to peak in South Central Asia and extends across Central Asia and South Asia and near Iran. It doesn't seem to extend into the North Caucasus but is there in the Swedish individual. I don't know what it could represent other than some kind of South Asian-specific archaic ANE signal (besides Ust-Ishim being literally an archaic sort of ANE I guess...). It could indicate that some of the earliest ANE populations moved south?

The WHG/UHG in Davidski's run is not total WHG, but excess WHG above and beyond the EEF proxy Stuttgart's ~40% WHG type stuff http://bga101.blogspot.com/2014/09/eurogenes-ane-k7.html

"Indeed, a really important thing to understand about the Lazaridis et al. study is that it doesn't actually attempt to estimate overall WHG-UHG ancestry in Europeans, but rather the excess WHG-UHG on top of what is already present in the EEF proxy Stuttgart"

"Western European/Unknown Hunter-Gatherer (WHG-UHG): this essentially looks like a West Eurasian forager component, and includes the forager-like stuff carried by Neolithic farmers (Oetzi the Iceman has 40% of it)."

I believe that Chad's run should include a proxy (Bedouin) that captures basal Eurasian (Stuttgart is only partially basal Eurasian)

Anabasis
05-25-2015, 02:06 PM
I added a Swedish individual (highest yet) and Caucasus individual.

One of the issues here is that WHG in K7 and in other Admixture runs becomes inflated/drifted relative to ENF. So even Yamnaya becomes overwhelmingly similar to Karelia/Samara HG. You'd need to do whatever it is David did to make his West Eurasia K8. If I had that calculator, I could make the proportions better.

I tried it with only Ust-Ishim or only EHG and the results are mostly the same, off by 1 or 2%. For example, with only Ust-Ishim I had 8.93% of it and with only EHG I had 6.02%, but together I had 8.11% and 5.27%. The components are very stable.

Ust-Ishim doesn't look Central Asia, but seems to peak in South Central Asia and extends across Central Asia and South Asia and near Iran. It doesn't seem to extend into the North Caucasus but is there in the Swedish individual. I don't know what it could represent other than some kind of South Asian-specific archaic ANE signal (besides Ust-Ishim being literally an archaic sort of ANE I guess...). It could indicate that some of the earliest ANE populations moved south?

It seems you add mine or Zephs results as caucuas individual. I would like to remind you something that while mine WHG scores 8.25% in Eugenes k9, i score WHG 1,6% Euroasian K8

Here my k8 results.

ANE 0.151945
South_Eurasian 0.013973
Near_Eastern 0.815005
East_Eurasian 1E-005
WHG 0.015921
Oceanian 1E-005
Pygmy 0.003126
Sub-Saharan 1E-005

I think David optimimised WHG in K8.

Chad Rohlfsen
05-25-2015, 02:12 PM
Bedouins aren't that much more basal than EEF. The distance is greater due to south Eurasian and African admixture.

Dr_McNinja
05-25-2015, 02:19 PM
I think David optimimised WHG in K8.Yes, this is based off K7 where WHG was inflated. K8 was not made public.

Anabasis
05-25-2015, 02:32 PM
We need Central Asians/South Siberians from 8-10kya. There could be something there.

Some of them on way

http://infobaikal.ru/news/s179/n127173/

https://www.google.com/maps/d/viewer?mid=zPTFaw2rnx-E.kfoCf5XA8Lgw

Kurd
05-25-2015, 02:40 PM
I suggested Bedouin, because a few samples are needed to sort out basal Eurasian allele frequencies, and think Bedouin would be helpful in that regard, short of ancient samples

Dr_McNinja
05-25-2015, 02:58 PM
I think David used some ancient samples to sort out ENF from WHG in K8:

http://eurogenes.blogspot.ca/2014/12/ane-is-primary-cause-of-west-to-east.html


an improved version of my ANE K7 ancestry test. Let's call it the West Eurasia K8. This one gives more accurate estimates of Western European Hunter-Gatherer (WHG) and Near Eastern admixture proportions, thanks to the use of new ancient samples.Some ANE does boost WHG in K7. Loschbour is 99% WHG in K8 but only 88% in K7 IIRC, with 8% in ANE. Likewise, our ANE scores went up from K7 to K8 as WHG went down.

Dr_McNinja
05-25-2015, 03:01 PM
Btw, Chad, have you noticed weird effects from including African components? I'm working on another calculator and any African components with basal Eurasian affinity screw everything up, even if they're only attracting 1-2% admixture. It causes a huge swing in other unrelated components. But it takes a lot longer to calculate results without the African components.

J Man
05-25-2015, 03:06 PM
It would be nice if true WHG ancestry could be separated from the UHG ancestry at some point in the future. That way we can figure out our true Mesolithic European ancestry. We may need to wait for Mesolithic and Neolithic genomes from the Near East though before this can be done.

Chad Rohlfsen
05-25-2015, 03:33 PM
Yeah, it does make things a little odd. That's why i stick to Yorubas, as they supposedly have the least Eurasian backflow. I can also make West Asians as a mix of EEF, EHG, MA1, and Yoruba, without Bedouin, but South and South Central Asians lose a lot of farmer ancestry.

