PDA

View Full Version : Post Your EEF, WHG and ANE Admixture Proportions



Pages : [1] 2

ZephyrousMandaru
12-28-2013, 10:33 AM
New calculator created by Day Tripper.

How to use it.

1. Download this spreadsheet.

https://app.box.com/s/nr09lj4xbgivblr4tpm5

2. Compute your ADMIXTURE Proportions Using Eurogenes K13.

3. Input your percentages in each corresponding labeled columns (Besides the "You").

My Results rounded to the nearest percent.

EEF 101%
WHG -7.61%
ANE 6.6%

http://bga101.blogspot.com.au/2013/12/eef-whg-ane-test-for-europeans.html

DMXX
12-28-2013, 10:47 AM
My results:

EEF- 84.8%
WHG - 4.8%
ANE - 10.4%

Humanist
12-28-2013, 10:53 AM
My results:


99.44__EEF
-6.45__WHG
7.02___ANE


Results for the two Iraqi Mandaeans:


98.71__EEF
-5.85__WHG
7.14___ANE


101.33__EEF
-8.12___WHG
6.79____ANE

Anglecynn
12-28-2013, 11:05 AM
Posted it in another thread first:

Makes sense really. My grandfather is the least Neolithic and my mother the most. My grandmother and aunt are closest to the average.

Me:
EEF 50.74727444
WHG 34.53641099
ANE 14.71631456

Father:
EEF 48.71762975
WHG 36.31198132
ANE 14.97038893

Grandfather:
EEF 47.17876269
WHG 37.38628388
ANE 15.43495343

Grandmother:
EEF 49.18717416
WHG 35.99146358
ANE 14.82136226

Mother:
EEF 51.5446288
WHG 33.80660099
ANE 14.64877021

Aunt:
EEF 50.04961238
WHG 35.29891573
ANE 14.65147188

Sein
12-28-2013, 11:06 AM
This calculator is somewhat inapplicable to me, since 5%-15% of my genetic ancestry is derived from an ancient population of South Eurasians, and they aren't accounted for in this model. Nevertheless, I'm predominantly West Eurasian, and my largest component for K13 is "West Asian", so I thought it might be fun to see what this makes of me.

Results:
EEF 59.45%
WHG 24.20%
ANE 16.33%

I'm tempted to treat my "ANE" percentage as capturing all of my non-West Eurasian admixture, since I'm consistently 4%-6% Siberian-Amerindian on most calculators, along with the 10% "ASI". But that wouldn't be accurate, since I'm 9.39% Baltic, and 1.53% North_Atlantic.

DMXX
12-28-2013, 11:15 AM
I'm tempted to treat my "ANE" percentage as capturing all of my non-West Eurasian admixture, since I'm consistently 4%-6% Siberian-Amerindian on most calculators, along with the 10% "ASI". But that wouldn't be accurate, since I'm 9.39% Baltic, and 1.53% North_Atlantic.

I've been treating ANE as a surrogate for North Euro components outside of Europe as they seemed tightly correlated in the original study. Even in my case, my Baltic+N. Atlantic combination (11.4%) seems to match ANE (10.4%) quite well.

Perhaps your non-West Eurasian ancestry is distorting the results to begin with? This calculator isn't built for us non-European West Eurasians anyway. The negative scores the Assyrians and Mandeans tells us that much.

Humanist
12-28-2013, 11:18 AM
Average values for the two Assyrians, and two Iraqi Mandaeans:


xxx ASY MAN
EEF 100.22 100.02
WHG -7.03 -6.99
ANE 6.81 6.97

Sein
12-28-2013, 11:28 AM
I've been treating ANE as a surrogate for North Euro components outside of Europe as they seemed tightly correlated in the original study. Even in my case, my Baltic+N. Atlantic combination (11.4%) seems to match ANE (10.4%) quite well.

Perhaps your non-West Eurasian ancestry is distorting the results to begin with? This calculator isn't built for us non-European West Eurasians anyway. The negative scores the Assyrians and Mandeans tells us that much.

A very good point. I'm thinking this is really geared towards individuals who fit into the main cline of modern European genetic variation.

ZephyrousMandaru
12-28-2013, 11:30 AM
My exact percentage for EEF, WHG and ANE are as follows.

EEF: 101.06%

WHG: -7.60

ANE: 6.53%

Mirror mirror on the wall, who's the farmest of them all? Well, so far, it's Iraqi Mandaean 2 at 101.33%.

MfA
12-28-2013, 11:49 AM
Here's mine and Anatolian Kurdish average(n=10)


EEF: 91.07108
WHG: -0.23218
ANE: 9.161104

Mean Range
EEF: 89.891057 (85.42166 95.15498)
WHG: 0.6981958 (-3.39783 4.37971)
ANE: 9.4272976 (8.242842 10.48169)

ZephyrousMandaru
12-28-2013, 11:54 AM
Here's mine and Anatolian Kurdish average(n=10)


EEF: 91.07108
WHG: -0.23218
ANE: 9.161104

Mean Range
EEF: 89.891057 (85.42166 95.15498)
WHG: 0.6981958 (-3.39783 4.37971)
ANE: 9.4272976 (8.242842 10.48169)

You think you can create an MDS Plot based on this?

DMXX
12-28-2013, 11:58 AM
Here's mine and Anatolian Kurdish average(n=10)


EEF: 91.07108
WHG: -0.23218
ANE: 9.161104

Mean Range
EEF: 89.891057 (85.42166 95.15498)
WHG: 0.6981958 (-3.39783 4.37971)
ANE: 9.4272976 (8.242842 10.48169)

This is useful for me. Note the lowest EEF, highest WHG and highest ANE scores in the Anatolian Kurdish range are identical to my results. Looks like I'm (unsurprisingly) in the same genetic zone as northern Kurds.

MfA
12-28-2013, 12:01 PM
It is obsolate for West Asians


EEF WHG ANE
Abhkasian 87,02415 6,748984 6,22687
Chechen 73,14048 13,40275 13,45677
Cypriot 100,9694 -6,16872 5,199286
Druze 102,9161 -8,33524 5,419118
Lezgin 83,62527 9,769327 6,605407


from the paper
http://abload.de/img/desktop_2013_12_27_22nef0i.png


You think you can create an MDS Plot based on this?

I think I can, need time though.

Humanist
12-28-2013, 12:09 PM
xxx ENG ASY MAN IRN KRD PST
EEF 50.7 100.2 100.0 84.8 89.9 59.5
WHG 34.5 -7.0 -7.0 4.8 0.7 24.2
ANE 14.7 6.8 7.0 10.4 9.4 16.3

N= 1 2 2 1 10 1

ENG = English
ASY = Assyrian
MAN = Iraqi Mandaean
IRN = Iranian
KRD = Anatolian Kurd
PST = Pashtun

Anglecynn
12-28-2013, 12:15 PM
xxx ENG ASY MAN IRN KRD PST
EEF 50.7 100.2 100.0 84.8 89.9 59.5
WHG 34.5 -7.0 -7.0 4.8 0.7 24.2
ANE 14.7 6.8 7.0 10.4 9.4 16.3

N= 1 2 2 1 10 1

ENG = English
ASY = Assyrian
MAN = Iraqi Mandaean
IRN = Iranian
KRD = Anatolian Kurd
PST = Pashtun

Hmm, seems my mother is the only one who is clearly more Farmer, everyone else in my family apart from myself seems to be less, although all the figures are in the right ballpark so it seems like a good job.

newtoboard
12-28-2013, 12:29 PM
I've been treating ANE as a surrogate for North Euro components outside of Europe as they seemed tightly correlated in the original study. Even in my case, my Baltic+N. Atlantic combination (11.4%) seems to match ANE (10.4%) quite well.

Perhaps your non-West Eurasian ancestry is distorting the results to begin with? This calculator isn't built for us non-European West Eurasians anyway. The negative scores the Assyrians and Mandeans tells us that much.
So what does WHG correlate to for non Europeans?

Hanna
12-28-2013, 12:30 PM
Me:

EEF 91.4258530593
WHG -0.1698049135
ANE 8.7439518542

My aunt:

EEF 95.7477159116
WHG -3.6731744332
ANE 7.9254585216

Hanna
12-28-2013, 12:33 PM
Here's mine and Anatolian Kurdish average(n=10)
EEF: 91.07108
WHG: -0.23218
ANE: 9.161104

We have similar results.

DMXX
12-28-2013, 12:45 PM
So what does WHG correlate to for non Europeans?

Going by the limited results we have, it's hard to say. Sein is the only non-European with a significant amount of it. At present, a reasonable guess is it represents a generalised "catch all" estimation of non-Near-Eastern, but still West Eurasian, hunter-gatherer stock. But I have very little confidence in that, given the small sample size, the calculator's non-specificity outside Europe, and the fact that those three components (EEF, WHG and ANE) were deduced from European samples exclusively.

rms2
12-28-2013, 01:00 PM
Here are my results:

EEF 46.49814568
WHG 38.03949482
ANE 15.4623595

That was fun. Thanks!

newtoboard
12-28-2013, 01:15 PM
Going by the limited results we have, it's hard to say. Sein is the only non-European with a significant amount of it. At present, a reasonable guess is it represents a generalised "catch all" estimation of non-Near-Eastern, but still West Eurasian, hunter-gatherer stock. But I have very little confidence in that, given the small sample size, the calculator's non-specificity outside Europe, and the fact that those three components (EEF, WHG and ANE) were deduced from European samples exclusively.
Maybe it represents a portion of Gedrosia which is more Northern like in More eastern regions and possibly linked to pre Neolithic Central Asian ancestry. I know David has proposed something similar about Gedrosia being more northern.

rms2
12-28-2013, 01:22 PM
It seems Anglecynn's grandfather and I have profiles that are pretty close:

Anglecynn's Grandfather:
EEF 47.17876269
WHG 37.38628388
ANE 15.43495343

Me:
EEF 46.49814568
WHG 38.03949482
ANE 15.4623595

MJost
12-28-2013, 02:02 PM
Results M093609
EEF 46.28923483
WHG 38.08808574
ANE 15.62267943

MJost

Hanna
12-28-2013, 02:05 PM
My grandfather:

EEF: 93.9816723878
WHG: -2.5059076158
ANE: 8.524235228

rms2
12-28-2013, 02:09 PM
Results M093609
EEF 46.28923483
WHG 38.08808574
ANE 15.62267943

MJost

Whoa, Mark! We're not too far off! You, me, and Anglecynn's grandfather: same tribe? ;)

Solothurn
12-28-2013, 02:25 PM
Here are mine and mother's based on our K13 Gedmatch results. We are British with all known recent British ancestry :)

Mine
46.82% EEF (Early European Farmer)
37.84% WHG (West European Hunter-Gatherer)
15.35% ANE (Ancient North Eurasian)

Mother's
47.37% EEF (Early European Farmer)
37.43% WHG (West European Hunter-Gatherer)
15.20% ANE (Ancient North Eurasian)

They are probably in the British 'norm' :)

* Slightly off topic, but I ran my paternal cousin's and got

46.59
37.73
15.68

On PF at FTDNA Both myself and cousin are 96% Orcs and 4% ME, whereas my mum is 100% Orc!

MJost
12-28-2013, 02:28 PM
Results M093609
EEF 46.28923483
WHG 38.08808574
ANE 15.62267943

MJost

Here is my half maternal sister M093357 who has a Norwegian paternal admix and my half maternal brother M104509 (HG-U106) (a NW German Jost) paternal

Results M093357
EEF 44.58216245
WHG 39.41561543
ANE 16.00222212

Results M104509
EEF 65.06077319
WHG 25.00169437
ANE 9.937532437




MJost

DMXX
12-28-2013, 02:30 PM
Here are mine and mother's based on our K13 Gedmatch results. We are British with all known recent British ancestry :)

Mine
46.82% EEF (Early European Farmer)
37.84% WHG (West European Hunter-Gatherer)
15.35% ANE (Ancient North Eurasian)

Mother's
47.37% EEF (Early European Farmer)
37.43% WHG (West European Hunter-Gatherer)
15.20% ANE (Ancient North Eurasian)

You and your mother's scores are, of course, very similar (barely differing by 1% each component). Although I couldn't help notice your EEF is lower while WHG and ANE are higher when compared with your mother. Going by the general pattern inferred when the study was discussed earlier on this forum, is your father's known ancestry from somewhere a bit more northern and/or eastern in Britain compared to your mother's, by any chance?

MJost
12-28-2013, 02:34 PM
Whoa, Mark! We're not too far off! You, me, and Anglecynn's grandfather: same tribe? ;)

Hey I was just thinking something is similar about us when I posted my data.

MJost

rms2
12-28-2013, 02:41 PM
Here is my half maternal sister M093357 who has a Norwegian paternal admix and my half maternal brother M104509 (HG-U106) (a NW German Jost) paternal

Results M093357
EEF 44.58216245
WHG 39.41561543
ANE 16.00222212

Results M104509
EEF 65.06077319
WHG 25.00169437
ANE 9.937532437




MJost

Your half maternal brother's result is interesting, especially in contrast to yours, mine, Anglecynn's and his family's, and solothurn's (all of which are British). The continental German (your half brother) has much higher EEF and much lower WHG and ANE scores than we do.

Looking at the German Language Area Y-DNA Project, it looks like the percentage of Neolithic y lineages is much higher than in the Isles.

Il Papŕ
12-28-2013, 02:45 PM
EEF 61,04695597
WHG 27,02233076
ANE 11,93071327

R.Rocca
12-28-2013, 03:01 PM
Seems that for people of European ancestry, the calculator is spot on.

I am 5/8 Italian (mostly southern) and 3/8 Spanish (mostly northern)

EEF: 78.02
WHG: 13.59
ANE: 8.39

Not surprisingly, my closest matches are with Spaniards and Tuscans:

Spanish:
EEF: 80.9
WHG: 6.8
ANE: 12.3

Tuscan:
EEF: 74.6
WHG: 13.6
ANE: 11.8

Sapporo
12-28-2013, 03:02 PM
I'm not Euro but decided to try it out for fun:

Results
EEF 52.16553515
WHG 30.15350707
ANE 17.68095779


CALCS
One: -23.2504558 52.37535471
Two: 15.9937849 30.27479013
Three: 40.5735816 17.75207392

Solothurn
12-28-2013, 03:04 PM
Unfortunately I don't have my father's results to compare, but his ancestry is NE compared to my mother's amazingly:)

I do have his cousin which shows more of what you stated!

45.09 EEF
39.02 WHG
15.89 ANE





You and your mother's scores are, of course, very similar (barely differing by 1% each component). Although I couldn't help notice your EEF is lower while WHG and ANE are higher when compared with your mother. Going by the general pattern inferred when the study was discussed earlier on this forum, is your father's known ancestry from somewhere a bit more northern and/or eastern in Britain compared to your mother's, by any chance?

Anglecynn
12-28-2013, 03:12 PM
Here is my half maternal sister M093357 who has a Norwegian paternal admix and my half maternal brother M104509 (HG-U106) (a NW German Jost) paternal

Results M093357
EEF 44.58216245
WHG 39.41561543
ANE 16.00222212

Results M104509
EEF 65.06077319
WHG 25.00169437
ANE 9.937532437




MJost

Hmm the results of M104509 are strange, in most calculators he's pretty similar to other people who have around 50% EEF but in this he is much, much more EEF.

DMXX
12-28-2013, 03:16 PM
Unfortunately I don't have my father's results to compare, but his ancestry is NE compared to my mother's amazingly:)

I do have his cousin which shows more of what you stated!


Not the first time I've pinpointed someone's ancestry like that. I should leave my career path and become a genetic shaman...

Calamus
12-28-2013, 03:16 PM
My result:
41.24 EEF
41.9 WHG
16.86 ANE

Very close to the Norwegian average (41.1 EEF, 42.8 WHG, 16.1 ANE)

NK19191
12-28-2013, 03:20 PM
My Results

Results

EEF 84.9
WHG 4.46
ANE 10.62

DMXX
12-28-2013, 03:21 PM
My Results

Results

EEF 84.9
WHG 4.46
ANE 10.62

Once again, our results are practically the same. :) It seems N-NW Iranians are on the more "northern" end of things compared to Kurmanji Kurds, referring again to the range MfA posted.

Solothurn
12-28-2013, 03:39 PM
:beerchug:

Do you have the British average?


Not the first time I've pinpointed someone's ancestry like that. I should leave my career path and become a genetic shaman...

soulblighter
12-28-2013, 03:46 PM
Here are my results:

EEF 47.45681494
WHG 34.75186065
ANE 17.7913244

My K13 results are in another post, but just for completeness:



North_Atlantic Baltic West_Med West_Asian East_Med Red_Sea South_Asian East_Asian Siberian Amerindian Oceanian Northeast_African Sub-Saharan
You 4.13 3.6 0 23.03 2.05 1.25 61.13 1.59 1.3 0 0.67 0 1.24

newtoboard
12-28-2013, 04:05 PM
Here are my results:

EEF 47.45681494
WHG 34.75186065
ANE 17.7913244

My K13 results are in another post, but just for completeness:



North_Atlantic Baltic West_Med West_Asian East_Med Red_Sea South_Asian East_Asian Siberian Amerindian Oceanian Northeast_African Sub-Saharan
You 4.13 3.6 0 23.03 2.05 1.25 61.13 1.59 1.3 0 0.67 0 1.24




I had a feeling WHG was actually somewhat ASI for non europeans when Sein first posted his results. Thanks for confirming it.

soulblighter
12-28-2013, 04:33 PM
I had a feeling WHG was actually somewhat ASI for non europeans when Sein first posted his results. Thanks for confirming it.

I think this is just testing for finnish/estonian ancestry as proxy for ANE in Europeans.
For non-Europeans, I have the largest ANE, and so does not seem like a test for any "northern" component?

newtoboard
12-28-2013, 04:39 PM
I think this is just testing for finnish/estonian ancestry as proxy for ANE in Europeans.
For non-Europeans, I have the largest ANE, and so does not seem like a test for any "northern euro" component?

How about ANE as a measure of Northern European + East Eurasian in your case?

soulblighter
12-28-2013, 04:57 PM
How about ANE as a measure of Northern European + East Eurasian in your case?

Shouldn't Sein have the highest ANE then?

parasar
12-28-2013, 04:59 PM
I think this is just testing for finnish/estonian ancestry as proxy for ANE in Europeans.
For non-Europeans, I have the largest ANE, and so does not seem like a test for any "northern" component?

That is because you are from India, and likely in the past from Northern India.

Your proportions are essentially identical to mine:
EEF 47.99684431
WHG 34.19761276
ANE 17.80554294

Hanna
12-28-2013, 05:55 PM
So how accurate is this for Middle Easterners because we seem to plot far away from EEF on the plot:

http://3.bp.blogspot.com/-fc2W-6tR-HA/Urigqts3hwI/AAAAAAAAJbg/hqZiV1TOGgc/s1600/europe.png

rms2
12-28-2013, 06:06 PM
Seems that for people of European ancestry, the calculator is spot on.

I am 5/8 Italian (mostly southern) and 3/8 Spanish (mostly northern)

EEF: 78.02
WHG: 13.59
ANE: 8.39

Not surprisingly, my closest matches are with Spaniards and Tuscans:

Spanish:
EEF: 80.9
WHG: 6.8
ANE: 12.3

Tuscan:
EEF: 74.6
WHG: 13.6
ANE: 11.8

At the risk of sounding ignorant, where did you get the averages for Spaniards and Tuscans, and is the same thing available for British Isles populations?

Update: Never mind. I found them here: http://img24.imageshack.us/img24/2240/q1is.png.

vettor
12-28-2013, 06:07 PM
my results below

EEF 69.21
WHG 20.20
ANE 10.59

AJL
12-28-2013, 06:15 PM
You can see why I end up in the Austrian Alps:

EEF 68.64%
WHG 20.26%
ANE 11.10%

My Ashkenazi grandfather is 89% EEF while my north Euro mother is only 50%.

vettor
12-28-2013, 06:20 PM
You can see why I end up in the Austrian Alps:

EEF 68.64%
WHG 20.26%
ANE 11.10%

My Ashkenazi grandfather is 89% EEF while my north Euro mother is only 50%.

we seem to have very close results to each other , but I have only a max of 1 % of ashkenazi in my chromosomes ( #6 actually)

R.Rocca
12-28-2013, 06:23 PM
At the risk of sounding ignorant, where did you get the averages for Spaniards and Tuscans, and is the same thing available for British Isles populations?

It is available in "Extended Data Table 3" here: http://biorxiv.org/content/biorxiv/suppl/2013/12/23/001552.DC1/001552-1.pdf

They have values for English and Scottish.

vettor
12-28-2013, 06:29 PM
It is available in "Extended Data Table 3" here: http://biorxiv.org/content/biorxiv/suppl/2013/12/23/001552.DC1/001552-1.pdf

They have values for English and Scottish.


thanks

so its safe to say I fit between

Bergamo 0.715 0.177 0.108
French_South 0.675 0.195 0.13


EEF 69.21
WHG 20.20
ANE 10.59

rms2
12-28-2013, 06:32 PM
My results seem to fall closest to the Orcadian and Ukrainian.

Me:
EEF 46.49814568
WHG 38.03949482
ANE 15.4623595

Orcadian:
EEF 45.7
WHG 38.5
ANE 15.8

Ukrainian
EEF 46.2
WHG 38.7
ANE 15.1

evon
12-28-2013, 07:13 PM
My result:
41.24 EEF
41.9 WHG
16.86 ANE

Very close to the Norwegian average (41.1 EEF, 42.8 WHG, 16.1 ANE)

Western Norwegians, with Romani ancestry on my Maternal side, and some German on my Paternal side:

Me:

EEF 43.7884769991
WHG 40.2029194712
ANE 16.0086035297

Maternal Uncle:
EEF 45.1625982483
WHG 39.0140182467
ANE 15.823383505

Maternal Grandmother:
EEF 44.1847904162
WHG 39.6001568675
ANE 16.2150527162

Paternal Aunt:
EEF 43.395219053
WHG 40.5189307459
ANE 16.0858502012

My Guess is the Norwegian Samples are from eastern Norway..

AJL
12-28-2013, 07:13 PM
we seem to have very close results to each other , but I have only a max of 1 % of ashkenazi in my chromosomes ( #6 actually)

Yes I make a fairly convincing Austrian or Slovenian or Swiss person, on average.

AJL
12-28-2013, 07:16 PM
It is available in "Extended Data Table 3" here: http://biorxiv.org/content/biorxiv/suppl/2013/12/23/001552.DC1/001552-1.pdf

They have values for English and Scottish.

Thanks. My mother is almost exactly the English average.

Hok
12-28-2013, 07:28 PM
EEF 47,96296546
WHG 36,697801
ANE 15,33923353

evon
12-28-2013, 07:36 PM
Some Norwegian Romani cousins of ours:

1:
EEF 47.2230628265
WHG 36.9877516678
ANE 15.7891855057

2:
EEF 44.6981794717
WHG 39.3648669787
ANE 15.9369535496

3.
EEF 44.8680493228
WHG 39.1775481926
ANE 15.9544024845

5.
EEF 47.9724514717
WHG 36.5801590945
ANE 15.4473894338

MfA
12-28-2013, 07:47 PM
It is obsolate for West Asians


EEF WHG ANE
Abhkasian 87,02415 6,748984 6,22687
Chechen 73,14048 13,40275 13,45677
Cypriot 100,9694 -6,16872 5,199286
Druze 102,9161 -8,33524 5,419118
Lezgin 83,62527 9,769327 6,605407


from the paper
http://abload.de/img/desktop_2013_12_27_22nef0i.png


Not only West Asians/Non-Europeans but also Iberian(xBasque) scores are quite off

EEF WHG ANE
Spanish_Andalucia 70,40227 20,0858 9,511934
Spanish_Aragon 67,38159 22,73376 9,884652
Spanish_Cantabria 65,5994 24,03028 10,37032
Spanish_Castilla_La_M 68,42474 21,68777 9,887488
Spanish_Castilla_Y_Leon 67,58369 22,23754 10,17877
Spanish_Cataluna 65,64939 23,70233 10,64828
Spanish_Extremadura 69,92058 20,31071 9,768708
Spanish_Galicia 67,19306 22,49292 10,31403
Spanish_Murcia 68,90913 21,16533 9,925541
Spanish_Valencia 67,61798 22,32716 10,05486
Portuguese 68,08301 21,73482 10,18217


http://abload.de/img/desktop_2013_12_28_215pul7.png

Táltos
12-28-2013, 07:59 PM
My results:
EEF=55.67025
WHG=30.18771
ANE=14.14204

Stellaritic
12-28-2013, 08:52 PM
EEF 100.969986651782
WHG -5.16723462211605
ANE 4.19724797033397

newtoboard
12-28-2013, 09:52 PM
I think this is just testing for finnish/estonian ancestry as proxy for ANE in Europeans.
For non-Europeans, I have the largest ANE, and so does not seem like a test for any "northern" component?

Well these three components can't encompass all the ancestry that South Asians have (because of some non West Eurasian ancestry) have. So some of the ASI is probably being added to the WHG category although it doesn't really fit well anywhere imo (and some of it might be inflating ANE as well). Similar to how extra Siberian ancestry likely inflates the ANE values near the Baltic. On the other hand I think the WHG that DMXX and other West Asians (and the negative association for Assyrians is also true) is probably real as I would be very surprised if all Indo-European groups did not carry a mixture of all three components (you can probably infer from uniparental markers that would have been the case if R1 represents ANE, U2e/U4/U5 represents WHG and lineages like HV, H, T and K represent EEF -we know all of these existed side by side just by looking at Andronovo samples).

Yggdrasil
12-28-2013, 09:57 PM
Father (Mid-Norwegian)
EEF: 42,5
WHG: 40,9
ANE: 16,6

Cousin (Mid-Norwegian)
EEF: 40,9
WHG: 42,4
ANE: 16,6

Mother (West-Norwegian)
EEF: 42,9
WHG: 40,7
ANE: 16,3

(Norwegian average 41.1 EEF, 42.8 WHG, 16.1 ANE)

DMXX
12-28-2013, 10:05 PM
So how accurate is this for Middle Easterners because we seem to plot far away from EEF on the plot:


Perhaps not in the best way possible, but we have to remember this PCA will show different patterns if viewed through different dimensions. The relationships are difficult to ascertain, but farming technology did enter Europe from the southeast, so some sort of relationship between modern Near-Easterners and EEF should be expected.

As the Assyrian and Mandean results show, something is certainly amiss, given the EEF itself is partially composed of WHG, when the aforementioned groups score 101% EEF but actually have negative values for WHG.

Clearly the three-way model doesn't apply particularly well for Near-Easterners. Different ancestral groups are likely responsible for the differences there. It will be pretty difficult acquiring these given sample preservation issues in the region thanks to climate. Arguably, forming a simplified model (as is presented here) will be even more difficult given the crossroads nature of the region as well as the numerous intrusive movements that have resulted in permanent demic events (SSA geneflow through Arabia, Turkish geneflow through West Asia, prehistoric Indo-Iranian input, Mongolian depopulation etc.)

Hok
12-28-2013, 10:07 PM
Some Norwegian Romani cousins of ours:

5.
EEF 47.9724514717
WHG 36.5801590945
ANE 15.4473894338

Wow, this one is very similar to mine.

Humanist
12-28-2013, 10:29 PM
My results:


99.44__EEF
-6.45__WHG
7.02___ANE

My K13 values (rounded to nearest %):


38_East_Med
35_West_Asian
10_West_Med
08_Red_Sea
04_South_Asian
02_North_Atlantic
01_Siberian
01_Baltic

newtoboard
12-28-2013, 11:10 PM
Daytripper just posted the correlation of these components to components from other projects on the ABF forum.

And as expected WHG's correlation to South Asian is second only to the Baltic component.

Stephen1986
12-29-2013, 12:04 AM
Myself -

EEF 45.93
WHG 38.49
ANE 15.57

My brother -

EEF 46.88
WHG 37.63
ANE 15.48

rossa
12-29-2013, 12:08 AM
Mine are

EEF 44.2872057
WHG 39.86229591
ANE 15.85049839

leonardo
12-29-2013, 01:12 AM
This calculator is somewhat inapplicable to me, since 5%-15% of my genetic ancestry is derived from an ancient population of South Eurasians, and they aren't accounted for in this model. Nevertheless, I'm predominantly West Eurasian, and my largest component for K13 is "West Asian", so I thought it might be fun to see what this makes of me.

Results:
EEF 59.45%
WHG 24.20%
ANE 16.33%

I'm tempted to treat my "ANE" percentage as capturing all of my non-West Eurasian admixture, since I'm consistently 4%-6% Siberian-Amerindian on most calculators, along with the 10% "ASI". But that wouldn't be accurate, since I'm 9.39% Baltic, and 1.53% North_Atlantic.

My father's results are similar to yours. He has 25% Italian ancestry and possibly a small portion (5-10%?) of recent West Asian ancestry. The rest is western and northern European ancestry:
EEF 56.83897
WHG 29.28624
ANE 13.87479

evon
12-29-2013, 02:24 AM
Wow, this one is very similar to mine.

What is your ancestry?

Joe B
12-29-2013, 03:45 AM
"Shaken, not stirred"
EEF 56.04
WHG 29.86
ANE 14.10



Early European Farmer (EEF)
West European Hunter-Gatherer (WHG)
Ancient North Eurasian (ANE)

seferhabahir
12-29-2013, 05:42 AM
My results:

EEF- 84.8%
WHG - 4.8%
ANE - 10.4%

I have somewhat similar numbers to DMXX (but probably a lot closer to AJL's Ashkenazi grandfather)...

EEF - 89.12%
WHG - 3.76%
ANE - 7.12%

parasar
12-29-2013, 06:33 AM
Southern India
soulblighter
EEF 47.45681494
WHG 34.75186065
ANE 17.7913244

Eastern India
parasar
EEF 47.99684431
WHG 34.19761276
ANE 17.80554294

North Western India
Sapporo
EEF 52.16553515
WHG 30.15350707
ANE 17.68095779

Pakhtoon
Sein
EEF 59.45%
WHG 24.20%
ANE 16.33%

A discernible increase in EEF/WHG ratio towards the Northwest.

parasar
12-29-2013, 06:40 AM
Southern India
soulblighter
EEF 47.45681494
WHG 34.75186065
ANE 17.7913244

Eastern India
parasar
EEF 47.99684431
WHG 34.19761276
ANE 17.80554294

North Western India
Sapporo
EEF 52.16553515
WHG 30.15350707
ANE 17.68095779

Pakhtoon
Sein
EEF 59.45%
WHG 24.20%
ANE 16.33%

A discernible increase in EEF/WHG ratio towards the Northwest.

For comparison:
Eastern Europe, Poland
PL1
EEF 42.012706
WHG 40.52702615
ANE 17.46026785
http://www.forumbiodiversity.com/showthread.php/40732-EEF-WHG-ANE-test-for-Europeans-split-mod

seferhabahir
12-29-2013, 07:04 AM
I have somewhat similar numbers to DMXX (but probably a lot closer to AJL's Ashkenazi grandfather)...

EEF - 89.12%
WHG - 3.76%
ANE - 7.12%

And these are close to Ponto's posted Maltese values (Maltese and Ashkenazi numbers in the paper's Extended Data Table 3 are almost identical)

Maltese
EEF - 90.96%
WHG - 2.11%
ANE - 6.93%

http://bga101.blogspot.com.au/2013/12/eef-whg-ane-test-for-europeans.html

Hok
12-29-2013, 01:06 PM
What is your ancestry?

Dutch(Western/Northern regions)&German.(Northern regions)

Clinton P
12-29-2013, 03:48 PM
Here are my results, my ancestry is British Isles (Predominately English, with small amounts of Welsh and Irish).


EEF 49.8%
WHG 35.4%
ANE 14.8%


Clinton P

MJost
12-29-2013, 08:53 PM
Here is my half maternal sister M093357 who has a Norwegian paternal admix and my half maternal brother M104509 (HG-U106) (a NW German Jost) paternal

Results M093357
EEF 44.58216245
WHG 39.41561543
ANE 16.00222212

Results M104509
EEF 65.06077319
WHG 25.00169437
ANE 9.937532437

MJost

Ok, just to show how these results do vary between my other three half maternal Jost siblings, here they are:

Results M092900
EEF 52.67800531
WHG 33.025656
ANE 14.29633869

Results M092052
EEF 50.92815793
WHG 34.63574435
ANE 14.43609772

Results M092630
EEF 50.68941508
WHG 34.89973677
ANE 14.41084815

Here is my Wife's brother who is Scot I2b1 a Clan MacLeod Sept and MtDNA H15a

Results M065559
EEF 50.49807513
WHG 34.80381694
ANE 14.69810792


MJost

basque
12-29-2013, 09:44 PM
:help: I have downloaded the spreadsheet but it wont let me add my results :confused:

My results are:
North Atlantic 48. 40%, Baltic 22.90%, West Med 15.79%. West Asian 6.71%, East Med 2.82%, South Asian 2.13%, Siberian 0.27%, Amerindian 0.99%

My Dads:
North Atlantic 48.98%, Baltic 21.19%, West Med 15.92%, West Asian 7.19%, East Med 3.21%, South Asian 2.40%, Amerindian 1.11%

I would be forever grateful if someone could put them on a spreadsheet for me and post my results :rolleyes:

basque

emmental
12-29-2013, 09:47 PM
Me

EEF 53.78
WHG 32.24
ANE 13.98

my Daughter

EEF 51.29
WHG 34.21
ANE 14.50

MJost
12-29-2013, 09:55 PM
:help: I have downloaded the spreadsheet but it wont let me add my results :confused:

My results are:
North Atlantic 48. 40%, Baltic 22.90%, West Med 15.79%. West Asian 6.71%, East Med 2.82%, South Asian 2.13%, Siberian 0.27%, Amerindian 0.99%

My Dads:
North Atlantic 48.98%, Baltic 21.19%, West Med 15.92%, West Asian 7.19%, East Med 3.21%, South Asian 2.40%, Amerindian 1.11%

I would be forever grateful if someone could put them on a spreadsheet for me and post my results :rolleyes:

basque

Results Basque
EEF 48.3116158
WHG 36.62701613
ANE 15.06136807

Results Basque's Dad
EEF 49.19217259
WHG 35.93217186
ANE 14.87565556


MJost

MJost
12-29-2013, 09:56 PM
Basque, It didn't change much as your mother must be very similar as well.