Chad Rohlfsen
05-25-2015, 04:02 PM
Some of them on way

http://infobaikal.ru/news/s179/n127173/

https://www.google.com/maps/d/viewer?mid=zPTFaw2rnx-E.kfoCf5XA8Lgw

That individual will likely be East Eurasian, without much West Eurasian. I'm thinking more towards the Urals in that timeframe.

Kurd
05-25-2015, 04:15 PM
The creator of the PuntDNA Eurasian K10 calculator, appears to have isolated a component that he calls W Semitic http://www.anthrogenica.com/showthread.php?4531-New-Calculator-puntDNAL/page9 which he says is based on basal Eurasian, and according to his spreadsheet it peaks at about 84% in Bedouin.

In his blog at http://bga101.blogspot.com/, David briefly mentions it:

"However, I suspect that West Eurasia will have to be modeled in a different way from Europe, with, amongst other things, the so called Basal Eurasian component replacing ENF. But for this to happen we'll need at least one ancient genome that is in large-part of Basal Eurasian origin"

Chad Rohlfsen
05-25-2015, 05:55 PM
I don't think I can include many tribals. They keep wanting to jump on either MA1 or Nivkh. I've ditched the Nivkh, to see if that helps. If they grab onto MA1 again, I'll have to drop to just the 4 Paniyas and a couple others, possibly.

Krefter
05-25-2015, 06:20 PM
I'm not seeing any particular problem with that. If Ust-Ishim's group branches out into ANE and East Asians, and then West Eurasians mix with ANE to make MA1 and the like.

I don't think Ust-Ishim is equidistant to East and West Eurasians because he is the ancestor of both, but rather because he is so minimally derived from the root of East and West Eurasians (probably in the East direction) as to not be statistically significant.

There's plenty of prove in formal stats that MA1 is more related to K14 and Loschbour than to East Asians. K14 who lived 14,000 years earlier was already branching out of Crown Eurasian. There's a clade they all descend from. So, MA1 certainly is young enough to have a specific relation to certain Eurasians.

I haven't seen evidence Ust_Ishim has a relation to MA1, beyond being "Crown Eurasian".

Chad Rohlfsen
05-25-2015, 06:30 PM
The creator of the PuntDNA Eurasian K10 calculator, appears to have isolated a component that he calls W Semitic http://www.anthrogenica.com/showthread.php?4531-New-Calculator-puntDNAL/page9 which he says is based on basal Eurasian, and according to his spreadsheet it peaks at about 84% in Bedouin.

In his blog at http://bga101.blogspot.com/, David briefly mentions it:

"However, I suspect that West Eurasia will have to be modeled in a different way from Europe, with, amongst other things, the so called Basal Eurasian component replacing ENF. But for this to happen we'll need at least one ancient genome that is in large-part of Basal Eurasian origin"

That's pretty close to the score of EEF, in Bedouins, when using K02, LBKT, and Stuttgart. The rest is EHG and Yoruba. Probably not a great fit, but it worked and didn't make them as their own component.

Kurd
05-25-2015, 10:35 PM
That's pretty close to the score of EEF, in Bedouins, when using K02, LBKT, and Stuttgart. The rest is EHG and Yoruba. Probably not a great fit, but it worked and didn't make them as their own component.

It kind of did make them their own component if considered in the larger context of the whole spreadsheet, which shows ME pops and Sardinians at 25-35% levels. Stuttgart was only at 30%. Whereas in the latest W Eurasia K8 spreadsheet by David, Stuttgart is at 72%, Sardinians at 65%, and some ME pops >85% ENF.

Chad Rohlfsen
05-25-2015, 10:42 PM
It kind of did make them their own component if considered in the larger context of the whole spreadsheet, which shows ME pops and Sardinians at 25-35% levels. Stuttgart was only at 30%. Whereas in the latest W Eurasia K8 spreadsheet by David, Stuttgart is at 72%, Sardinians at 65%, and some ME pops >85% ENF.

No, sorry. I was referring to the supervised run I did, without a Bedouin component. I'll show it in just a bit.

tamilgangster
05-25-2015, 11:50 PM
The creator of the PuntDNA Eurasian K10 calculator, appears to have isolated a component that he calls W Semitic http://www.anthrogenica.com/showthread.php?4531-New-Calculator-puntDNAL/page9 which he says is based on basal Eurasian, and according to his spreadsheet it peaks at about 84% in Bedouin.

In his blog at http://bga101.blogspot.com/, David briefly mentions it:

"However, I suspect that West Eurasia will have to be modeled in a different way from Europe, with, amongst other things, the so called Basal Eurasian component replacing ENF. But for this to happen we'll need at least one ancient genome that is in large-part of Basal Eurasian origin"

THe West semetic accounts for all OOA populations, Its the sandawe component which accounts for basal eurasian its found in higher ratios among SW asians and SOuth asians

Kale
05-26-2015, 12:56 AM
There's plenty of prove in formal stats that MA1 is more related to K14 and Loschbour than to East Asians.

No disagreement there. Supposing Davidski's extraction of ANE from Karitiana is meaningful in any way, and ff ANE were a third branch, MA1 is still 60% West Eurasian, so it stands he'd be closer to West than East.

Krefter
05-26-2015, 01:07 AM
No disagreement there. Supposing Davidski's extraction of ANE from Karitiana is meaningful in any way, and ff ANE were a third branch, MA1 is still 60% West Eurasian, so it stands he'd be closer to West than East.

What are you suggesting the other 40% is?

Kale
05-26-2015, 01:09 AM
What are you suggesting the other 40% is?

Perhaps a homogenized blend of West Eurasian and a third branch...which shows up as ANE.

Kind of weird that in all the ancient samples, and in all the modern peaks...its always 40%.

Karitiana 40%
MA1 40%
Karelia_HG 40%
Samara_HG 40%
North India/Central Asia peak 40%

Krefter
05-26-2015, 01:47 AM
Perhaps a homogenized blend of West Eurasian and a third branch...which shows up as ANE.

Kind of weird that in all the ancient samples, and in all the modern peaks...its always 40%.

Karitiana 40%
MA1 40%
Karelia_HG 40%
Samara_HG 40%
North India/Central Asia peak 40%

Is this from an ADMIXTURE test you created?

tamilgangster
05-26-2015, 02:19 AM
There's plenty of prove in formal stats that MA1 is more related to K14 and Loschbour than to East Asians. K14 who lived 14,000 years earlier was already branching out of Crown Eurasian. There's a clade they all descend from. So, MA1 certainly is young enough to have a specific relation to certain Eurasians.

I haven't seen evidence Ust_Ishim has a relation to MA1, beyond being "Crown Eurasian".

I believe that Ust Ishim is more related to WHG than it is to ANE.


Interestingly, the Ust’-Ishim individual probably lived during a warm period that has been proposed to be a time of expansion of modern humans into Europe
This individual would then represent an early modern human radiation into Europe and Central Asia that may have failed to leave descendants among present-day population

parasar
05-26-2015, 03:06 AM
I believe that Ust Ishim is more related to WHG than it is to ANE.

A smidgen perhaps, but not with any statistical confidence.
Please see La Brana (WHG) vs Mal'ta (ANE)
http://www.nature.com/nature/journal/v514/n7523/images_article/nature13810-f3.jpg

Ust-Ishim is far closer to Non-Africans than to Africans, and significantly (statistical) closer to Other non-Africans than to Present-day Europeans.

Krefter
05-26-2015, 03:17 AM
A smidgen perhaps, but not with any statistical confidence.
Please see La Brana (WHG) vs Mal'ta (ANE)
http://www.nature.com/nature/journal/v514/n7523/images_article/nature13810-f3.jpg

Ust-Ishim is far closer to Non-Africans than to Africans, and significantly (statistical) closer to Other non-Africans than to Present-day Europeans.

He's a little closer to East Asians and Oceanians(who both share ancestry, beyond what they share with MA1, K14, WHG) than to MA1 and La Brana-1. Maybe he's related to the ancestors(or ancestral) of East Asians and Oceanians.

Kurd
05-26-2015, 04:10 AM
http://i.imgur.com/Mq3zBUC.jpg

Europeans & W Asians derive some ancestry from Basal Eurasians which places them further from Ust-Ishim than E Eurasians, who is located near the Eastern non African/ "Old' WHG node.

The flowchart also shows Stuttgart to have around 44% Basal Eurasian ancestry, who according to David's W Eurasia K8 spreadsheet is around 72% ENF. This would suggest that Bedouin like populations who are close to 100% ENF should be around 61% Basal Eurasian.

parasar
05-26-2015, 04:25 AM
Hey Chad, an Abstract just posted on Eurogenes might be of interest to you.

450 diverse high coverage whole genome sequences reveal ancient population admixture in modern human populations

Luca Pagani

"Finally, ChromoPainter (Lawson et al. 2012) and MSMC (Schiffels and Durbin 2014) have cemented genetic evidence of an early African origin for the people currently inhabiting Papua New Guinea. Our results are compatible with a first migration out of Africa of these Oceanian populations, which subsequently experienced 80% of gene flow from populations coming from the second, main Eurasian out of Africa."

"significantly higher amounts of Neanderthal gene flows in Island South East Asian and Oceanian populations"
Do we know if this came with the second - 80% - or the first, or both?

Megalophias
05-26-2015, 04:35 AM
I believe that Ust Ishim is more related to WHG than it is to ANE.

Why?

D MA1, La Brana; Ust_Ishim, Yoruba 0.0097, 1.348
D Chimp, Ust_Ishim; Karelia_HG, MA1 0.0088, 1.029
D Chimp, Ust_Ishim; Samara_HG, MA1 0.0041, 0.411
D Mbuti, Ust_Ishim; Karitiana, Loschbour 0.0008, 0.128
D Mbuti, Ust_Ishim; Ket, Loschbour 0.0022, 0.0488

Generalissimo
05-26-2015, 05:15 AM
Those K7 and K8 files are available now...

Link (http://eurogenes.blogspot.com.au/2015/05/tracking-asi-on-ancient-steppe.html?showComment=1432371053806#c28183184194 0448544)

puntDNALKing
05-26-2015, 05:36 AM
http://i.imgur.com/Mq3zBUC.jpg

Europeans & W Asians derive some ancestry from Basal Eurasians which places them further from Ust-Ishim than E Eurasians, who is located near the Eastern non African/ "Old' WHG node.

The flowchart also shows Stuttgart to have around 44% Basal Eurasian ancestry, who according to David's W Eurasia K8 spreadsheet is around 72% ENF. This would suggest that Bedouin like populations who are close to 100% ENF should be around 61% Basal Eurasian.