MJost

Táltos
12-30-2013, 05:48 AM
"Shaken, not stirred"
EEF 56.04
WHG 29.86
ANE 14.10

Early European Farmer (EEF)
West European Hunter-Gatherer (WHG)
Ancient North Eurasian (ANE)
Hey Joe B so far your results look closest to mine! B)
Here are mine again to compare.
EEF=55.67025
WHG=30.18771
ANE=14.14204

Isidro
12-30-2013, 07:09 AM
Mine , Spain:

23andMe
EEF 65.91
WHG 23.61
ANE 10.47

FTDNA
EEF 66.08
WHG 23.46
ANE 10.44

vettor
12-30-2013, 08:30 AM
I forgot I had 2 accounts at Gedmatch

23andme results below
Population
North_Atlantic 29.98%
Baltic 16.99%
West_Med 23.08%
West_Asian 8.11%
East_Med 17.30%
Red_Sea 2.29%
South_Asian 1.16%
East_Asian 0.56%
Siberian -
Amerindian -
Oceanian 0.22%
Northeast_African -
Sub-Saharan 0.31%

EEF 69.20886831
WHG 20.20271451
ANE 10.58841718


and fdna below
Population
North_Atlantic 29.88%
Baltic 17.04%
West_Med 22.94%
West_Asian 8.44%
East_Med 17.23%
Red_Sea 2.34%
South_Asian 0.90%
East_Asian 0.64%
Siberian -
Amerindian -
Oceanian 0.27%
Northeast_African -
Sub-Saharan 0.32%

EEF 69.32388376
WHG 20.08270538
ANE 10.59341086


both sit between Bergamo and South France

jeanL
12-30-2013, 03:05 PM
My mom

Results
EEF 70.24863631
WHG 20.11682292
ANE 9.634540767

My Dad

Results
EEF 69.67465653
WHG 20.36619107
ANE 9.959152399

Me

Results
EEF 69.74456793
WHG 20.39039517
ANE 9.865036897

MfA
12-30-2013, 03:17 PM
I optimized the excel file for West Asians, Here's the results and comparison to study


Near East ANE
Abhkasian 81.624 18.376
Chechen 74.115 25.885
Cypriot 87.699 12.301
Druze 87.671 12.329
Lezgin 73.664 26.336


http://abload.de/img/desktop_2013_12_27_22nef0i.png


You can download the file from here (https://docs.google.com/spreadsheet/ccc?key=0Aqr2nbGXpVFndEZxZHF1OVNJOEtrcVpyLXVmd1Qze EE&usp=sharing)

Dr_McNinja
12-30-2013, 03:18 PM
Old Spreadsheet:

I'll refer to Harappa ID numbers where available:

Me (HRP0349):

EEF 49.19584011
WHG 33.05320861
ANE 17.75095128

My mother (HRP0374):

EEF 50.7085168
WHG 31.71624891
ANE 17.57523429

Punjabi Jatt Muslim (Pahari regions) (HRP0283):

EEF 54.68825073
WHG 28.09780688
ANE 17.21394239

Afghan Pashtun Durrani (HRP0370):

EEF 62.40525211
WHG 21.94774568
ANE 15.64700221

MfA
12-30-2013, 03:29 PM
I optimized the excel file for West Asians, Here's the results and comparison to study


Near East ANE
Abhkasian 81.624 18.376
Chechen 74.115 25.885
Cypriot 87.699 12.301
Druze 87.671 12.329
Lezgin 73.664 26.336


http://abload.de/img/desktop_2013_12_27_22nef0i.png


You can download the file from here (https://docs.google.com/spreadsheet/ccc?key=0Aqr2nbGXpVFndEZxZHF1OVNJOEtrcVpyLXVmd1Qze EE&usp=sharing)

My new results and Anatolian Kurds and Eurogenes Kurds medians respectively

Near East 82.265
ANE 17.735

Near East 81.805
ANE 18.195

Near East 81.681
ANE 18.319

Dr_McNinja
12-30-2013, 03:30 PM
New spreadsheet:

Me (HRP0349):

Near East 61.015
ANE 38.985

My mother (HRP0374):

Near East 61.913
ANE 38.087

Punjabi Jatt Muslim (Pahari regions) (HRP0283):

Near East 64.043
ANE 35.957

Afghan Pashtun Durrani (HRP0370):

Near East 68.609
ANE 31.391

Sapporo (HRP0341):

Near East 62.465
ANE 37.535

Sein (HRP0282):

Near East 66.833
ANE 33.167

Parasar (HRP0003):

Near East 60.371
ANE 39.629

Isidro
12-30-2013, 04:07 PM
50% Spanish 25% German 25% Scandinavian Ancestry,


Son:

EEF 53.15
WHG 32.82
ANE 14.01


Daughter:

EEF 54.71
WHG 31.98
ANE 13.30

vettor
12-30-2013, 05:20 PM
My mom

Results
EEF 70.24863631
WHG 20.11682292
ANE 9.634540767

My Dad

Results
EEF 69.67465653
WHG 20.36619107
ANE 9.959152399

Me

Results
EEF 69.74456793
WHG 20.39039517
ANE 9.865036897

we have similar numbers, a guess looking at your marker and my known ancestry puts us anywhere in a range between modern Southern Germany and Northern Italy, then again I could be wrong

MfA
12-30-2013, 05:22 PM
http://abload.de/img/desktop_2013_12_30_185krbt.png


Near East ANE
Abhkasian 81.624 18.376
Adygei 77.277 22.723
Armenian 85.268 14.732
Assyrian 86.060 13.940
Azeri_Jewish 85.193 14.807
Balkar 76.829 23.171
Bedouin 88.518 11.482
Chechen 74.115 25.885
Cyprian 87.699 12.301
Egyptian 89.463 10.537
Georgian 81.341 18.659
Georgian_Jewish 85.408 14.592
Iranian 81.214 18.786
Jordanian 88.066 11.934
Kumyk 76.692 23.308
Kurdish 81.681 18.319
Leb_Christian 89.343 10.657
Leb_Druze 87.671 12.329
Leb_Muslim 87.091 12.909
Lezgin 73.664 26.336
North_Ossetian 76.558 23.442
Palestinian 88.460 11.540
Samaritan 89.480 10.520
Saudi 90.089 9.911
Syrian 87.186 12.814
Turkish 82.305 17.695
Yemenite_Jewish 90.991 9.009

DMXX
12-30-2013, 05:31 PM
Text

Simple re-ordering of the data MfA's kindly shared focused on ANE in order of decreasing value.

It is apparent the frequency of ANE decreases as one moves further away from the North Caucasus. A near-perfect geographical axis beginning with the North Caucasus and ending with the southern Arabian peninsula exists.

Also, MfA, you mentioned optimising the calculator for West Asians. Have you shared it here?



Near-East ANE
Lezgin 73.664 26.336
Chechen 74.115 25.885
North_Ossetian 76.558 23.442
Kumyk 76.692 23.308
Balkar 76.829 23.171
Adygei 77.277 22.723
Iranian 81.214 18.786
Georgian 81.341 18.659
Abhkasian 81.624 18.376
Kurdish 81.681 18.319
Turkish 82.305 17.695
Azeri_Jewish 85.193 14.807
Armenian 85.268 14.732
Georgian_Jewish 85.408 14.592
Assyrian 86.060 13.94
Leb_Muslim 87.091 12.909
Syrian 87.186 12.814
Leb_Druze 87.671 12.329
Cyprian 87.699 12.301
Jordanian 88.066 11.934
Palestinian 88.460 11.54
Bedouin 88.518 11.482
Leb_Christian 89.343 10.657
Egyptian 89.463 10.537
Samaritan 89.480 10.52
Saudi 90.089 9.911
Yemenite_Jewish 90.991 9.009

MfA
12-30-2013, 05:32 PM
Also, MfA, you mentioned optimising the calculator for West Asians. Have you shared it here?


Yes, at previous page.

emmental
12-30-2013, 05:43 PM
50% Spanish 25% German 25% Scandinavian Ancestry,


Son:

EEF 53.15
WHG 32.82
ANE 14.01


Daughter:

EEF 54.71
WHG 31.98
ANE 13.30

This is very close to mine. I'm about 70% Swiss, 30% Dutch/North German (Low German); my daughter is mostly the same basic gene pool with 12.5% West Prussian (Kulm) mixed in.

DMXX
12-30-2013, 05:56 PM
Yes, at previous page.

My corrected West Asian result:

Near East 79.886
ANE 20.114

Looking at the list you provided, my value is between the Adyghei and Iranian values.

MfA
12-30-2013, 06:04 PM
My corrected West Asian result:

Near East 79.886
ANE 20.114

Looking at the list you provided, my value is between the Adyghei and Iranian values.

There is a Kurdish guy with 1/8 Adygei ancestry, his results similar to yours.
Near East 79.929
ANE 20.71

DMXX
12-30-2013, 06:07 PM
There is a Kurdish guy with 1/8 Adygei ancestry, his results similar to yours.
Near East 79.929
ANE 20.71

If you're referring to the gentleman who I think you're referring to (StarDS9?), then I'm not surprised. Him, myself and NK19191 on this forum (as well as a Sheikh_Ubayd on ABF) would regularly score results within 1% of each other.

MfA
12-30-2013, 06:14 PM
If you're referring to the gentleman who I think you're referring to (StarDS9?), then I'm not surprised. Him and NK19191 on this forum (as well as a Sheikh_Ubayd on ABF) would regularly score results within 1% of each other.

Yes that's him, he is a bit different than rest of the Kurds though. Most hovers between ANE 17-19%, thus median is 18%. There is also a 16% guy at opposite pole.

DMXX
12-30-2013, 06:29 PM
Yes that's him, he is a bit different than rest of the Kurds though. Most hovers between ANE 17-19%, thus median is 18%. There is also a 16% guy at opposite pole.

Yes, same goes to the Sheikh_Ubayd I was discussing. He too has some ancestry from the Caucasus, Georgia if I remember correctly. The same applies with myself (oral tradition of N Caucasian ancestry confirmed by certain genetic results), as with NK19191.

So, all four of us are essentially of West Iranic-speaking backgrounds with some minor contribution from various parts of the Caucasus. Despite the diversity in backgrounds, our results come out similarly.

parasar
12-30-2013, 06:40 PM
Here are my results:

EEF 47.45681494
WHG 34.75186065
ANE 17.7913244

My K13 results are in another post, but just for completeness:



North_Atlantic Baltic West_Med West_Asian East_Med Red_Sea South_Asian East_Asian Siberian Amerindian Oceanian Northeast_African Sub-Saharan
You 4.13 3.6 0 23.03 2.05 1.25 61.13 1.59 1.3 0 0.67 0 1.24


soulblighter:
West Asian NE/ANE Calculator
Near East 60.142
ANE 39.858

Sein
12-30-2013, 06:49 PM
So far, it seems I'm most similar to the Afghan Durrani Pashtun, HRP0370.

newtoboard
12-30-2013, 06:54 PM
Yes, same goes to the Sheikh_Ubayd I was discussing. He too has some ancestry from the Caucasus, Georgia if I remember correctly. The same applies with myself (oral tradition of N Caucasian ancestry confirmed by certain genetic results), as with NK19191.

So, all four of us are essentially of West Iranic-speaking backgrounds with some minor contribution from various parts of the Caucasus. Despite the diversity in backgrounds, our results come out similarly.


Which North Caucasian ethnicity do you have ancestry from? The North Caucasian numbers are interesting. It probably indicates even if ANE was initially associated with R1, men of other Y-DNAs probably spread it around early.

DMXX
12-30-2013, 07:05 PM
Which North Caucasian ethnicity do you have ancestry from? The North Caucasian numbers are interesting. It probably indicates even if ANE was initially associated with R1, men of other Y-DNAs probably spread it around early.

Without derailing this thread too much; it appears to be the Ossetians. I first noted an autosomal affinity with North Caucasians on 23andMe's Compare Genes that West Asians (including fellow Iranians) seemed to lack. Later on, Dienekes carried out a fastIBD analysis on West Asians (http://dodecad.blogspot.com/2012/01/fastibd-analysis-of-balkanswest-asia.html) and I was included (DOD010) (https://docs.google.com/spreadsheet/ccc?key=0ArAJcY18g2GadFJ4UFFFUlVWUG5zaXM0LVNaQUROa lE). I am one of the few to have a significant positive association with his North Ossetian reference (+0.74) despite having a negative association with all other Caucasians, including others from the north (e.g. Abkhazians, -0.31). The results weren't a fluke, given I shared well with Iranians* and Kurds. I even shared more segments here with Ossetians than Turks (+0.74 vs. +0.61).

All of this affirmed the oral tradition on my paternal side claiming we were "Iranians from outside Iran [meaning above the river Aras (Araxes)], near Russia". Perfectly so, I should add.

* Iranian_D sample is composed of individuals from across the country... Didn't see the sense in including them. Had a negative score with this.

MitchellSince1893
12-30-2013, 07:06 PM
Mine:
EEF 44.96294
WHG 39.0575
ANE 16.0311

Father's:
EEF 47.57097
WHG 37.06065
ANE 15.36838

Mother's
EEF 45.87243
WHG 38.38094
ANE 15.74664

Looks like my closest matches are Rossa (1.713084 difference), my mom (1.81897 difference) and RMS2 (3.070412 difference)

My closest country matches are:
-Ukrainian (2.474 difference)
-Orcadian (5.33706 difference)
-Norwegian (7.725879)

Orkney and Norway make sense as Eurogenes typically plots me in the North Sea. MDLP and Dodecade plot me in Western Germany near Belgian and Dutch borders.




Mine are

EEF 44.2872057
WHG 39.86229591
ANE 15.85049839


My results seem to fall closest to the Orcadian and Ukrainian.

Me:
EEF 46.49814568
WHG 38.03949482
ANE 15.4623595

Orcadian:
EEF 45.7
WHG 38.5
ANE 15.8

Ukrainian
EEF 46.2
WHG 38.7
ANE 15.1

Sapporo
12-30-2013, 07:30 PM
Interesting and logical for North Caucasians like Chechens and Lezgins to score the highest ANE of West Asian populations considering they also tend to have the most Northern European autosomal admixture by far of West Asians on various calculators.

For us from around South Asia, I wonder how much the South Asian component on Eurogenes K13 effects the final results since it is a heavily mixed component between ancient West Eurasian and East Eurasian like affinities.

parasar
12-30-2013, 08:04 PM
Interesting and logical for North Caucasians like Chechens and Lezgins to score the highest ANE of West Asian populations considering they also tend to have the most Northern European autosomal admixture by far of West Asians on various calculators.

For us from around South Asia, I wonder how much the South Asian component on Eurogenes K13 effects the final results since it is a heavily mixed component between ancient West Eurasian and East Eurasian like affinities.

WHG best correlates with early northern Europeans, EEF with southern Europeans and western Asians, and ANE with southern Asians (via Siberians of the MA-1 type).

Sapporo
12-30-2013, 08:06 PM
WHG best correlates with early northern Europeans, EEF with southern Europeans and western Asians, and ANE with southern Asians (via Siberians of the MA-1 type).

Interesting. What would explain the higher ANE in Lezgins and Chechens versus other West Asians in your opinion then?

parasar
12-30-2013, 08:30 PM
Interesting. What would explain the higher ANE in Lezgins and Chechens versus other West Asians in your opinion then?

There may be multiple explanations - one a very early one which brought ANE to eastern Europe and to the Caucasus.


Another more recent Indic connection to the eastern regions of the Black Sea where Indian tribes had settled, is also possible.


Remembrance of immigration ... there are the (Grk.) sindoi people on the R. Kuban, north of the Caucasus, and there is the (Grk.) sindEs ... Both sindoi and sindEs preserve, with their s-, a pre-Iranian form of the name (details in Witzel 1999)[N.34] that reminds of Vedic sindhu...
http://www.people.fas.harvard.edu/~witzel/EJVS-7-3.htm

While Dr. Witzel thinks that the above Sindoi were remnants of Indians who migrated to India, a Greek text clearly call them Indian (Hesychius: Sintoi ethnos Indikon and Kerketai ethnos Indikon http://books.google.com/books?id=aAcT_D6t_YwC&pg=PA74 )

http://books.google.com/books?id=aAcT_D6t_YwC&pg=PA74&img=1&zoom=3&hl=en&sig=ACfU3U2cf7czrPFQOGy8ScLWUwfjIWku8g&ci=101%2C210%2C839%2C1101&edge=0

please see also http://books.google.com/books?id=J6QUUhU8EN0C&pg=PA293

Plus as Humanist noted:
http://i1096.photobucket.com/albums/g326/dok101/Faces/sindu.jpg
http://www.anthrogenica.com/showthread.php?1519-Languages-and-Y-DNA-lineages&p=22803&viewfull=1#post22803

That the the river Kuban itself derives from Kubha (now Kabul) is quite possible.

Stellaritic
12-30-2013, 09:06 PM
MfA's spreadsheet :

Near East 89.622
ANE 10.378

vettor
12-30-2013, 09:13 PM
It seems to me , viewing results from this forum and other forums that numbers ranging for
EEF 69
WHG 20
ANE 10
and close to these numbers represents, areas ranging from
north Portugal through northern Spain, through southern france, through Northern italy and other alpine areas ( swiss, austrian southern germany)..............ending on the fringes of Hungaria

Dr_McNinja
12-30-2013, 09:28 PM
So far, it seems I'm most similar to the Afghan Durrani Pashtun, HRP0370.
There's a pretty nice gradient going from West to East:

Afghan Pashtun (HRP0370): Near East 68.609 / ANE 31.391
Pakistan Pashtun (HRP0282): Near East 66.833 / ANE 33.167
Pakistan Pahari Punjabi Jatt Muslim (HRP0283): Near East 64.043 / ANE 35.957
India Punjabi Jatt Sikh (HRP0341): Near East 62.465 / ANE 37.535
India Bihar Brahmin (HRP0003): Near East 60.371 / ANE 39.629

the SUN child
12-30-2013, 09:45 PM
I think this new calculator is pretty useless for anybody outside the European continent without the European roots or not native to Europe.