Thanks for letting me know about this interesting thread. I agree with your assessment fully. I used a lot of Bedouin samples in order to obtain a pure Southwest Asian Component. The actual Western_Semitic cluster is very much in line with EEF when compared with modern people, but with ancient DNA, it becomes a Basal Eurasian cluster. Its kind of weird. The only reason I add this Eurasian feature to my calculator was to isolate the non-african component from Horn of Africans.

I ran LBK (F999916), which should have 44% +/- 10% Basal Eurasian. The Western_Semitic Components acts as Basal Eurasian.

Here are the Results:
# Population Percent
1 Indo_European 69.42
2 Western_Semitic 30.58

Then, I ran the La Brana-Arintero (F999915) sample from Leon, Spain, about 7000 years old, represents a pre-agricultural European human genome – in other words, before the agriculturists from the Near East arrived. Notice, there is no Western Semitic component (Basal Eurasian).

Here are the Results:
# Population Percent
1 Indo_European 94.47
2 Sino_Tibetan 3.42
3 Dravidian 1.92

Then ran the Loschbour sample (F999918), with zero Basal Eurasian.
# Population Percent
1 Indo_European 95.25
2 Sino_Tibetan 3.03
3 Amerindian 1.21

Finally I ran the Mal’ta sample, from Siberia, about 24,000 years of age. Although Native Americans are consider to be closer to East Asians, my results as well as the MA-1 mitochondrial genome belongs to haplogroup U, which has also been found at high frequency among Upper Palaeolithic and Mesolithic European hunter-gatherers10, 11, 12, and the Y chromosome of MA-1 is basal to modern-day western Eurasians and near the root of most Native American lineages.

Here are the Results:

# Population Percent
1 Indo_European 57.89
2 Amerindian 19.08
3 Dravidian 18.31
4 Sino_Tibetan 3.03
5 Niger_Congo 1.69

tamilgangster
05-26-2015, 09:11 AM
He's a little closer to East Asians and Oceanians(who both share ancestry, beyond what they share with MA1, K14, WHG) than to MA1 and La Brana-1. Maybe he's related to the ancestors(or ancestral) of East Asians and Oceanians.

Ancestral to east asians yes, oceanians no. It makes more sense than east asians being descended from onge. In the paper it says that hes more closley related to east asians than to onge. They also say its a distinct crown eurasian lineage which is neither west or east eurasian, it is possible that SOME of the east eurasian lineanage is derived from Ust Ishim.

tamilgangster
05-26-2015, 09:36 AM
http://i.imgur.com/Mq3zBUC.jpg

Europeans & W Asians derive some ancestry from Basal Eurasians which places them further from Ust-Ishim than E Eurasians, who is located near the Eastern non African/ "Old' WHG node.

The flowchart also shows Stuttgart to have around 44% Basal Eurasian ancestry, who according to David's W Eurasia K8 spreadsheet is around 72% ENF. This would suggest that Bedouin like populations who are close to 100% ENF should be around 61% Basal Eurasian.

Europeans do most likely derive some of their ancestry from a population related to ust ishim, but they did not leave any pure blood descendents. But you are correct its most likely the basal eurasian that distances them from Ust Ishim.

Kale
05-26-2015, 01:14 PM
They also say its a distinct crown eurasian lineage which is neither west or east eurasian, it is possible that SOME of the east eurasian lineanage is derived from Ust Ishim.

They say that because in their modeling they don't take into account the minor East Eurasian admixture that exists in their West Eurasian samples (Malta, Loschbour...and as a result Stuttgart). They considered it as an alternative explanation, but did not wish to explore further with the data they had.

MbutiPygmy Ust'-Ishim Kostenki14 Dai 0.0006 0.096
MbutiPygmy Ust'-Ishim Kostenki14 Loschbour 0.0004 0.055
Kostenki has nothing basal to Ishim, none of the basal Eurasian found in Stuttgart
MbutiPygmy Ust'-Ishim Kostenki14 Stuttgart -0.0252 -3.503

Yet we see this...
Chimp Dai Loschbour Kostenki14 -0.0187 -3.034 16025 15438 329585
Chimp Japanese Loschbour Kostenki14 -0.0211 -3.488 16065 15401 329585
A sharing between East Eurasians and Loschbour...

Chimp Dai MA1 Loschbour -0.0074 -1.11 12233 12052 251027
Chimp Han MA1 Loschbour -0.005 -0.752 12187 12066 251027
MA1 sharing no significantly different amount.

Chad Rohlfsen
05-26-2015, 02:18 PM
I'm getting closer to a good k run. Actually, pretty damn good. I'll try to have it fine tuned and up in the next couple of days.

parasar
05-26-2015, 02:38 PM
Europeans do most likely derive some of their ancestry from a population related to ust ishim, but they did not leave any pure blood descendents. But you are correct its most likely the basal eurasian that distances them from Ust Ishim.

No doubt from what we know as of now.

"There is 100% bootstrap support for Ust’-Ishim being in the same clade as Papuan, Australian, Karitiana, Mixe, Han and Dai ... TreeMix continues to place the Ust’-Ishim individual in the same clade as Papuan, Australian, Karitiana, Mixe, Han and Dai with 100% bootstrap support ...
"When all populations from Panel A and B are included, we see that Ust-Ishim is "on the eastern non-African lineage."