Sapporo
12-30-2013, 09:48 PM
There's a pretty nice gradient going from West to East:

Afghan Pashtun (HRP0370): Near East 68.609 / ANE 31.391
Pakistan Pashtun (HRP0282): Near East 66.833 / ANE 33.167
Pakistan Pahari Punjabi Jatt Muslim (HRP0283): Near East 64.043 / ANE 35.957
India Punjabi Jatt Sikh (HRP0341): Near East 62.465 / ANE 37.535
India Bihar Brahmin (HRP0003): Near East 60.371 / ANE 39.629

Could my higher N Euro vs. HRP0283's higher Caucasus explain his lower ANE? Otherwise, our other components don't differ much.

MitchellSince1893
12-30-2013, 09:53 PM
Mine:
EEF 44.96294
WHG 39.0575
ANE 16.0311


If you take the average of England & Scotland it's close to my numbers

EEF = 44.25
WHG = 39.6
ANE = 16.15

Or England, Scotland, & Orkney average is spot on my numbers...only .46 difference.

EEF = 44.73
WHG = 39.23
ANE = 16.03

Which would make a lot of sense looking at my signature I.e. I'm probably 80 to 85%British.

ZephyrousMandaru
12-30-2013, 10:28 PM
Near East 86.323
ANE 13.677

parasar
12-30-2013, 10:37 PM
I optimized the excel file for West Asians, Here's the results and comparison to study


Near East ANE
Abhkasian 81.624 18.376
Chechen 74.115 25.885
Cypriot 87.699 12.301
Druze 87.671 12.329
Lezgin 73.664 26.336


http://abload.de/img/desktop_2013_12_27_22nef0i.png


You can download the file from here (https://docs.google.com/spreadsheet/ccc?key=0Aqr2nbGXpVFndEZxZHF1OVNJOEtrcVpyLXVmd1Qze EE&usp=sharing)

NE Europeans when fitted to Near East/ANE should show even higher ANE than North Caucasians, right?

Sein
12-31-2013, 12:10 AM
Please see "Reason for Editing".

MfA
12-31-2013, 12:50 AM
NE Europeans when fitted to Near East/ANE should show even higher ANE than North Caucasians, right?

I don't have a perfect answer.. Probably that is because of WHG and ANE is on the different poles of the same cline and distinct from EEF/(hypothetical Near East) and hNE still shares some common origin with that though. So on lower resolution majority of WHG merges with ANE and minority with hNE.. So we have elevated ANE for Europeans when using K2.

Hanna
12-31-2013, 11:55 AM
My new results:

NE: 82.774
ANE: 17.256

My aunt's results:

NE: 84.226
ANE: 15.774

NK19191
12-31-2013, 04:05 PM
My corrected West Asian result:

Near East 79.886
ANE 20.114

Looking at the list you provided, my value is between the Adyghei and Iranian values.


Here is mine

Results

Near East 80.205
ANE 19.795

Wulf Talented
01-01-2014, 12:00 PM
Results
EEF 47.36552359
WHG 37.16203532
ANE 15.47244109

seferhabahir
01-01-2014, 08:45 PM
My newer values with the West Asian calculator:

EEF 84.14
ANE 15.86

This would be in between the following two (although Ashkenazi has no easily recognizable geographic center):

Turkish 82.305 17.695
Azeri_Jewish 85.193 14.807

My Gedrosia and Southwest Asian K12b components are higher than typical Ashkenazi values.

R.Rocca
01-03-2014, 04:02 AM
My wife is of 100% Polish ancestry by way of her 4 grandparents. Her paternal grandfather is from Zalesie, Podlasie, Poland and is I2a1b-M423 like the Loschbour hunter-gatherer. Here are my wife's results:

EEF 41.83587054
WHG 40.99387637
ANE 17.17025309

Not surprisingly, she clusters closest to Belorussians and Norwegians

Belorussian
EEF 0.418
WHG 0.431
ANE 0.151

Norwegian
EEF 0.411
WHG 0.428
ANE 0.161

Michał
01-03-2014, 10:53 AM
My wife is of 100% Polish ancestry by way of her 4 grandparents. Her paternal grandfather is from Zalesie, Podlasie, Poland and is I2a1b-M423 like the Loschbour hunter-gatherer. Here are my wife's results:

EEF 41.83587054
WHG 40.99387637
ANE 17.17025309

Here are the results for four unrelated Poles plus one (myself, 198727) who is related to the first two people. All their known ancestors (including all four grandparents in each case) are of Polish origin.

FTDNA 216607
EEF 38.71758
WHG 43.4736
ANE 17.80882

FTDNA 225210
EEF 40.36797
WHG 42.26724
ANE 17.36479

FTDNA 225211
EEF 44.386
WHG 39.09559
ANE 16.51841

FTDNA 225212
EEF 39.71178
WHG 42.85374
ANE 17.43448

FTDNA 198727
EEF 41.02334
WHG 41.55785
ANE 17.41882

lifeisdandy
01-03-2014, 03:30 PM
Results
EEF 112.0844477
WHG -14.66282829
ANE 2.578380572


Results
Near East 91.393
ANE 8.607

Silesian
01-03-2014, 05:25 PM
Here are the results for four unrelated Poles plus one (myself, 198727) who is related to the first two people. All their known ancestors (including all four grandparents in each case) are of Polish origin.

FTDNA 216607
EEF 38.71758
WHG 43.4736
ANE 17.80882

FTDNA 225210
EEF 40.36797
WHG 42.26724
ANE 17.36479

FTDNA 225211
EEF 44.386
WHG 39.09559
ANE 16.51841

FTDNA 225212
EEF 39.71178
WHG 42.85374
ANE 17.43448

FTDNA 198727
EEF 41.02334
WHG 41.55785
ANE 17.41882

Poland, Silesia:

Father yDNA R1b1a2a1 + mtDNA H2a1

EEF 42.59
WHG 40.4124
ANE 16.99

My results R1b1a2a1 + U5b2a2

EEF 45.15
WHG 38.67
ANE 16.17

Scarlet Ibis
01-05-2014, 03:18 AM
Discussion interpreting the Lazaridis et al paper was moved here: http://www.anthrogenica.com/showthread.php?1756-Ancient-human-genomes-suggest-three-ancestral-populations-for-Europeans/page29

parasar
01-05-2014, 05:16 AM
I had a feeling WHG was actually somewhat ASI for non europeans when Sein first posted his results. Thanks for confirming it.


Here are all the correlations:

EEF
East Med (+0.94855)
Red Sea (+0.83288)
West Med (+0.67335)
Northeast African (+0.5254)
West Asian (+0.41363)
Sub-Saharan (+0.33439)
East Asian (+0.30893)
Oceanian (-0.37898)
North Atlantic (-0.5244)
Siberian (-0.56267)
South Asian (-0.6387)
Amerindian (-0.64406)
Baltic (-0.82668)

WHG
Baltic (+0.78009)
South Asian (+0.60554)
Amerindian (+0.57381)
North Atlantic (+0.53109)
Siberian (+0.52934)
Oceanian (+0.40564)
East Asian (-0.34429)
Sub-Saharan (-0.35298)
Northeast African (-0.54025)
West Asian (-0.55277)
West Med (-0.56)
Red Sea (-0.85739)
East Med (-0.95655)

ANE
Baltic (+0.77605)
Amerindian (+0.71262)
South Asian (+0.59046)
Siberian (+0.53338)
North Atlantic (+0.37778)
Oceanian (+0.20167)
West Asian (+0.12404)
East Asian (-0.1242)
Sub-Saharan (-0.19735)
Northeast African (-0.35262)
Red Sea (-0.55686)
East Med (-0.69661)
West Med (-0.86965)
http://www.forumbiodiversity.com/showthread.php/40732-EEF-WHG-ANE-test-for-Europeans-split-mod/page5

Anglecynn
01-05-2014, 05:19 AM
http://www.forumbiodiversity.com/showthread.php/40732-EEF-WHG-ANE-test-for-Europeans-split-mod/page5

This is interesting because on admixture tests a lot of northern Europeans have noise-level amounts of South Asian, Amerindian, Oceanian - Sometimes totalling a couple of percent, and then also Siberian although that seems to be a bit more localised to Fennoscandia.

Perhaps something to do with WHG?

parasar
01-05-2014, 06:00 AM
This is interesting because on admixture tests a lot of northern Europeans have noise-level amounts of South Asian, Amerindian, Oceanian - Sometimes totalling a couple of percent, and then also Siberian although that seems to be a bit more localised to Fennoscandia.

Perhaps something to do with WHG?

Both WHG and ANE, I suppose.

everest59
01-06-2014, 01:27 AM
How exactly can ANE be calculated for South Asians? The formula used by Lazaridis et al for North Caucasians was:
f4 (Test,Stuttgart; Karitiana, Onge) / f4 (MA1, Stuttgart; Karitiana, Onge)

Can this formula be applied to South Asians as well? I ask this because using the Onge as an outgroup will produce a different result from using Yoruba as an outgroup.

Using Onge as the outgroup produces the following:
UP Brahmin : 4 %
Kshatriya : None
Kashmiri Pandits: 3%
Pathans : 7%
Irula: None

Using Yoruba on the other hand produces the following:
UP Brahmin: 21.2%
Kshatriya: 20.5%
Kashmiri Pandit : 21%
Pathan: 18%
Onge :29%

(There were some tribals in the 30's)

One difference from the Laziridis calculation is that I do not have the Stuttgart data. I substituted with Sardinian. For comparison, here are some European results using Onge as outgroup:

Orcadian: 14.2%
French: 13.3%
Tuscan : 9.9%

Slight difference from the paper, probably because Sardinians have a tiny bit of ANE.

I think the first calculation makes more sense because it ignores the Onge. The ANE is calculated from the West Eurasian portion only.

Sein
01-06-2014, 02:08 AM
I have to agree with you, the first calculation makes sense when looking at geography.

everest59
01-06-2014, 02:47 AM
I have to agree with you, the first calculation makes sense when looking at geography.

The SNP count was a little low though, at only 11k. However, I think it's still pretty accurate looking at the European results. I wasn't able to include the 23andme data.

parasar
01-06-2014, 03:19 AM
How exactly can ANE be calculated for South Asians? The formula used by Lazaridis et al for North Caucasians was:
f4 (Test,Stuttgart; Karitiana, Onge) / f4 (MA1, Stuttgart; Karitiana, Onge)

Can this formula be applied to South Asians as well? I ask this because using the Onge as an outgroup will produce a different result from using Yoruba as an outgroup.

Using Onge as the outgroup produces the following:
UP Brahmin : 4 %
Kshatriya : None
Kashmiri Pandits: 3%
Pathans : 7%
Irula: None

Using Yoruba on the other hand produces the following:
UP Brahmin: 21.2%
Kshatriya: 20.5%
Kashmiri Pandit : 21%
Pathan: 18%
Onge :29%

(There were some tribals in the 30's)

One difference from the Laziridis calculation is that I do not have the Stuttgart data. I substituted with Sardinian. For comparison, here are some European results using Onge as outgroup:

Orcadian: 14.2%
French: 13.3%
Tuscan : 9.9%

Slight difference from the paper, probably because Sardinians have a tiny bit of ANE.

I think the first calculation makes more sense because it ignores the Onge. The ANE is calculated from the West Eurasian portion only.

If we see http://biorxiv.org/content/biorxiv/suppl/2013/12/23/001552.DC1/001552-2.pdf MA-1 has significant Ongee at K=16 (lowest error), much higher than most South Asians at K=16. The early Siberian also show connection to Papuan at K=16.

Logically I would think that for Europeans and Near Easterners, the Ongee are a legitimate outgroup.

"We also identified a list of 13 of these populations which show no evidence of either European or Near Eastern ancestry ... Ami, Atayal, Bougainville, Esan, Han, Ju_hoan_North, Karitiana, Mbuti, Naxi, Onge, Papuan, She, Yoruba"


But for South Asians, I'm not sure that the Ongee are a good choice. Perhaps the Yaruba may be the better outgroup as the Ongee are on the ASI side of the branching, and ASI is a significant component on all populations of Reich's 'Indian Cline.'
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2842210/bin/nihms137159f4.jpg

everest59
01-06-2014, 03:44 AM
Yeah, with the first calculation, the assumption is that the Onge has no relationship with ANE, which may not be accurate. As far as Papuans, check this out:

alpha std. err Z (null=0)
result: Karitiana Yoruba Papuan Sardinian : Karitiana Yoruba Mal'ta Sardinian 0.183545 0.039626 4.632
## end of run

Admixture of 18% with Standard Error of 3.96% and Z-score of 4.632.

BTW, the formula won't work for East Asians like the Han due to the fact that Karitiana has East Asian admixture (I'm getting Admixture of >60%). So I'll need to find a population to replace Karitiana.

So basically, if the first calculation is correct, ANE is more related to Onge than Papuan (but related to Papuan as well). I have found some Indian tribals whose admixture was in the 30's. For the Irula, however, their ANE was similar to UP Brahmins at 21 %.

parasar
01-06-2014, 04:05 AM
Yeah, with the first calculation, the assumption is that the Onge has no relationship with ANE, which may not be accurate. As far as Papuans, check this out:

alpha std. err Z (null=0)
result: Karitiana Yoruba Papuan Sardinian : Karitiana Yoruba Mal'ta Sardinian 0.183545 0.039626 4.632
## end of run

Admixture of 18% with Standard Error of 3.96% and Z-score of 4.632.

BTW, the formula won't work for East Asians like the Han due to the fact that Karitiana has East Asian admixture (I'm getting Admixture of >60%). So I'll need to find a population to replace Karitiana.

So basically, if the first calculation is correct, ANE is more related to Onge than Papuan (but related to Papuan as well). I have found some Indian tribals whose admixture was in the 30's. For the Irula, however, their ANE was similar to UP Brahmins at 21 %.

My thinking is that MA-1 is descended of an ancestral south east asian population that was entering South Asia and Eastern Europe at the same time from SE Asia. http://www.anthrogenica.com/showthread.php?1483-24-000-year-old-Y-DNA-R-AND-MTDNA-U-FOUND-IN-SIBERIA!!!!&p=17781&viewfull=1#post17781

The ones in South Asia stayed put, while the ones in Eastern Europe abandoned the region in LGM. MA-1 is one the latter. So the greater proximity of MA-1 to some Indian tribals is to be expected.

gravetti
01-06-2014, 06:37 PM
New calculator created by Day Tripper.

How to use it.

1. Download this spreadsheet.

https://app.box.com/s/nr09lj4xbgivblr4tpm5

2. Compute your ADMIXTURE Proportions Using Eurogenes K13.

3. Input your percentages in each corresponding labeled columns (Besides the "You").

My Results rounded to the nearest percent.

EEF 101%
WHG -7.61%
ANE 6.6%

http://bga101.blogspot.com.au/2013/12/eef-whg-ane-test-for-europeans.html

I can´t input my procentages in the corresponding columns.Why?

J Man
01-07-2014, 12:23 AM
I am sick and tired of a lot of the people on a forum I frequent (not this one) saying that this test is absolutely no good for far Southern European people such as Southern Italians and Greeks, no good for Near Eastern people and also that it is no good for far Northeastern European people such as Finns. My question now concentrates on the results of Southern European peoples and Near Eastern peoples and their relation to the WHG component. So in the EEF-WHG-ANE test that you designed many Southern Europeans and especially Near Eastern/West Asian people such as Turks and Assyrians tend to score negative results for the WHG component. This is in line with the paper itself when they say that Near Eastern people score zero for the WHG component. So then it this EEF-WHG-ANE test useful for Southern Europeans and Near Eastern people in the that they score very low amounts of even negative amounts of the WHG component? I mean it shows what the paper did in that they tend to score zero. I know someone created a special test for Near Eastern and Southeastern European people but is this one useful in showing that they score pretty much zero or even less for the WHG component? I hope this makes sense what I am saying and asking.

Scarlet Ibis
01-08-2014, 06:57 AM
Not very relevant for me, but here are my results just for fun:

EEF: 55.61%
WHG: 30.13%
ANE: 14.26%

Invisible Sun
01-10-2014, 10:57 PM
Here are my results with both Calculators. Known ancestry: 7/8 West Sicilian + 1/8 North Italian.

EEF 84,711%
WHG 7,710%
ANE 7,579%

Near East 82,946%
ANE 17,054%

vettor
01-11-2014, 12:47 AM
I can´t input my procentages in the corresponding columns.Why?

replace the numbers along the "you" row with your numbers

gravetti
01-11-2014, 03:37 PM
replace the numbers along the "you" row with your numbers

Thanks,vettor.

My results (I am I2a1b3a+ H1a1-without exact matches) :

EEF: 60.3792
WHG: 26.5262
ANE: 13.0945

Arbogan
02-18-2014, 03:10 PM
Regarding the ANE test. Does the modified ANE-EFF test for south-europeans and west-asians, control for ENA(South-asian, Siberian, sinid and other related groups)? Because we know that ENA groups have high correlations with ANE. It would make sense to add such a component to the test.

Update:
Apparently they used Onge and Dai as controls for ENA ancestry. But they failed to build a model for caucasians/west-asians and hence couldn't infer the amount of ANE ancestry.