"The finding that the Ust’-Ishim individual is equally closely related to present-day Asians and to 8,000-to 24,000-year-old individuals from western Eurasia, but not to present-day Europeans, is compatible with the hypothesis that present-day Europeans derive some of their ancestry from a population that did not participate in the initial dispersals of modern humans into Europe and Asia11"*[referring to Lazaridis et al.]

*"In a different, geographical, sense the category “West Eurasian” could be transferred to the “basal Eurasian” element instead, as it is the only one whose presence we can detect only in West Eurasia, while the common ancestry of both MA1 and Loschbour with eastern non-Africans (drift non_African-->X) raises the alternative possibility of an eastern sojourn of their ancestors and a temporal priority of “basal Eurasians” in western Eurasia."

Dr_McNinja
05-26-2015, 04:31 PM
Chad, how would you describe Onge in terms of admixture? How much Southeast Asian/ASE, how much general East Eurasian, how much Papuan, etc?

And how would you describe Papuan in terms of K7/K8 components (minus Oceanian)? How much Southeast Asian, general or north East Asian, basal Eurasian, WHG, African, etc?

Chad Rohlfsen
05-26-2015, 09:06 PM
Chad, how would you describe Onge in terms of admixture? How much Southeast Asian/ASE, how much general East Eurasian, how much Papuan, etc?

And how would you describe Papuan in terms of K7/K8 components (minus Oceanian)? How much Southeast Asian, general or north East Asian, basal Eurasian, WHG, African, etc?

I haven't attacked them yet. I have them as their own component. In unsupervised runs they either get full on Papuan or top the South Asian.

Chad Rohlfsen
05-26-2015, 10:06 PM
I'm tempted to try that Basal as Ena model, where I'd remove EEF and Bedouins as a component.

tamilgangster
05-27-2015, 01:30 AM
wrong thread

Sein
05-27-2015, 03:42 AM
Chad, how would you describe Onge in terms of admixture? How much Southeast Asian/ASE, how much general East Eurasian, how much Papuan, etc?

And how would you describe Papuan in terms of K7/K8 components (minus Oceanian)? How much Southeast Asian, general or north East Asian, basal Eurasian, WHG, African, etc?

For what it's worth, the Onge can be quite easily characterized, using formal stats.

Kale posted this:

"Chimp Onge Loschbour Japanese 0.0586 12.384 16786 18874 351075
Chimp Kostenki14 Japanese Onge -0.0045 -1.106 15836 15693 332566
Chimp Loschbour Japanese Onge -0.0057 -1.466 16978 16786 351075

Onge significantly prefer East Eurasian over West Eurasian, and West Eurasian has no significant preference between East Eurasian and Onge."

Basically, Onge are very close to East Asians, and ancient West Eurasian samples are equally related to East Asians and Onge.

parasar
05-27-2015, 04:01 AM
..

Basically, Onge are very close to East Asians, and ancient West Eurasian samples are equally related to East Asians and Onge.

The same can be said for Ust-Ishim and most likely for Tianyuan* too, and it makes complete sense.

*Edit
Tianyuan's relationship to select modern Africans, E. Asians, Papuans, Amerindians, and Europeans.
The closest modern population is the Karitiana. The slightly higher difference the Papuan is likely due to the elevated Denisovan in Papuans.
http://4.bp.blogspot.com/-Ygeyjg0KKjM/UQA-xfvF89I/AAAAAAAABjo/RY19NiL3ETg/s1600/TianyuanPairwise.png

Dr_McNinja
05-27-2015, 02:27 PM
Going from the K7/K8 breakdown of other calculator components and various individuals, it looks like there's some East Eurasian that is correlated to ASE whenever a South Asian component is formed. However, that is lacking in the Baloch component itself which only has ASE and not East Eurasian. It's also lacking in other ancient populations or components which don't have that ASE-entangled East Eurasian.

Going by general K7 results, it looks like it should be around a 90% ASE, 10% East Eurasian composition. Hitting as high as 92% in individuals with the least East Eurasian (Punjabi Arain, Punjabi Gujjar, Baloch, some Sindhis).

Not addressing this admixture results in it pulling ASE away from a South Asian component and becoming "Southeast Asian". Malay are around 50-60% ASE I think, highest in a standalone population, and the HAP South Indian component has 70-72% ASE (out of ASE+East Eurasian total). That made me think Onge would be around 70 ASE as well. More than 60 at least, otherwise the Malay would be more ASE than the Onge.

But that would mean the source of ASE in a lot of India was not Onge or an Onge-like population because there just isn't enough East Eurasian.

So how's this picture look?

Northeast Asian: 3.71% ANE, 11.52% ASE, 0.06% WHG, 83.86% East Eurasian, 0.21% West African, 0.32% East African, 0.31% ENF

Himalayan/Indo-Tibetan?: 30% ASE, 70% East Eurasian?

Southeast Asian: 1.79% ANE, 50.28% ASE, 1.70% WHG, 43.57% East Eurasian, 0.13% West African, 0.61% East African, 1.93% ENF

ASI?: 70% ASE, 30% East Eurasian?