MfA
03-08-2014, 10:54 AM
Ancient North Eurasian (ANE) levels across Asia (http://eurogenes.blogspot.com/2014/03/ancient-north-eurasian-ane-levels.html)

My results:
0.196525 ANE(MA-1)
0.803455 ME(South Bedouin)
0.00001 ENA(Han)
0.00001 SSA(Yoruba)

Sein
03-08-2014, 03:35 PM
My results for a very accurate PCA based test by David:
ME=47.3%
ANE=36.2%
ENA=16.5%
Here are other results for comparison:
https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdGE5bjNCdl8yci0yNGN4TWk4M3NUW VE&usp=sharing#gid=0

This test was extremely accurate, but it was very time consuming. He had to run each sample once at a time. This new test is much less accurate, but still very nice, and gets the ballpark:
https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdC01RENtMGpRc3FHTGZpSUFKR3hHY 1E#gid=0

My new results for this Admixture run:
ME=53.8%
ANE=31.9%
ENA=11.6%
SSA=2.5%

As you can see, the Admixture based results are very nice, but less accurate. I have less ENA now, and I'm even scoring some SSA.

everest59
03-08-2014, 03:55 PM
Well, as David acknowledged, ANE is inflated in South/South-Central Asians. If you look at the Paniya for example, even they are at 24%. Pashtuns, Jats, Kalash as well as the Brahmins are in the 30's range, so it's not completely inaccurate as we would definitely assume it would be higher in these groups.
IMO, it will be in the 10-20% range, and IMO certainly less than say the North Caucasians.
This is what David said, and it is what I've been saying:

Nevertheless, I think there is an issue with this ANE component, which is that it contains South Asian admixture, but more on that soon.

http://eurogenes.blogspot.com/2014/03/ancient-north-eurasian-ane-levels.html

ADMIXTURE cannot properly differentiaete ANE, ASI, ENA and probably WHG (because David said that if he included Northern Europeans, due to higher WHG in them, they would end up as having the highest ANE).

Sein
03-08-2014, 04:01 PM
Well, as David acknowledged, ANE is inflated in South/South-Central Asians. If you look at the Paniya for example, even they are at 24%. Pashtuns, Jats, Kalash as well as the Brahmins are in the 30's range, so it's not completely inaccurate as we would definitely assume it would be higher in these groups.
IMO, it will be in the 10-20% range, and IMO certainly less than say the North Caucasians.
This is what David said, and it is what I've been saying:


http://eurogenes.blogspot.com/2014/03/ancient-north-eurasian-ane-levels.html

ADMIXTURE cannot properly differentiaete ANE, ASI, ENA and probably WHG (because David said that if he included Northern Europeans, due to higher WHG in them, they would end up as having the highest ANE).

Yes, I do think the ADMIXTURE test is less accurate. But the PCA test was solid. I honestly think it was even more accurate than the values found in the paper. The correspondence was perfect, and the ENA numbers made sense. But it would have taken forever, so he had to make a choice between accuracy and speed. If our reference point is the paper, I probably am around 37% ANE. And since in this run, many Pashtuns, and other South Central Asians, have more than myself, I wouldn't be surprised if they turn out 40% ANE! There is a big peak in South-Central Asia, almost equal to that found among Native Americans.

Also, I believe South Indians didn't register any ANE in the PCA test (I'll ask him about this).

everest59
03-08-2014, 04:17 PM
Yes, I do think the ADMIXTURE test is less accurate. But the PCA test was solid. I honestly think it was even more accurate than the values found in the paper. The correspondence was perfect, and the ENA numbers made sense. But it would have taken forever, so he had to make a choice between accuracy and speed. If our reference point is the paper, I probably am around 37% ANE. And since in this run, many Pashtuns, and other South Central Asians, have more than myself, I wouldn't be surprised if they turn out 40% ANE! There is a big peak in South-Central Asia, almost equal to that found among Native Americans.

Also, I believe South Indians didn't register any ANE in the PCA test (I'll ask him about this).

Here is shared genetic drift I calculated a while ago:
Source 1 Source 2 Target f_3 std. err Z SNPs
result: Mal'ta Orcadian Yoruba 0.196754 0.003360 58.562 72978
result: Mal'ta French Yoruba 0.190634 0.003317 57.472 73047
result: Mal'ta Russian Yoruba 0.200768 0.003310 60.648 73058
result: Mal'ta Adygei Yoruba 0.188852 0.003274 57.676 72999
result: Mal'ta Brahmin Yoruba 0.193249 0.004721 40.936 11512
result: Mal'ta Kashmiri_Pandit Yoruba 0.191860 0.004697 40.843 11526
result: Mal'ta Onge Yoruba 0.178263 0.005797 30.752 11369
result: Mal'ta Karitiana Yoruba 0.232218 0.004235 54.838 72423
result: Mal'ta Pathan Yoruba 0.193482 0.003106 62.296 73034
result: Mal'ta Gond Yoruba 0.181407 0.004666 38.880 11506
result: Mal'ta Irula Yoruba 0.183800 0.005097 36.063 11452
result: Mal'ta Meghawal Yoruba 0.188390 0.004895 38.490 11465
result: Mal'ta Basque Yoruba 0.188448 0.003348 56.289 73013

You can see that Russians and the Karitiana have the highest shared genetic drift with Mal'ta. The upper caste Indians as well as South Central Asians are up there, but I really can't see Pashtuns/NW South Asians having higher ANE than certain groups like North Caucasians even if basal european admixture is masking real ANE ancestry (which is the case in the Adygei).

Sein
03-08-2014, 04:24 PM
But in terms of phylogeny, ANE is much closer to ENA than it is to Basal Eurasian. Just assuming that I'm 47% ME, that would make around 20%-24% Basal Eurasian. And this was calculated with Bedouins, whose Basal Eurasian is likely even higher. So, taking myself as an example, I could be around 25%-30% Basal Eurasian. This will strongly confound things. Also, WHG is extremely close to ANE. So, Karitiana and Russians should be on top of the list, as most of their ancestry is phylogenetically closer to ANE. In addition, there are certain specific subtleties in the interpretation of drift statistics, it just isn't clear cut.

And even taking that into account, the HGDP Pashtuns share more drift with MA-1 than the French. Most of the ancestry of the French is closer to ANE, as Pashtuns have substantial Basal Eurasian. Yet, Pashtuns still share more drift than the French. This should tell us something. Also, the Adygei should have a lot of ANE (I think, correct me if I'm wrong), but look at their score.

everest59
03-08-2014, 04:27 PM
I can definitely see it in the 25-30% range, no question. However, we don't know with certainty if the West Eurasian in South Asians even has basal eurasian(it probably does, but it all depends on when West Asians migrated to South Asia).
Like David acknowledged, we need to wait for aDNA from South Asia. Right now, we don't even know the true ASI in various populations.

Sein
03-08-2014, 04:35 PM
I can definitely see it in the 25-30% range, no question. However, we don't know with certainty if the West Eurasian in South Asians even has basal eurasian(it probably does, but it all depends on when West Asians migrated to South Asia).
Like David acknowledged, we need to wait for aDNA from South Asia. Right now, we don't even know the true ASI in various populations.

All West Eurasian populations outside Europe lack WHG, so they should be modeled as mixtures between EEF and ANE. And EEF is almost 50% Basal Eurasian. Also, another note about drift statistics. I believe North Caucasians are lower on the list than Europeans, despite having more ANE than any European population. And South Central Asians are ahead of those populations, which themselves have more ANE than Europeans. If Lezgians share less drift than the HGDP Pastuns, and are around 28% ANE, than 35%-40% ANE for Pashtuns makes sense. This goes for other South Central Asians.

But I absolutely agree, we do need ancient DNA from South Asia. I think we can expect results from the Farmana team, so that is exciting.

MfA
03-08-2014, 04:48 PM
Global MDS plot
http://abload.de/img/ane3ukiza.png
http://abload.de/img/ane3ukiza.png

Sapporo
03-08-2014, 05:36 PM
My results for a very accurate PCA based test by David:
ME=47%
ANE=37%
ENA=16%
Here are other results for comparison:
https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdGE5bjNCdl8yci0yNGN4TWk4M3NUW VE&usp=sharing#gid=0

This test was extremely accurate, but it was very time consuming. He had to run each sample once at a time. This new test is much less accurate, but still very nice, and gets the ballpark:
https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdC01RENtMGpRc3FHTGZpSUFKR3hHY 1E#gid=0

My new results for this Admixture run:
ME=54%
ANE=32%
ENA=12%
SSA=2%

As you can see, the Admixture based results are very nice, but less accurate. I have less ENA now, and I'm even scoring some SSA.

My score on the less accurate new test:

ME=50.955% or 51%
ANE=33.7008% or 34%
ENA=13.2214% or 13%
SSA=2.1228% or 2%

Pretty close to yours overall Sein. I expect my more "accurate" results would be roughly similar with higher ANE likely.

Sein
03-08-2014, 05:53 PM
My score on the less accurate new test:

ME=50.955% or 51%
ANE=33.7008% or 34%
ENA=13.2214% or 13%
SSA=2.1228% or 2%

Pretty close to yours overall Sein. I expect my more "accurate" results would be roughly similar with higher ANE likely.

On the more accurate PCA-based test, you might even be around 38%-40% ANE! I always expected your ANE to be higher than mine. You have more Gedrosia/Baloch, and more Northeastern European, so this a solid result. At the same time, your ENA is slightly higher, which I also think is reasonable. Mine is 11.68%. So, on the other one, you would probably be 18% ENA, as I was 16%.

Also, if you look at the spreadsheet, I think it proves that the Di Cristofaro Pashtuns are admixed with Tajiks and Uzbeks. Compare their ANE with the HGDP Pashtuns. The only unadmixed Pashtun is probably Pashtun10_17Af. Although, they don't seem very different from the HGDP Pashtuns in terms of ENA.

everest59
03-08-2014, 08:02 PM
All West Eurasian populations outside Europe lack WHG, so they should be modeled as mixtures between EEF and ANE. And EEF is almost 50% Basal Eurasian. Also, another note about drift statistics. I believe North Caucasians are lower on the list than Europeans, despite having more ANE than any European population. And South Central Asians are ahead of those populations, which themselves have more ANE than Europeans. If Lezgians share less drift than the HGDP Pastuns, and are around 28% ANE, than 35%-40% ANE for Pashtuns makes sense. This goes for other South Central Asians.

But I absolutely agree, we do need ancient DNA from South Asia. I think we can expect results from the Farmana team, so that is exciting.

It's mainly the West Asian portion that has basal Eurasian. WHG and ANE don't have them. Looking at the data, the Chechens are 51% Caucasian and 20% Baloch for a total of 71% West Asian (not counting the Northern Euro). So since they are more West Asian, they will be affected by basal Eurasian more.

Also, if you run a regression using just North Caucasian types, Pashtuns will still end up with a lower number.
So it seems to me that ASI is being captured by ANE a bit. We may be able to resolve the issue a little bit better once Stuttgart data is available.

If Paniyas are ending up as 24% ANE, there is something wrong.
Note: Even the Paniyas will have some ANE. I would say in the 5-6% range, because looking at the tribal types, it looks like they may have some ANE ancestry.

Sein
03-08-2014, 08:16 PM
It's mainly the West Asian portion that has basal Eurasian. WHG and ANE don't have them. Looking at the data, the Chechens are 51% Caucasian and 20% Baloch for a total of 71% West Asian (not counting the Northern Euro). So since they are more West Asian, they will be affected by basal Eurasian more.

Also, if you run a regression using just North Caucasian types, Pashtuns will still end up with a lower number.
So it seems to me that ASI is being captured by ANE a bit. We may be able to resolve the issue a little bit better once Stuttgart data is available.

If Paniyas are ending up as 24% ANE, there is something wrong.
Note: Even the Paniyas will have some ANE. I would say in the 5-6% range, because looking at the tribal types, it looks like they may have some ANE ancestry.

Your thinking in terms of ADMIXTURE components. And you've somewhat repeated what I've already said. EEF is around 50% Basal Eurasian. The other 50% is similar to WHG. No one outside Europe has WHG in addition to what they already possess in their EEF ancestry. As a result, you have to construe all non-European West Eurasians as EEF+ANE mixes (although, Bedouins don't have any ANE admixture, based on personal communication with Dr. Reich). That will depress the statistic, which is why the Adygei are behind Europeans, even though they probably have more ANE admixture. And despite the Adygei having more ANE admixture than Europeans, South-Central Asians are still higher on the list. That is a pretty good indication that South-Central Asians are around 35%-40% ANE.

And yes, the Paniya result is probably incorrect. But this experiment is less accurate, because it has underestimated ENA for everyone, and it has underestimated ANE for some populations. And this isn't my assessment, but David's. Here is what he said:

"it's not as accurate as the PCA based test I was running initially, but a lot quicker."

And in the more accurate ANE experiment, I was 36%-37% ANE. Sapporo would probably be 38%-40% ANE. So, with less accuracy, we are getting less ANE. But we are dealing with ADMIXTURE here, so this is no surprise.

But I agree, Stuttgart should settle things. The Iceman belongs to the same meta-population. So, perhaps you could try some things with the Iceman's genome?

everest59
03-08-2014, 08:26 PM
Your thinking in terms of ADMIXTURE components. And you've somewhat repeated what I've already said. EEF is around 50% Basal Eurasian. The other 50% is similar to WHG. No one outside Europe has WHG in addition to what they already possess in their EEF ancestry. As a result, you have to construe all non-European West Eurasians as EEF+ANE mixes. That will depress the statistic, which is why the Adygei are behind Europeans, even though they probably have more ANE admixture. And despite the Adygei having more ANE admixture than Europeans, South-Central Asians are still higher on the list. That is a pretty good indication that South-Central Asians are around 35%-40% ANE.

And yes, the Paniya result is probably incorrect. But this experiment is less accurate, because it has underestimated ENA for everyone, and it has underestimated ANE for some populations. And this isn't my assessment, but David's. Here is what he said:

"it's not as accurate as the PCA based test I was running initially, but a lot quicker."

And in the more accurate ANE experiment, I was 36%-37% ANE. Sapporo would probably be 38%-40% ANE. And our ENA scores should be higher.

Construing all non-European West Eurasians as EEF and ANE mixes may be a major issue. If you have a population that has another mix (ASI), you can't calculate ANE properly. You will need to know the level of ASI, and calculate certain populations as a 3-way mix, and maybe even 4-way. And that Moorjani calculation is highly unreliable.
I would be happy if it turns out South Asians have high ANE. However, I'm not sure if the numbers are accurate.
I will wait for Generalissimo to comment on inflated South Asian admixture in the ANE component.
P.S. There is a reason they avoided South Asians from their paper. Since we have a more complex history, the calculation will require more complexity.

Sein
03-08-2014, 08:35 PM
It's not a question of being "happy" that South Asians have high ANE. It's a matter of answering a few very important and big questions about Eurasian prehistory. If these results are correct, they tie in with a few other pieces of the puzzle. For example, the position of South Asia in regards to the differentiation of R lineages. And this raises a few more tantalizing questions, which may be answered by recent archeological work in South Asia. In terms of paradigm, everything is coming full circle, which is why I'm so confident about this. But, if I'm wrong, and you are right, it's not a big deal. We are all just trying to figure out how things took place before written records. That's all. Right now, I am confident in these results. And if it is a fluke, I will be the first person to give you credit for being right about this.

And I'm sorry if I came off as being dogmatic. :)

everest59
03-08-2014, 08:43 PM
Well, keep in mind that Generalissimo himself isn't confident in his numbers for South Asians. So I really am not confident tbh.

But let me add that, assuming his number are accurate, the correlation may be with Gedrosia/Baloch.
Let me edit that. This is not true as the Baloch don't have ANE as high as Pashtuns do.

I'm not saying the numbers are completely useless, since it does show more West Eurasian groups of South Asia (like the Pashtuns and Jats) with higher ANE. This is what we would expect.

Sein
03-08-2014, 08:55 PM
An excellent point. I really wish they could do some work on South Asia in the actual paper, but you are absolutely correct, I guess our part of the world is too complex.

And this a very perspicuous/brilliant observation, I didn't even think of this. For example, I tend to be 37% Gedrosia/Baloch, so there seems to be a strong correlation, assuming the results are correct.

You know, Everest, perhaps you could replicate the original, but time consuming, experiment on one of the Paniya samples? I'm sure David could easily explain his methodology to you. If the Paniya show substantial ANE admixture in that sort of experiment, you are probably right, we shouldn't be confident in these results. If not, I guess the results are solid. Do you think it can be done, if you eventually get the sufficient time and computational resources?

Edit: The Baloch also have an affinity to Southwest Asia (very strong, usually 10%), and to a lesser extent, Southern Europe. So, maybe that explains this.

everest59
03-08-2014, 08:59 PM
An excellent point. I really wish they could do some work on South Asia in the actual paper, but you are absolutely correct, I guess our part of the world is too complex.

And this a very perspicuous/brilliant observation, I didn't even think of this. For example, I tend to be 37% Gedrosia/Baloch, so there seems to be a strong correlation, assuming the results are correct.

You know, Everest, perhaps you could replicate the original, but time consuming, experiment on one of the Paniya samples? I'm sure David could easily explain his methodology to you. If the Paniya show substantial ANE admixture in that sort of experiment, you are probably right, we shouldn't be confident in these results. If not, I guess the results are solid. Do you think it can be done, if you eventually get the sufficient time and computational resources?

I was thinking of creating Zombies out of ANE, WHG and Stuttgart. I want to do this first with ANI/ASI. It seems the idea is pretty simple, just tedious. For example, if you have 14 different equations, you need to create 14 different files, at least with the method I'm gonna pursue. I just need some time first.

Sein
03-08-2014, 09:08 PM
That sounds exciting, this should prove to be very interesting.

everest59
03-08-2014, 09:20 PM
I know the process, but I've just been too lazy. As you can see, I have enough time as I've been typing on this board since morning...