A confounding factor is archaic admixture. I've only been able to isolate Ust-Ishim which I suspect is basically just tying up the archaic signal seen previously in Oceanian with some WHG/ANE/ENF. We can see hints of this archaic signal perhaps in Southeast Asian above. It should be high in Himalayan/Indo-Tibetan as well, but we have no idea what it would be like in ASI.

Kale
05-27-2015, 02:31 PM
The same can be said for Ust-Ishim and most likely for Tianyuan* too, and it makes complete sense.

*Edit
Tianyuan's relationship to select modern Africans, E. Asians, Papuans, Amerindians, and Europeans.
The closest modern population is the Karitiana. The slightly higher difference the Papuan is likely due to the elevated Denisovan in Papuans.
http://4.bp.blogspot.com/-Ygeyjg0KKjM/UQA-xfvF89I/AAAAAAAABjo/RY19NiL3ETg/s1600/TianyuanPairwise.png

Wow. That's surprising. It's a small sample for comparison, but I think what shows there would actually put Tianyuan closer to West Eurasians than East Eurasians. Han and Dai are the same distance from Tianyuan as French and Sardinian. And that is WITH the modern Euros containing Basal Eurasian, which should drag them further away. Karitiana is closer than the other East Eurasians...at 40kya I don't think there would be enough divergence in the East Eurasian clade for Tianyuan to have any preference among them (except maybe against Papuans or Onge). So the only other thing that could potentially bring Karitiana closer is the ~40% West Eurasian ancestry.

parasar
05-27-2015, 02:39 PM
...

Not addressing this admixture results in it pulling ASE away from a South Asian component and becoming "Southeast Asian". Malay are around 50-60% ASE I think, highest in a standalone population, and the HAP South Indian component has 70-72% ASE (out of ASE+East Eurasian total). That made me think Onge would be around 70 ASE as well. More than 60 at least, otherwise the Malay would be more ASE than the Onge.

But that would mean the source of ASE in a lot of India was not Onge or an Onge-like population because there just isn't enough East Eurasian.

...

A confounding factor is archaic admixture. I've only been able to isolate Ust-Ishim which I suspect is basically just tying up the archaic signal seen previously in Oceanian with some WHG/ANE/ENF. We can see hints of this archaic signal perhaps in Southeast Asian above. It should be high in Himalayan/Indo-Tibetan as well, but we have no idea what it would be like in ASI.

Definitely not the Ongee. Ongee like, perhaps, but I think limited to Austro Asiatics. If you recall Reich used the Santhal as an ASI proxy, and the Santhal would not even fit on his Indian cline.

parasar
05-27-2015, 02:52 PM
Wow. That's surprising. It's a small sample for comparison, but I think what shows there would actually put Tianyuan closer to West Eurasians than East Eurasians. Han and Dai are the same distance from Tianyuan as French and Sardinian. And that is WITH the modern Euros containing Basal Eurasian, which should drag them further away. Karitiana is closer than the other East Eurasians...at 40kya I don't think there would be enough divergence in the East Eurasian clade for Tianyuan to have any preference among them (except maybe against Papuans or Onge). So the only other thing that could potentially bring Karitiana closer is the ~40% West Eurasian ancestry.

I think if we remove the Basal portion from the French and Sardinians they will be as close to Tianyuan as the Karitiana are.
Remember Tianyuan is ~40000 years old and is only about 300 generations closer to the ENA side splitting point (42500ybp into East Asian, ASI, Ongee, ANE-WHG) than to Basal. A 50000 year old human specimen will be identically close to almost all modern non-Africans (if we eliminate archaic admixture form Denisovans and Neandertals).

Kale
05-27-2015, 04:24 PM
Well we have Ust-Ishim at 45,000 years, who shows effectively neutral Eurasian affinity, but is arguably an early East Eurasian.
And at 37,000 years we have Kostenki which is unambiguously West Eurasian, and possibly even derived in a WHG direction, as opposed to ANE.

At 40,000 years we should have definitive, although still small, deviation towards one branch (but no subbranch). The small sample available from Tianyuan might not produce much of a result...but it is still worth looking into. I'd like to see a direct comparison of Tianyuan to say Kostenki, Loschbour, and MA1.

Chad Rohlfsen
05-27-2015, 04:53 PM
Doing some unsupervised runs too. Onge did get broken up at k5 with some near eastern. I'll post k5-10 this afternoon.

Dr_McNinja
05-27-2015, 05:37 PM
So, here's a makeshift calculator I'm tweaking:

HRP0349:

0.87% SE-Asian
0.11% NE-Asian
0.30% Indo-Tibetan
10.57% ASI
7.84% Western-Steppe
7.89% Eastern-Steppe
66.81% Gedrosian
2.27% Caucasus
0.00% Mideast
0.02% Europe
0.34% West-African
2.98% East-African

HRP0393 (Haryana Jatt):

0.57% SE-Asian
0.04% NE-Asian
0.15% Indo-Tibetan
9.22% ASI
24.44% Western-Steppe
3.48% Eastern-Steppe
55.55% Gedrosian
3.48% Caucasus
0.02% Mideast
0.16% Europe
0.00% West-African
2.88% East-African

HRP0341:

0.52% SE-Asian
0.04% NE-Asian
0.15% Indo-Tibetan
8.14% ASI
10.73% Western-Steppe
3.86% Eastern-Steppe
70.29% Gedrosian
3.57% Caucasus
0.01% Mideast
0.05% Europe
0.00% West-African
2.64% East-African

East-African is like basal Eurasian (it's from Eurogenes K7).