Also, in my opinion, I have a hard time believing that South Asians don't have some WHG ancestry even if really low. Keep in mind that Mal'ta is a >20,000 year old individiual. Surely there was some mixing between WHG and ANE types, and probably some EEF before they arrived in the subcontinent.
Another possibility is that ANE came to the subcontinent before Indo-European speakers , in which case they had no WHG.
But then who brought Indo-European languages to South Asia?
I would have thought people with high ANE and some WHG (low but significant). Throw some West Asian ancestry in there.

Dr_McNinja
03-08-2014, 09:28 PM
Did ANE come from South Asia or to South Asia? I always figured the former. That would explain why it contains so much South Asian.

I strongly disagree with any Finno-Ugric/Uralic connection to South Asia after the development of ANE in North Eurasia without more evidence. Admixture confuses ENA-type admixture in the presence of Western European-like admixture and converts it into a Finnish-type (equivalent to even more boosted ANE/WHG). If you look at the other Admixture experimentation thread I got it to easily flow admixture between all the components (including between East and West Eurasian).

I don't see any evidence in history (or archaeological research) for a sizable ANE contribution to South Asia unless it piggybacked on the Indo-European migrations, but those can be primarily tracked by the EEF-type admixture still present in South Asians in limited traces. And it would be proportionate to the EEF signature. Even logically it would have to be treated as such otherwise you have no way of knowing if your ENA-type South Indian/East Asian is leaking into ANE.

Secondly, the data can only be manipulated so much. Genetic drift among populations has resulted in admixture that in principle can not be recognizable, so if we're somehow manipulating it into being recognizable, odds are we're just coming up with nonsense. So even though getting more ancient DNA will shed a lot of light on history and the "story" aspect of things, I doubt it will be of significant help for software analysis of modern people. The genomes of modern populations themselves are kind of too far gone from the ancestral components to be able to distinguish those components which resemble each other. So we could compare, say, Native American, African, Western European, East Asian, but distinguishing between Eurasian populations is always going to be an endeavor fraught with error.

Since EEF is the northern (relative to South Asia) admixture that has left a seemingly indelible trace, figuring out more accurate ways of measuring that and then coming up with formulas based on population averages is probably the best bet to track ANE. It's like the reverse of the Onge component, where calculating that meant you could double it and get a rough estimate of South Indian. But since ANE and ENA type components are mixed up together, this is a similar approach from the opposite end.

(EDIT by EEF I mean EEF/WHG mixture)

Sein
03-08-2014, 09:40 PM
I know the process, but I've just been too lazy. As you can see, I have enough time as I've been typing on this board since morning...

Also, in my opinion, I have a hard time believing that South Asians don't have some WHG ancestry even if really low. Keep in mind that Mal'ta is a >20,000 year old individiual. Surely there was some mixing between WHG and ANE types, and probably some EEF before they arrived in the subcontinent.
Another possibility is that ANE came to the subcontinent before Indo-European speakers , in which case they had no WHG.
But then who brought Indo-European languages to South Asia?
I would have thought people with high ANE and some WHG (low but significant). Throw some West Asian ancestry in there.

Lol, same here bro. I've been at this through my whole Saturday morning, and I only took a pause to buy some groceries. :P

This is a very pertinent point, and it's something I think the authors of the paper should address, before it gets published. Based on what I read, they find no evidence of any WHG admixture, outside of Europe. Although, WHG has to have an "Asiatic" equivalent, since about 50% of EEF is something similar to WHG. But for some reason, they think WHG is a European-only thing. That's just what the authors think. Theoretically, South and West Asians should have no WHG whatsoever. But the question is, are they correct?

Sein
03-08-2014, 09:46 PM
Did ANE come from South Asia or to South Asia? I always figured the former. That would explain why it contains so much South Asian.

I strongly disagree with any Finno-Ugric/Uralic connection to South Asia after the development of ANE in North Eurasia without more evidence. Admixture confuses ENA-type admixture in the presence of Western European-like admixture and converts it into a Finnish-type (equivalent to even more boosted ANE/WHG). If you look at the other Admixture experimentation thread I got it to easily flow admixture between all the components (including between East and West Eurasian).

I don't see any evidence in history (or archaeological research) for a sizable ANE contribution to South Asia unless it piggybacked on the Indo-European migrations, but those can be primarily tracked by the EEF-type admixture still present in South Asians in limited traces. And it would be proportionate to the EEF signature. Even logically it would have to be treated as such otherwise you have no way of knowing if your ENA-type South Indian/East Asian is leaking into ANE.

Secondly, the data can only be manipulated so much. Genetic drift among populations has resulted in admixture that in principle can not be recognizable, so if we're somehow manipulating it into being recognizable, odds are we're just coming up with nonsense. So even though getting more ancient DNA will shed a lot of light on history and the "story" aspect of things, I doubt it will be of significant help for software analysis of modern people. The genomes of modern populations themselves are kind of too far gone from the ancestral components to be able to distinguish those components which resemble each other. So we could compare, say, Native American, African, Western European, East Asian, but distinguishing between Eurasian populations is always going to be an endeavor fraught with error.

Since EEF is the northern (relative to South Asia) admixture that has left a seemingly indelible trace, figuring out more accurate ways of measuring that and then coming up with formulas based on population averages is probably the best bet to track ANE. It's like the reverse of the Onge component, where calculating that meant you could double it and get a rough estimate of South Indian. But since ANE and ENA type components are mixed up together, this is a similar approach from the opposite end.

(EDIT by EEF I mean EEF/WHG mixture)

I think we should remember that ANE might actually be somewhat of a misnomer. If ANE ancestry really peaks in South-Central Asia, followed by the Caucasus, and than in Northeastern Europe, we might need to rename it. And the name can confuse people. Some people assume that ANE is closer to East Eurasia. But this is somewhat of an illusion. MA-1 just lacks the African shift common to all living West Eurasians, so he is just genetically more "Eurasian". But he still belongs to the "Western" clade of peoples.

Sein
03-08-2014, 09:52 PM
Also, another interesting detail. The most common U linage in South Asia is U2. This is the same mtDNA haplogroup found for a 30,000-year-old hunter-gatherer found in Russia, Kostenki. I think U2 is the most common single mtDNA linage among Pashtuns. Ties in nicely with R1a1, and R2.

Dr_McNinja
03-08-2014, 09:56 PM
If ANE ancestry really peaks in South-Central Asia, followed by the Caucasus, and than in Northeastern Europe, we might need to rename it.This is a great idea. Especially if more ancient samples are found to further develop a baseline for ANE and it'd be interesting to track it outside of South Asia. But it still leaves the problem of sorting out ANE within South Asia where the signals have become muddled.

It'd be interesting to see how ANE's geographic spread correlates with components like West Asian/Gedrosia-Caucasus. I think we (Northwestern Indian/Pakistani/etc) have a lot more Gedrosian than current results suggest. I think the ANE admixture is confounding some of the Gedrosian and Western European into appearing Northeastern European and Caucasian. Kind of like connecting the dots when the dots aren't supposed to be genetically connected.

Also it's high time they find some remains from the early Indo-European days so we can get an actual sample of one of the possible sources of this admixture in South Asia. There's got to be some sitting around somewhere waiting to be found, that'd be exciting.

everest59
03-08-2014, 10:32 PM
Lol, same here bro. I've been at this through my whole Saturday morning, and I only took a pause to buy some groceries. :P

This is a very pertinent point, and it's something I think the authors of the paper should address, before it gets published. Based on what I read, they find no evidence of any WHG admixture, outside of Europe. Although, WHG has to have an "Asiatic" equivalent, since about 50% of EEF is something similar to WHG. But for some reason, they think WHG is a European-only thing. That's just what the authors think. Theoretically, South and West Asians should have no WHG whatsoever. But the question is, are they correct?

You're going to like this chart below. Very interesting, to say the least.

Sein
03-08-2014, 10:40 PM
This is quite interesting. PC1, very far away from all components, but closest to "South Asian". PC2, very far away from all components, but closest to "Gedrosia" and "Northern European".

everest59
03-08-2014, 10:46 PM
Yeah, that's what I noticed as well. Typically speaking, you see Northern European and Gedrosia facing what we perceive as Siberian/South Asian, but when I removed all the East Asianlike components, the Northern European and Gedrosia both are facing Mal'ta. The South Asian has some Mal'ta in it IMO.

Sein
03-08-2014, 11:17 PM
Yeah, that's what I noticed as well. Typically speaking, you see Northern European and Gedrosia facing what we perceive as Siberian/South Asian, but when I removed all the East Asianlike components, the Northern European and Gedrosia both are facing Mal'ta. The South Asian has some Mal'ta in it IMO.

Everest, could you try a supervised ADMIXTURE experiment with MA-1 as one component, La Brana as a separate component, Bedouins as another component, Han as another component, Yoruba as another component, and finally, Onge as another reference component? ADMIXTURE does seem to be less reliable, but we could still compare the results to David's run.

everest59
03-08-2014, 11:29 PM
I tried it yesterday. Got garbage results. I think I need to use files with better read depth.

alan
03-08-2014, 11:39 PM
It is certainly interesting that south Asia could be IE-ised without some WHG getting mixed in in the steppes. Maybe it is a sign that WHG wasnt common on the steppes at one time. One explanation would be if WHG is pretty well exclusively the imprint of the western expansion of hunters from the Magdallenian and their offshoot cultures and not of the eastern and epu-Gravettian groups who had lived in SE Europe, the Ukraine etc . If that was the case then you would expect WHG to be big in western, central and northern Europe but lacking in SE Europe, Ukraine, south Russia etc. If that was the case then the present WHG presence in those areas would be down to west-east movements into the area apparently after 3000BC perhaps partly linked to corded ware derived elements but also partly down to far later Slavic expansions.


I know the process, but I've just been too lazy. As you can see, I have enough time as I've been typing on this board since morning...

Also, in my opinion, I have a hard time believing that South Asians don't have some WHG ancestry even if really low. Keep in mind that Mal'ta is a >20,000 year old individiual. Surely there was some mixing between WHG and ANE types, and probably some EEF before they arrived in the subcontinent.
Another possibility is that ANE came to the subcontinent before Indo-European speakers , in which case they had no WHG.
But then who brought Indo-European languages to South Asia?
I would have thought people with high ANE and some WHG (low but significant). Throw some West Asian ancestry in there.

Dr_McNinja
03-08-2014, 11:52 PM
Everest, could you try a supervised ADMIXTURE experiment with MA-1 as one component, La Brana as a separate component, Bedouins as another component, Han as another component, Yoruba as another component, and finally, Onge as another reference component? ADMIXTURE does seem to be less reliable, but we could still compare the results to David's run.I was thinking of trying that. Where can you get the MA-1 genome data? The study data only has the samples they used for comparison, not MA-1 himself.

everest59
03-08-2014, 11:59 PM
I'll give it to you. Will email you data.

Sein
03-09-2014, 01:34 AM
I'd like to add a final data point to this discussion. If the peak in ANE ancestry in South Central Asia+Northwestern South Asian was a fluke, MA-1 wouldn't consistently display South Asian percentages in unsupervised ADMIXTURE runs. He also scores substantial Native American percentages. In fact, this was the angle used in the original paper. They reported that he looks like a mix between South Asians, Native Americans, and Northern Europeans. But he never shows East Asian admixture. I think this further augments these two particular experiments done by David, both the exceedingly nice PCA-based one, and the less accurate (but nonetheless, very robust) ADMIXTURE experiment. I think we should always have expected that the highest ANE ancestry would occur among Native Americans, and South Central Asians+Northwestern South Asians. Just my final thought for today.

Generalissimo
03-09-2014, 05:44 AM
I wouldn't recommend running some of the more time intensive methods, because you never know what you're getting unless you can run a lot of samples at once. For instance, you might spend three days running 500 samples, and then you'll stumble upon a group that doesn't fit the model, which means there might be a problem with the analysis, and you have to start again. So running ADMIXTURE in supervised mode, and using a few Karitiana to help MA-1 along, is really the only practical method of doing this, and definitely worth exploring further to see if the results can be improved. It only takes a few minutes to run a test like that.

One thing though, for anyone wanting to try this, just make sure that you only keep the least related Karitiana, otherwise they'll just create an Amerindian cluster, and also limit the markers to x2 read depth (they don't have to be pruned for LD though, because reducing markers creates more noise).

Honestly, I don't know what the deal is with the high ANE in South Asia. It might be a similar phenomenon to that in Eastern Europe, where ANE looks like it's through the roof when WHG ancestry isn't accounted for in the model. But even if that's the case, uniparental markers suggest that ANE should indeed be present across South Asia at relatively high frequencies. Is it a coincidence, for example, that in Eurasia ANE peaks in the Hindu Kush, where Y-DNA R1a-Z93 also reaches ridiculous frequencies?

In any case, I suspect we're missing a few things at this stage, and our assumptions are likely to change when we finally get ancient genomes from Eastern Europe and South Asia. For one, I'm sceptical that WHG and ANE were separate clades and that prehistoric Eastern Europeans were a mixture of WHG and ANE, rather than something in between (with some EEF thrown in as well). Secondly, I don't think we've heard the last word about what the so called Ancestral South Asian (ASI) component really is.

everest59
03-09-2014, 07:06 PM
Hey, have you tried your experiment using x1 read depth? It seems like the paper used original read depth. They were able to use >60k snp's.

parasar
03-09-2014, 09:03 PM
...

Also, I believe South Indians didn't register any ANE in the PCA test (I'll ask him about this).
Did you get a response on this?

My own feeling based on Genetiker's analysis is that ANE is also South Indian ("Veddoid") derived. In the MA1 timeframe it had already transitioned towards WHG by admixture or both share significant common ancestry (see mtDNA U). So MA1 should share more ancient but overall fewer alleles with southern Indian Veddoid populations.

Sein
03-09-2014, 09:18 PM
Did you get a response on this?

My own feeling based on Genetiker's analysis is that ANE is also South Indian ("Veddoid") derived. In the MA1 timeframe it had already transitioned towards WHG by admixture or both share significant common ancestry (see mtDNA U). So MA1 should share more ancient but overall fewer alleles with southern Indian Veddoid populations.

This is a very interesting notion.

I never got to ask him this question. But, I'm sure he'll see this post, and address this question.

icebreaker
03-09-2014, 09:44 PM
My results:
EEF 86,43822162
WHG 3,94986869
ANE 9,611909692

ZephyrousMandaru
03-09-2014, 11:08 PM
It appears there's been an update for this test.

ANE (MA-1) 13.53%
ME (South Bedouin) 86.02%
ENA (Han) 0%
SSA (Yoruba) 0.4%

MfA
03-09-2014, 11:18 PM
It appears there's been an update for this test.

ANE (MA-1) 13.53%
ME (South Bedouin) 86.02%
ENA (Han) 0%
SSA (Yoruba) 0.4%

http://eurogenes.blogspot.co.uk/2014/03/ancient-north-eurasian-ane-levels.html

You are late bro :p

Ancient North Eurasian (ANE) levels across Asia (http://eurogenes.blogspot.com/2014/03/ancient-north-eurasian-ane-levels.html)

My results:
0.196525 ANE(MA-1)
0.803455 ME(South Bedouin)
0.00001 ENA(Han)
0.00001 SSA(Yoruba)

Davidski gonna replace Yorubas with Pygmies to fix the issue of South Central Asians score some SSA.. While updating, adding Turks would be informative aswell, for some reason they're left out.

ZephyrousMandaru
03-09-2014, 11:28 PM
You are late bro :p


Davidski gonna replace Yorubas with Pygmies to fix the issue of South Central Asians score some SSA.. While updating, adding Turks would be informative aswell, for some reason they're left out.

Since an ENA component is present, it might be able to separate it from the ANE. But the fact that East Asian admixture goes into ANE is suspect, what gives ANE it's eastward pull?

Sein
03-10-2014, 03:49 AM
David just updated the entry. A comparison between my first ADMIXTURE results, and new results with around 100K SNPs (the first run had 53K SNPs, I think).

K4_ADMIXTURE_Test
ME=53.85%
ANE=31.94%
ENA=11.68%
SSA=2.51%

K4_ADMIXTURE_Test4_x1_depth_read
ME=53.18%
ANE=32.30%
ENA=12.22%
SSA=2.28%

Almost identical, so the results seem pretty robust across runs.

Also, I just realized that EEF is a European-specific construct. Just goes to show how perceptive I am. It took quite a while for that fact to sink in. :P

So, if we are dealing with only "Near Eastern" ancestry for South Asians, without any WHG, I think people from South Central Asia and Northwestern South Asia probably are around 35%-40% ANE. But perhaps Everest is right, maybe there is some WHG in South Asia? Probably not, but now, I am much more willing to accept the existence of some WHG-like admixture in South Asia. But as far as the paper goes, I think the working assumption is that WHG is only specific to Europe.

(Anyone with a better understanding of the paper, please do chime in)

Generalissimo
03-10-2014, 10:50 AM
My own feeling based on Genetiker's analysis is that ANE is also South Indian ("Veddoid") derived. In the MA1 timeframe it had already transitioned towards WHG by admixture or both share significant common ancestry (see mtDNA U). So MA1 should share more ancient but overall fewer alleles with southern Indian Veddoid populations.

Yes, the South Indians show ANE admix in the other test, at around 20-25%.

But anyway, what do you make of the fact that some Malays and Cambodians have above noise ME, but no ANE? Didn't they get that ME from India?

MfA
03-10-2014, 12:15 PM
Ami, Atayal genomes out yet? which form their own component at k19 and k20, shows up in Chinese, Japanese, Cambodians, etc[1] (http://biorxiv.org/content/biorxiv/suppl/2013/12/23/001552.DC1/001552-2.pdf).. Might be useful using Atayal as reference to ENA instead of Han..

Dr_McNinja
03-10-2014, 12:25 PM
I posted some supervised admixture experimentation results in the other thread. There's some kind of link between European and South Indians/SE-Asian components. If you introduce a North Indian component (Kshatriya and Uttar Pradesh Brahmin populations), European populations will start scoring a little bit of it. And if you don't include it, they'll score a little Southeast Asian. It's hard to tell because it seems like it gets split up among Harappa-type South Indian, SE-Asian, and Papuan in calculators with a lot of components but it's about 1-2% on average and it goes up the more Indian populations you introduce (like North Indian and South Indian together).

The Northern Europeans are scoring ~5% Uralic and Northeast Europeans over 10% Uralic (I have no NE-Euro component in there, just Basque and Sardinian), so it could be bleed off from that. I have to do another run to check fst distances.

alan
03-10-2014, 12:35 PM
I am a bit confused by the abbreviations. I assume these are Asian autosomal DNA terms. What are:

ME
ANE
ENA
SSA


David just updated the entry. A comparison between my first ADMIXTURE results, and new results with around 100K SNPs (the first run had 53K SNPs, I think).

K4_ADMIXTURE_Test
ME=53.85%
ANE=31.94%
ENA=11.68%
SSA=2.51%

K4_ADMIXTURE_Test4_x1_depth_read
ME=53.18%
ANE=32.30%
ENA=12.22%
SSA=2.28%

Almost identical, so the results seem pretty robust across runs.

Also, I just realized that EEF is a European-specific construct. Just goes to show how perceptive I am. It took quite a while for that fact to sink in. :P

So, if we are dealing with only "Near Eastern" ancestry for South Asians, without any WHG, I think people from South Central Asia and Northwestern South Asia probably are around 35%-40% ANE. But perhaps Everest is right, maybe there is some WHG in South Asia? Probably not, but now, I am much more willing to accept the existence of some WHG-like admixture in South Asia. But as far as the paper goes, I think the working assumption is that WHG is only specific to Europe.

(Anyone with a better understanding of the paper, please do chime in)

Dr_McNinja
03-10-2014, 12:45 PM
I am a bit confused by the abbreviations. I assume these are Asian autosomal DNA terms. What are:

ME
ANE
ENA
SSAMiddle Eastern
Ancestral North Eurasian
Eastern Non-African
Sub-Saharan African

alan
03-10-2014, 05:20 PM
Thank you for that. Can I also ask if ancient north Eurasian is the same as Ancient north Asian as in Mal'ta?


Middle Eastern
Ancestral North Eurasian
Eastern Non-African
Sub-Saharan African

Dr_McNinja
03-10-2014, 06:29 PM
Thank you for that. Can I also ask if ancient north Eurasian is the same as Ancient north Asian as in Mal'ta?
Yes, Mal'ta is ANE

Sein
03-10-2014, 06:50 PM
Thank you for that. Can I also ask if ancient north Eurasian is the same as Ancient north Asian as in Mal'ta?

I think that in Europe, the highest ANE occurs among Estonians, Hungarians, Scots, Czech, and Greeks. I believe the numbers are 18.7%, 17.4%, 17.1%, 16.3%, and 16.1%. In the paper, the highest ANE they find in Eurasia is almost 29%, for the Lezgian people of the northern Caucasus. For comparison, when the values resemble that of the actual paper, David found me to be 36% ANE. Many people from my neck of the woods could get up to 38%-40%. So in Eurasia, the highest is in South Central Asia+Northwestern South Asia, followed by a more moderate peak in the northern Caucasus and West Asia, and finally, a more diluted but nonetheless appreciable peak among diverse European populations. My knowledge of European genetics (and history) is rather patchy, so I'm not sure what the European groups with high ANE share in terms of genetics and history?

Dr_McNinja
03-10-2014, 07:54 PM
I think that in Europe, the highest ANE occurs among Estonians, Hungarians, Scots, Czech, and Greeks. I believe the numbers are 18.7%, 17.4%, 17.1%, 16.3%, and 16.1%. In the paper, the highest ANE they find in Eurasia is almost 29%, for the Lezgian people of the northern Caucasus. For comparison, when the values resemble that of the actual paper, David found me to be 36% ANE. Many people from my neck of the woods could get up to 38%-40%. So in Eurasia, the highest is in South Central Asia+Northwestern South Asia, followed by a more moderate peak in the northern Caucasus and West Asia, and finally, a more diluted but nonetheless appreciable peak among diverse European populations. My knowledge of European genetics (and history) is rather patchy, so I'm not sure what the European groups with high ANE share in terms of genetics and history?Which European populations have low ANE?

Sein
03-10-2014, 08:17 PM
Which European populations have low ANE?

Based on Table S12.8, the Sardinians, Pais_Vasco, Basque, and Spanish. Respectively, 4.1%, 9.6%, 9.6%, and 11.5%. As you can see, these are also the populations that appear the least East Eurasian, and most West Eurasian, out of all Europeans. In fact, I think the Sardinians are the most "western" population of all. If the results we are seeing turn out to be correct, our heavy ANE admixture should make us somewhat allergic to these populations on an east-west axis, even though we aren't really highly admixed with East Eurasians (although, deciding how much is really "highly admixed" is a highly subjective topic. In my view, 30% of any ancestry is significant, 20% is moderate, and anything around 10% is minor, but that's all pretty subjective). If I add the African admixture to my ENA score in David's run, on the assumption that it is an ASI signal, I get around 13%-14% ENA admixture for myself. Most people from our part of the world get 9%-16%, excepting the Burusho. One HGDP Pashtun is only 7%, and some of the Baloch are only 4%-5% ENA. So, I think our ANE ancestry is pushing us way east of where we should cluster. Then again, these results could be a fluke. It'll be exciting to see if David obtains different results with the Onge, and with the Stuttgart and Loschbar genomes.

Edit: I think David has noted the possibility that these low ANE percentages might not actually be ANE, but WHG, especially for the Basques and Sardinians. Also, David just added Turks. They are pretty diverse in terms of ENA, but quite a few are around as ENA as the Baloch. And some almost reach Pashtun+Punjabi levels of ENA (one is 12% ENA), but they do cluster much to our west on PCA plots.

everest59
03-10-2014, 10:51 PM
I have no doubt now that the South Asian component of Dienekes' K12b has some Ancestral North Eurasian. What I'm doing is calculating shared drift with Malta. This was done on only around 16k snp's, but it shows something very interesting:

https://drive.google.com/file/d/0B3vEDdpZDjUpQ0FkSl96bkkzQ3M/edit?usp=sharing

The South Asian component of K12b actually has higher ANE than both the Caucasus and the Atlantic_Med. This could be the basal Eurasian effect. But IMO, South Asian will have some. The South Asian component peaks in tribal types, which means they are going to have some ANE as well. It's not zero.

Basically, all Western Eurasian components of K12b have some ANE with SW-Asian having the lowest (probably has some African in there, lowering the actual shared drift) and Northern European clearly the highest. Gedrosia is up there as well. East Asian will have some ANE, probably because some Karitiana-like elements got inferred as ANE by ADMIXTURE.

I'm not using a ton of snp's here because the Mal'ta file did not have that many common SNP's with K12b. I'm using read depth of x2, which lowered snp count some more.

Sein
03-10-2014, 11:07 PM
I have no doubt now that the South Asian component of Dienekes' K12b has some Ancestral North Eurasian. What I'm doing is calculating shared drift with Malta. This was done on only around 16k snp's, but it shows something very interesting:

https://drive.google.com/file/d/0B3vEDdpZDjUpQ0FkSl96bkkzQ3M/edit?usp=sharing

The South Asian component of K12b actually has higher ANE than both the Caucasus and the Atlantic_Med. This could be the basal Eurasian effect. But IMO, South Asian will have some. The South Asian component peaks in tribal types, which means they are going to have some ANE as well. It's not zero.

Basically, all Western Eurasian components of K12b have some ANE with SW-Asian having the lowest (probably has some African in there, lowering the actual shared drift) and Northern European clearly the highest. Gedrosia is up there as well. East Asian will have some ANE, probably because some Karitiana-like elements got inferred as ANE by ADMIXTURE.

I'm not using a ton of snp's here because the Mal'ta file did not have that many common SNP's with K12b. I'm using read depth of x2, which lowered snp count some more.

Thanks everest! This is a very interesting output.

Sapporo shares the most drift with MA-1 out of everyone, even more than the Northern European component (which is probably composed of a lot of ANE and WHG, and might even be very low on "Basal Eurasian"). I share less than the Northern European component, but I'm still ahead of the Gedrosia component. It's pretty suggestive that Sapporo shares more drift than the Northern European component, and I'm pretty close, despite the fact that (I think) we all probably have some seriously heavy "Basal Eurasian" admixture. Also, you definitely have more ANE admixture than West Asians. A lot of food for thought here.

everest59
03-10-2014, 11:13 PM
I would like to do it one more time using higher snp count. These things seem to be very sensitive.

Sein
03-10-2014, 11:17 PM
I would like to do it one more time using higher snp count. These things seem to be very sensitive.

A very good point. This should prove to be very interesting.

everest59
03-10-2014, 11:19 PM
A very good point. This should prove to be very interesting.

I don't think anything will change by a whole lot though. Just gonna make sure.

Sein
03-10-2014, 11:35 PM
I don't think anything will change by a whole lot though. Just gonna make sure.

Probably so.

It's pretty interesting that DMXX and Icebreaker share the most drift with MA-1, out of all West Asians. I believe DMXX has a higher Northern European, Gedrosia, and South Asian score relative to other West Asians, so that makes sense. I'm less familiar with Icebreaker's results, but I think he had an affinity to South Asia in his Family Finder results. Also, I think he resembles DMXX in regards to the aforementioned components (perhaps he could chime in). But, it is puzzling that NK19191, who I believe is also of Iranian background, shares much less than DMXX.

everest59
03-10-2014, 11:41 PM
That could be the low snp effect. Or maybe they are from different regions.

everest59
03-11-2014, 02:11 AM
The text file I linked above did not appear correctly when I switched to Windows. So I'll just paste it here:
parameter file: parqp3Pop
### THE INPUT PARAMETERS
##PARAMETER NAME: VALUE
indivname: Analyze2.ind
snpname: Analyze2.snp
genotypename: Analyze2.geno
popfilename: list_qp3Pop
## qp3Pop version: 204
nplist: 22
number of blocks for block jackknife: 521
snps: 16532
Source 1 Source 2 Target f_3 std. err Z SNPs
result: Malta Atlantic_Med Yoruba 0.170672 0.004225 40.395 16025
result: Malta Caucasus Yoruba 0.171615 0.004104 41.816 16025
result: Malta Gedrosia Yoruba 0.182254 0.004067 44.809 16033
result: Malta Northern_European Yoruba 0.187017 0.004202 44.503 16044
result: Malta South_Asian Yoruba 0.173849 0.004070 42.720 16014
result: Malta SW_Asian Yoruba 0.161680 0.004077 39.658 15997
result: Malta Sub_Saharan Yoruba 0.122563 0.003919 31.275 15695
result: Malta SE_Asian Yoruba 0.169820 0.004228 40.163 15988
result: Malta East_Asian Yoruba 0.172864 0.004385 39.425 15999
result: Malta Siberian Yoruba 0.180088 0.004381 41.103 16008
result: Malta NW_African Yoruba 0.150337 0.003764 39.943 15988
result: Malta E_African Yoruba 0.131627 0.003772 34.898 15907
result: Malta .Everest Yoruba 0.177387 0.005581 31.785 15743
result: Malta .Sein Yoruba 0.182808 0.005743 31.834 15754
result: Malta .Sapporo Yoruba 0.189016 0.005684 33.256 15767
result: Malta .Mcninja Yoruba 0.181600 0.005549 32.724 15763
result: Malta .Humanist Yoruba 0.162474 0.005829 27.874 15729
result: Malta .mfa Yoruba 0.168396 0.005562 30.276 15739
result: Malta .DMXX Yoruba 0.171073 0.005310 32.220 15730
result: Malta .Zephyrous Yoruba 0.165376 0.005402 30.613 15728
result: Malta .NK19 Yoruba 0.164836 0.005871 28.077 15733
result: Malta .Icebreaker Yoruba 0.174090 0.005868 29.668 15739
##end of qp3Pop

ZephyrousMandaru
03-11-2014, 07:10 AM
The text file I linked above did not appear correctly when I switched to Windows. So I'll just paste it here:
parameter file: parqp3Pop
### THE INPUT PARAMETERS
##PARAMETER NAME: VALUE
indivname: Analyze2.ind
snpname: Analyze2.snp
genotypename: Analyze2.geno
popfilename: list_qp3Pop
## qp3Pop version: 204
nplist: 22
number of blocks for block jackknife: 521
snps: 16532
Source 1 Source 2 Target f_3 std. err Z SNPs
result: Malta Atlantic_Med Yoruba 0.170672 0.004225 40.395 16025
result: Malta Caucasus Yoruba 0.171615 0.004104 41.816 16025
result: Malta Gedrosia Yoruba 0.182254 0.004067 44.809 16033
result: Malta Northern_European Yoruba 0.187017 0.004202 44.503 16044
result: Malta South_Asian Yoruba 0.173849 0.004070 42.720 16014
result: Malta SW_Asian Yoruba 0.161680 0.004077 39.658 15997
result: Malta Sub_Saharan Yoruba 0.122563 0.003919 31.275 15695
result: Malta SE_Asian Yoruba 0.169820 0.004228 40.163 15988
result: Malta East_Asian Yoruba 0.172864 0.004385 39.425 15999
result: Malta Siberian Yoruba 0.180088 0.004381 41.103 16008
result: Malta NW_African Yoruba 0.150337 0.003764 39.943 15988
result: Malta E_African Yoruba 0.131627 0.003772 34.898 15907
result: Malta .Everest Yoruba 0.177387 0.005581 31.785 15743
result: Malta .Sein Yoruba 0.182808 0.005743 31.834 15754
result: Malta .Sapporo Yoruba 0.189016 0.005684 33.256 15767
result: Malta .Mcninja Yoruba 0.181600 0.005549 32.724 15763
result: Malta .Humanist Yoruba 0.162474 0.005829 27.874 15729
result: Malta .mfa Yoruba 0.168396 0.005562 30.276 15739
result: Malta .DMXX Yoruba 0.171073 0.005310 32.220 15730
result: Malta .Zephyrous Yoruba 0.165376 0.005402 30.613 15728
result: Malta .NK19 Yoruba 0.164836 0.005871 28.077 15733
result: Malta .Icebreaker Yoruba 0.174090 0.005868 29.668 15739
##end of qp3Pop

Hey Everest, thanks for the data. How would I go by interpreting the results?

everest59
03-11-2014, 09:43 AM
Zeph you need to look at the first number, which in your case is .165. The higher that number the higher shared drift, which means more ANE admixture. However, I am also wondering how much African the SW Asian component has. It will dilute the number.

Dr_McNinja
03-11-2014, 12:26 PM
That's interesting. Sapporo had the highest match in my run as well, followed by another Jatt. And it appears South Asian is closer than even Caucasus (EDIT: And East Asian is closer than I expected too).

Hanna
03-11-2014, 12:53 PM
adding Turks would be informative aswell, for some reason they're left out.
I asked Polako to add Turkish samples and he did:

https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdG1KU0RzN2dVMHFsYVM3NHpsYXNVT Hc#gid=0

My results:

ANE(MA-1):19.29%
ME(South Bedouin): 77.94%
ENA(Han): 2.75%
SSA(Pygmy): 0.00%

parasar
03-11-2014, 02:56 PM
Mine should be in this range:

Brahmin_UP BR052
ANE 0.3378
ME 0.441849
ENA 0.197795
SSA 0.022557

Gujarati NA20854
ANE 0.338623
ME 0.442495
ENA 0.184612
SSA 0.034271

everest59
03-11-2014, 11:07 PM
Okay, I gave a shot at calculating ANE admixture. The first number is the Admixture %. I assumed SW_Asian as having no ANE. What you'll notice is that the South Asian component is only 12.9% ANE, so we must be on the right track. :

result: Sub_Saharan Karitiana Gedrosia SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.141917 0.014296 9.927
result: Sub_Saharan Karitiana Caucasus SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.099852 0.012808 7.796
result: Sub_Saharan Karitiana Northern_European SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.152819 0.013831 11.049
result: Sub_Saharan Karitiana South_Asian SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.129498 0.015682 8.258
result: Sub_Saharan Karitiana Atlantic_Med SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.064394 0.012156 5.297
result: Sub_Saharan Karitiana .Everest SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.305097 0.034917 8.738
result: Sub_Saharan Karitiana .Sein SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.271508 0.032390 8.383
result: Sub_Saharan Karitiana .Sapporo SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.245000 0.030527 8.026
result: Sub_Saharan Karitiana .Mcninja SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.255401 0.031986 7.985
result: Sub_Saharan Karitiana .Icebreaker SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.232185 0.032883 7.061
result: Sub_Saharan Karitiana .NK19 SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.077665 0.030869 2.516
result: Sub_Saharan Karitiana .mfa SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.140816 0.031695 4.443
result: Sub_Saharan Karitiana .DMXX SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.176674 0.031387 5.629
result: Sub_Saharan Karitiana .Zephyrous SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.065704 0.030801 2.133
result: Sub_Saharan Karitiana .Humanist SW_Asian : Sub_Saharan Karitiana Malta SW_Asian 0.150711 0.031172 4.835


However, it seems the East Asian inflates the value a little bit. Also, assuming that the SW Asian has some African, the actual admixture will be higher.
It was worth a shot...

parasar
03-15-2014, 03:36 AM
My results for a very accurate PCA based test by David:
ME=47.3%
ANE=36.2%
ENA=16.5%
Here are other results for comparison:
https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdGE5bjNCdl8yci0yNGN4TWk4M3NUW VE&usp=sharing#gid=0

This test was extremely accurate, but it was very time consuming. He had to run each sample once at a time. This new test is much less accurate, but still very nice, and gets the ballpark:
https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdC01RENtMGpRc3FHTGZpSUFKR3hHY 1E#gid=0

My new results for this Admixture run:
ME=53.8%
ANE=31.9%
ENA=11.6%
SSA=2.5%

As you can see, the Admixture based results are very nice, but less accurate. I have less ENA now, and I'm even scoring some SSA.