There's a "DIY" South Asian component, just multiply ASI by 3, then subtract the difference from Gedrosian. Since HAP S-Indian is basically a version of ASI mixed with Oceanian/East Eurasian mixed with, literally, HAP Baloch. The ratios even add up to make it unmistakable (31-35 ANE, rest ENF).

So I'd have:

0.87% SE-Asian
0.11% NE-Asian
0.30% Indo-Tibetan
10.57% ASI
7.84% Western-Steppe
7.89% Eastern-Steppe
66.81% Gedrosian
2.27% Caucasus
0.00% Mideast
0.02% Europe
0.34% West-African
2.98% East-African
(31.71% South Asian)
(45.67% Gedrosian)

and HRP0393 would have,

0.57% SE-Asian
0.04% NE-Asian
0.15% Indo-Tibetan
9.22% ASI
24.44% Western-Steppe
3.48% Eastern-Steppe
55.55% Gedrosian
3.48% Caucasus
0.02% Mideast
0.16% Europe
0.00% West-African
2.88% East-African
(27.66% South Asian)
(37.11% Gedrosian)

and HRP0341 would have:

0.52% SE-Asian
0.04% NE-Asian
0.15% Indo-Tibetan
8.14% ASI
10.73% Western-Steppe
3.86% Eastern-Steppe
70.29% Gedrosian
3.57% Caucasus
0.01% Mideast
0.05% Europe
0.00% West-African
2.64% East-African
(24.42% South Asian)
(54.01% Gedrosian)

Or you could combine ASI with Gedrosian and East-African in whatever proportions you'd like. It's not technically a proper admixture calculator that way since there's a bunch of ASE in Gedrosian which would then get pulled into a South Asian component that incorporated ANE/ENF, but it's probably just as accurate a method (in terms of application to reality).

I'm gonna tinker with it some more.

EDIT: Here's soulblighter (South Indian Brahmin),

Soulblighter

0.76% SE-Asian
0.12% NE-Asian
0.31% Indo-Tibetan
14.94% ASI
0.07% Western-Steppe
9.42% Eastern-Steppe
69.91% Gedrosian
0.23% Caucasus
0.00% Mideast
0.00% Europe
0.05% West-African
4.19% East-African

Dr_McNinja
05-27-2015, 05:41 PM
Doing some unsupervised runs too. Inge did get broken up at k5 with some near eastern. I'll post k5-10 this afternoon.

I got "ASI" to be something like 89-90% K7 ASE and 3-4% ENF and 3-4% WHG (probably all ENF), and 2-3% East Eurasian. So it could mean the East Eurasian influx was later, but early in its history. I kept it as 90% ASE, 10% East Eurasian though for the above runs.

Sapporo
05-27-2015, 05:41 PM
@Dr_McNinja

Interesting calculator. Curious as to why you and my Gedrosia are 67-70% while HRP0393 is only 55% though. Why is his Western Steppe 2.5-3x ours? I'd expect it to be higher but not by that amount.

Dr_McNinja
05-27-2015, 05:47 PM
Western Steppe is Yamnaya, almost exactly. Eastern Steppe is a mix of Western Steppe and what populations living there (between Kazakhstan/Mongolia/Siberia, or western Mongolia, NW-China nexus) could resemble. It was based off an average of Uyghur, Khakass, Tatar, Mongolia, Selkup, Ket, and Karitiana.

You're meant to average out Western Steppe (Yamnaya being the far left fringe of the actual Western steppe) and Eastern Steppe to get a location somewhere on the vast expanse of the Central Asian steppe.

Caucasus is probably not real Caucasus. Combined with East-African, it would become Middle Eastern.

The only reason we're getting any Caucasus in other calculators is because the South Asian component has Gedrosian in it. Otherwise it would mix with some of the ASE in the South Asian components to go into Gedrosian.

EDIT: Multiplying ASI by 3.25 or so gives closer to South Asian scores in other calculators. Sapporo clearly has greater Gedrosian affinity in every calculator run I've done compared to his HarappaWorld results. There's still 9% ASE in Gedrosian or thereabouts.

Dr_McNinja
05-27-2015, 05:50 PM
@Dr_McNinja

Interesting calculator. Curious as to why you and my Gedrosia are 67-70% while HRP0393 is only 55% though. Why is his Western Steppe 2.5-3x ours? I'd expect it to be higher but not by that amount.

His WHG is more than double ours and Western Steppe is especially WHG-limited. Yamnaya had 34-35% ANE, 4-5% ASE, and 35% WHG or so, so it strongly sucks up admixture from ANE/WHG-heavy people (i.e, EHG). I actually compensated for the WHG bias in K7 otherwise he'd be at well over 30% Western Steppe.

We should all be Eastern Steppe (representing the source of our original ANE) but Western Steppe could represent a more recent (Scythian-era) wave from the northwest.

EDIT: Western Steppe competes for ANE with Gedrosian which Northeastern Europe components never did (15% ANE or so). Being that WHG in South Asia likely came from the Steppe, accompanied by a lot of ANE, this is a more probable reflection of actual ancestry than anything with European components.

parasar
05-27-2015, 06:07 PM
Well we have Ust-Ishim at 45,000 years, who shows effectively neutral Eurasian affinity, but is arguably an early East Eurasian.
And at 37,000 years we have Kostenki which is unambiguously West Eurasian, and possibly even derived in a WHG direction, as opposed to ANE.

At 40,000 years we should have definitive, although still small, deviation towards one branch (but no subbranch). The small sample available from Tianyuan might not produce much of a result...but it is still worth looking into. I'd like to see a direct comparison of Tianyuan to say Kostenki, Loschbour, and MA1.

Yes that comparison would be useful.

Ust-Ishim is ~5000 years down from Basal and ~2500 above the ENA split. And since we not entitling Basal as West Eurasian, but ANE and WHG, I would say that that Ust-Ishim leans non-Basal rather that East Eurasian.

I think Tianyuan would be slightly closer to Loschbour than to MA1, and more distant to Kostenki 14 who has Basal.

The Fu paper put Tianyuan a bit on the non-Basal side - "The nuclear DNA sequences determined from this early modern human reveal that the Tianyuan individual derived from a population that was ancestral to many present-day Asians and Native Americans but postdated the divergence of Asians from Europeans" but ANE and WHG cross-over portions from non-Basal to Basal were not resolved.

Later Rasmussen et al. who had MA1 and Tianyuan, had a different take based on shared genetic history with Anzick-1:
http://csfa.tamu.edu/cfsa-publications/Rasmussen%20et%20al%202014%20-%20anzick%20genome.pdf
http://www.nature.com/nature/journal/v506/n7487/extref/nature13025-s1.pdf
"Tianyuan individual, which at 40,000-years-old seems equally related to a geographically wide range of Eurasian populations in this analysis (Extended Data Figure 5f and 5g)."
http://www.nature.com/nature/journal/v506/n7487/fig_tab/nature13025_SF5.html

Chad Rohlfsen
05-27-2015, 07:59 PM
I got "ASI" to be something like 89-90% K7 ASE and 3-4% ENF and 3-4% WHG (probably all ENF), and 2-3% East Eurasian. So it could mean the East Eurasian influx was later, but early in its history. I kept it as 90% ASE, 10% East Eurasian though for the above runs.

Let me check the unsupervised runs. If something hits about 64-65% in the Paniya, it may well be ASI.

Chad Rohlfsen
05-27-2015, 08:56 PM
I don't get WHG in South Asia, but I do get EHG in supervised runs. I think that WHG is compensating for the difference between EHG and ma1, but doesn't reflect any real WHG in EHG or South Asia.

Dr_McNinja
05-27-2015, 09:08 PM
Also, since it's based on Eurogenes K7, that calculator uses 230k SNPs.


I think that WHG is compensating for the difference between EHG and ma1, but doesn't reflect any real WHG in EHG or South Asia.It depends on how you want to classify admixture in the Steppe (the presumed source of this admixture in South Asia). As either 50% EHG or 35% WHG for example. EHG is the better fit to our knowledge but calculators seem to like WHG.

Chad Rohlfsen
05-27-2015, 09:54 PM
Unsupervised K5, where the Onge break up.

https://docs.google.com/spreadsheets/d/1XbeCJ8FNbEfXawXgvu58JAKMwteZ5DiHEGAuHtrdQLk/edit?usp=sharing

Chad Rohlfsen
05-27-2015, 09:57 PM
Unsupervised K6, Onge still broken up.

https://docs.google.com/spreadsheets/d/1cFjlwG4HhTO21CVxpyjQBqL8ByTUal_-Y3TwKhiJ4ks/edit?usp=sharing

Chad Rohlfsen
05-27-2015, 10:08 PM
Waiting on K7, but here is K8. We have an Onge component, that has been dropped about 3-4% for the Paniya. So, SC Asian must be about 15-20% ASI.

https://docs.google.com/spreadsheets/d/1o-Ah_g-44pTLsEfh5ObUbzNrhoHUBJI1Z8LecilliEc/edit?usp=sharing

Chad Rohlfsen
05-27-2015, 10:16 PM
K7 is about done. K9 and 10 may be a little while.

Dr_McNinja
05-27-2015, 10:25 PM
Can you run Onge through K7 and/or ask David to run them through K8? Perhaps they'll break down to 30/60 ASE:Oceanian in the latter too.

Chad Rohlfsen
05-27-2015, 10:31 PM
Can you run Onge through K7 and/or ask David to run them through K8? Perhaps they'll break down to 30/60 ASE:Oceanian in the latter too.

K7 will be done here in the next few minutes. These are unsupervised, so I have no idea what the Onge will do. I'll post it after it's done. If one looks pretty good, I may throw in the other 150+ samples I have of South Asians and Malays.

Chad Rohlfsen
05-27-2015, 10:48 PM
K7, ASI looks like it is about 10% Near East, and the rest is Onge, Papuan, some Siberian and Austronesian like.

https://docs.google.com/spreadsheets/d/1ljkxdB3IAcU-Nu-cQjvn3jPo8dgME32Z4knde8ClLgE/edit#gid=244244658

Chad Rohlfsen
05-27-2015, 11:53 PM
After K9 and K10, everyone feel free to chime in on what you'd like to see. Personally, I'm interested to see how things behave if I throw in the rest of the South Asians. Especially, the K7. It might get rid of that split between the three SE Asians, or at least clean it up.