My score on the less accurate new test:

ME=50.955% or 51%
ANE=33.7008% or 34%
ENA=13.2214% or 13%
SSA=2.1228% or 2%

Pretty close to yours overall Sein. I expect my more "accurate" results would be roughly similar with higher ANE likely.

Thanks to Davidski of Eurogenes Blog, my results:

ANE 0.337197
ME 0.445173
ENA 0.192069
SSA 0.025561

parasar
03-15-2014, 04:14 PM
Mine should be in this range:

Brahmin_UP BR052
ANE 0.3378
ME 0.441849
ENA 0.197795
SSA 0.022557

Gujarati NA20854
ANE 0.338623
ME 0.442495
ENA 0.184612
SSA 0.034271
https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdC01RENtMGpRc3FHTGZpSUFKR3hHY 1E#gid=0

Pakhtoon
Sein
ANE=31.9%
ME=53.8%
ENA=11.6%
SSA=2.5%

North Western India
Sapporo
ANE=33.7008% or 34%
ME=50.955% or 51%
ENA=13.2214% or 13%
SSA=2.1228% or 2%

Western India
Gujarati NA20854
ANE 0.331747
ME 0.439002
ENA 0.190722
SSA 0.038529

North Central India
Brahmin_UP BR052
ANE 0.329205
ME 0.452896
ENA 0.199666
SSA 0.018232

Eastern India
Mine:
ANE 0.337197
ME 0.445173
ENA 0.192069
SSA 0.025561

The only trend appears to be that ME reduces along NW to SE cline. ANE is almost constant. The slightly higher ANE in the Burusho may either point to a non-IE origin nature of ANE with the isolate Burusho retaining more of their composition and language or perhaps a second shot of ANE into the Burusho from east asia of the kind reported by Lazaridis et al explaining elevated levels of ANE in some populations:


gene flow into some northeastern Europeans after the initial ANE admixture, and may be related to the fact that Y-chromosome haplogroup N30, 31 is shared between Siberian and northeastern Europeans
32, 33 but not with western Europeans. There may in Downloaded from http://biorxiv.org/ on March 15, 201411 fact be multiple layers of Siberian gene flow into northeastern Europe after the initial ANE gene flow, as our analyses reported in SI 12 show that some Mordovians, Russians and Chuvash have Siberian-related admixture that is significantly more recent than that in Finns (SI12).

Dr_McNinja
03-15-2014, 05:51 PM
https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdC01RENtMGpRc3FHTGZpSUFKR3hHY 1E#gid=0

Pakhtoon
Sein
ANE=31.9%
ME=53.8%
ENA=11.6%
SSA=2.5%

North Western India
Sapporo
ANE=33.7008% or 34%
ME=50.955% or 51%
ENA=13.2214% or 13%
SSA=2.1228% or 2%

Western India
Gujarati NA20854
ANE 0.331747
ME 0.439002
ENA 0.190722
SSA 0.038529

North Central India
Brahmin_UP BR052
ANE 0.329205
ME 0.452896
ENA 0.199666
SSA 0.018232

Eastern India
Mine:
ANE 0.337197
ME 0.445173
ENA 0.192069
SSA 0.025561

The only trend appears to be that ME reduces along NW to SE cline. ANE is almost constant. The slightly higher ANE in the Burusho may either point to a non-IE origin nature of ANE with the isolate Burusho retaining more of their composition and language or perhaps a second shot of ANE into the Burusho from east asia of the kind reported by Lazaridis et al explaining elevated levels of ANE in some populations:Burusho just might be due to isolation and drift.

I think ANE was there long before Indo-Europeans. I still think it emerged in South Central Asia very early and then future Central Asians, Siberians, Amerindians, etc branched off from it and went north while the original ANE people stayed behind to become the indigenous inhabitants of South Central Asia. The first people had to have come from somewhere after all and the spread is throughout the subcontinent.

Sein
03-15-2014, 05:58 PM
Burusho just might be due to isolation and drift.

I think ANE was there long before Indo-Europeans. I still think it emerged in South Central Asia very early and then future Central Asians, Siberians, Amerindians, etc branched off from it and went north while the original ANE people stayed behind to become the indigenous inhabitants of South Central Asia. The first people had to have come from somewhere after all and the spread is throughout the subcontinent.

I'm beginning to think that ANE were the first West Eurasians to enter South Asia. Both they and the ASI populations must have been hunter gatherers, and they easily mixed in northern South Asia. This is actually the explanation the authors provide in the paper, for Northern Caucasians. They note that Northern Caucasians have more ANE than any European population, so it is possible that ANE hunter gatherers occupied all of West Asia, and eventually became submerged in an agricultural revolution brought about by "Ancient Near Easterners", who were rich in African-like "Basal Eurasian" (Natufians? They looked somewhat African, so perhaps they represent the Levantine twig of these "Ancient Near Easterners"?). And if South Central Asians+Northwestern South Asians have more ANE than Northern Caucasians, and are almost at Native American levels, perhaps we could apply the same logic to our part of the world? This is basically my extremely tentative/speculative picture of the situation:


ANE hunter gatherer populations constitute a continuum across the West Asia, from Anatolia to western Pakistan, and into the Caucasus. They encounter ASI hunter gatherers in Northwestern South Asia, and across the generations, mixture takes place. Then, during the Neolithic, "Ancient Near Easterners" absorb ANE hunter gatherers across West Asia, and eventually, they reach South Asia. They get absorbed into the ANE+ASI mixed populations of Northwestern South Asia, and these newly mixed populations create the Indus Valley Civilization. These "Harappans", who are mixtures between "Near Easterners"+ANE+ASI, (but probably mostly ANE+ASI, with heavy "Near Eastern" admixture) expand further into South Asia, where mixture slowly occurs with more peninsular ASI populations. Many South Indian tribal people can trace much of their West Eurasian ancestry to these processes. Due to climactic changes and overpopulation, the Indus Valley Civilization collapses. Due to the same exact climactic changes, Indo-Iranian populations rich in ANE enter South Asia via the northwestern passes in contemporary Afghanistan and northwestern Pakistan. They find a socio-political vacuum, devastated/empty urban complexes, and a highly reduced population. Mixture is inevitable, and Northwestern South Asians get more ANE admixture. More admixture with ANE-rich populations occurs across the centuries, with Scythian people, Kushans, and other similar groups who migrate into South Asia. Finally, "Near Eastern" ancestry slowly increases in the northwest, due to low-level gene flow from West Asia, across the centuries. This ANE and "Near Eastern" admixture slowly diffuses across South Asia, but isolated communities on the edge of Sanskritic culture are shielded from further West Eurasian admixture (for example, the Paniya). At the end of it all, people in the northwest basically become West Eurasians (because West Eurasian gene flow is constantly diffusing throughout the area), and people in the South of India basically become West Eurasian+East Eurasian hybrids, but slightly more West Eurasian. But there is more. Rice came from the northeast, so expansions occur across what is now Bangladesh. Austro-Asiatic speakers come from Southeast Asia, and mix with populations who are already ANE+ME+ASI. And amongst the Iranic peoples of Northwestern South Asia (Pashtuns and Baloch), there are some intense episodes of West Asian gene flow (if you read the Hellenthal et al. paper, it seems something happened to Pashtuns in the medieval era, some very heavy West Asian gene flow occurs at that time). Later, Turkic-Mongolic admixture occurs among Pashtun tribes, and African admixture occurs among the Balochistani populations. All in all, a very rich history.

This is basically my amateur description of South Asian genetic history. The question is, how much of this is right? I am no expert, and much of this is oversimplification. Any suggestions on how to improve this account, or if I made any errors?

Dr_McNinja
03-15-2014, 07:46 PM
I'm beginning to think that ANE were the first West Eurasians to enter South Asia. Both they and the ASI populations must have been hunter gatherers, and they easily mixed in northern South Asia. This is actually the explanation the authors provide in the paper, for Northern Caucasians. They note that Northern Caucasians have more ANE than any European population, so it is possible that ANE hunter gatherers occupied all of West Asia, and eventually became submerged in an agricultural revolution brought about by "Ancient Near Easterners", who were rich in African-like "Basal Eurasian" (Natufians? They looked somewhat African, so perhaps they represent the Levantine twig of these "Ancient Near Easterners"?). And if South Central Asians+Northwestern South Asians have more ANE than Northern Caucasians, and are almost at Native American levels, perhaps we could apply the same logic to our part of the world? This is basically my extremely tentative/speculative picture of the situation:


ANE hunter gatherer populations constitute a continuum across the West Asia, from Anatolia to western Pakistan, and into the Caucasus. They encounter ASI hunter gatherers in Northwestern South Asia, and across the generations, mixture takes place. Then, during the Neolithic, "Ancient Near Easterners" absorb ANE hunter gatherers across West Asia, and eventually, they reach South Asia. They get absorbed into the ANE+ASI mixed populations of Northwestern South Asia, and these newly mixed populations create the Indus Valley Civilization. These "Harappans", who are mixtures between "Near Easterners"+ANE+ASI, (but probably mostly ANE+ASI, with heavy "Near Eastern admixture") expand further into South Asia, where mixture slowly occurs with more peninsular ASI populations. Many South Indian tribal people can trace much of their West Eurasian ancestry to these processes. Due to climactic changes and overpopulation, the Indus Valley Civilization collapses. Due to the same exact climactic changes, Indo-Iranian populations rich in ANE enter South Asia via the northwestern passes in contemporary Afghanistan and northwestern Pakistan. They find a socio-political vacuum, devastated/empty urban complexes, and a highly reduced population. Mixture is inevitable, and Northwestern South Asians get more ANE admixture. More admixture with ANE-rich populations occurs across the centuries, with Scythian people, Kushans, and other similar groups who migrate into South Asia. Finally, "Near Eastern" ancestry slowly increases in the northwest, due to low-level gene flow from West Asia, across the centuries. This ANE and "Near Eastern" admixture slowly diffuses across South Asia, but isolated communities on the edge of Sanskritic culture are shielded from further West Eurasian admixture (for example, the Paniya). At the end of it all, people in the northwest basically become West Eurasians (because West Eurasian gene flow is constantly diffusing throughout the area), and people in the South of India basically become West Eurasian+East Eurasian hybrids, but slightly more West Eurasian. But there is more. Rice came from the northeast, so expansions occur across what is now Bangladesh. Austro-Asiatic speakers come from Southeast Asia, and mix with populations who are already ANE+ME+ASI. And amongst the Iranic peoples of Northwestern South Asia (Pashtuns and Baloch), there are some intense episodes of West Asian gene flow (if you read the Hellenthal et al. paper, it seems something happened to Pashtuns in the medieval era, some very heavy West Asian gene flow occurs at that time). Later, Turkic-Mongolic admixture occurs among Pashtun tribes, and African admixture occurs among the Balochistani populations. All in all, a very rich history.

This is basically my highly amateur description of South Asian genetic history. The question is, how much of this is right? I am no expert, and much of this is oversimplification. Any suggestions on how to improve this account, or if I made any errors?
Now correlate it to human migration out of Africa and the emergence/branching of major haplogroups like R and subclades. :P

Sapporo
03-15-2014, 08:26 PM
My official ANE results on David's run:

ANE: 34.317%
ME: 50.1756%
ENA: 13.4809%
SSA: 2.0265%

https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdG1KU0RzN2dVMHFsYVM3NHpsYXNVT Hc#gid=0

Sein
03-15-2014, 08:31 PM
Now correlate it to human migration out of Africa and the emergence/branching of major haplogroups like R and subclades. :P

:biggrin1:

Lol, I think all of us should get together and attempt to create a completely fresh, new version of Luigi Luca Cavalli-Sforza's "The History and Geography of Human Genes". B) ;)

parasar
03-15-2014, 09:45 PM
My official ANE results on David's run:

ANE: 34.317%
ME: 50.1756%
ENA: 13.4809%
SSA: 2.0265%

https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdG1KU0RzN2dVMHFsYVM3NHpsYXNVT Hc#gid=0

His comment on this run is pertinent to the point Dr_McNinja made above. I remember reading an analysis about a month back where MA1 was melding to Indus populations. The Amerindians share more drift but their EDAR period east Asiatic admixture, which MA1 did not have, keeps them more distant overall.


Update 15/03/2014: The above ADMIXTURE results suggest strong links between MA-1 and present-day South Central Asians. So to test this further, I ran a cluster analysis using the GAGA and Mclust algorithms on a global dataset including both MA-1 and La Brana-1, the Mesolithic genome from Iberia...
The results can be seen in the spreadsheet at the link below. MA-1 falls into cluster 21, along with La Brana-1 and the Chuvash, Mari and Tatars from the Volga-Ural region. Again, this result correlates well with the aforementioned TreeMix analysis from Raghavan et al., in which MA-1 is derived from a clade rooted closely to a Mari genome, but also not too far away from an Indian genome (see here).

GAGA + PCA + Mclust global cluster analysis

Please note, however, this doesn't mean that the Chuvash, Mari and Volga Tatars are genetically identical to MA-1 and La Brana-1. What it tells us is that all of the samples in this cluster show very similar affinities in terms of global genetic variation. Indeed, the result is misleading in one important way, because it fails to pick up on the fact that the Chuvash, Mari and Tatars are shifted towards East Asia and the Americas in large part because of recent East Eurasian admixture, while MA-1 and La Brana-1 are shifted in the same direction only because of shared (Mammoth steppe?) ancestry with Amerindians.

Dr_McNinja
03-16-2014, 06:02 AM
My results on David's run:


ANE 0.330996
ME 0.467526
ENA 0.172382
SSA 0.029096

Dr_McNinja
03-16-2014, 01:13 PM
David posted another analysis with results that are slightly off. Said he doesn't know why it happened that way. My ME seems to have jumped up to 0.493, a lot closer to the other Jat but a bunch of ANE leaked into ENA for everyone and more of mine did.

DMXX
03-16-2014, 01:27 PM
My results from Polako's ANE run:



ME 74.1
ANE 20.33
ENA 4.18
SSA 1.39


I like the approach he has taken (using populations which demonstrate ADMIXTURE component peaks as stand-ins against Mal'ta). It's a good working alternative until we get our hands on more aDNA.

Dr_McNinja
03-20-2014, 01:04 PM
Davidski posted some really interesting runs for MA-1 and LaBrana:

http://bga101.blogspot.com.au/2014/03/updated-eurogenes-k13-population.html

Also see the comments.

vettor
03-20-2014, 05:35 PM
Davidski posted some really interesting runs for MA-1 and LaBrana:

http://bga101.blogspot.com.au/2014/03/updated-eurogenes-k13-population.html

Also see the comments.

I still see an issue on why nobody splits the huge French marker ...............again see davidski split,.... french only joins with North-Italian which joins with Sardinian and then Tuscan........Is he saying northern french is associated with North-Italian and then sardinian? ...........its seems to me a split of french into a Northern-french ( germanic influence) and a southern-French ( gallic influence) would make these kinds of admixture runs far more clearer.

Or, Is north-Italian an isolated group which rarely moved over time and so needs to be associated with someone..............lets join it with a huge French marker, even including all northern France............makes little sense to me

ZephyrousMandaru
03-20-2014, 05:38 PM
Mammoth Steppe K=4 with La Brana

Mammoth_Steppe 10.19%
ME 85.97%
ENA (+ minor ANE) 3.22%
SSA 0.6%

Mammoth Steppe K=4 with MA-1

Mammoth_Steppe 10.38%
ME 85.75%
ENA (+ minor ANE) 3.26%
SSA 0.5%

Mammoth Steppe here is not purely ANE, it's more of a combination of ANE and WHG. But for us Middle Easterners, it's probably mostly, if not entirely ANE.

Dr_McNinja
03-23-2014, 06:39 AM
results for HRP0393 (Haryana Jatt)

ANE 34.1375%
ME 50.0629%
ENA 13.498%
SSA 2.3016%

Dr_McNinja
03-23-2014, 02:01 PM
results for HRP0402 (Punjabi Jatt of Pahari origin)

ANE 33.1381%
ME 51.9686%
ENA 11.6196%
SSA 3.2737%

parasar
03-27-2014, 02:13 AM
4 Ancestors Oracle results for Anzick-1, La Brana-1 and MA-1
...
it is possible, for instance, that a population very similar to present-day Oceanians was living somewhere on mainland Asia during the Upper Palaeolithic, and that's why La Brana-1 shows signals of ancestry from Melanesia and Malaya. Interestingly, he belonged to Y-chromosome haplogroup C, which today reaches extreme frequencies in Oceania and East Asia.

http://eurogenes.blogspot.com/

Dr_McNinja
03-27-2014, 02:35 AM
http://eurogenes.blogspot.com/Well that shows where the NE-Euro is coming from


MA-1

Using 2 populations approximation:
1 Brahmin_UP+Mari @ 26.092786
2 Kshatriya+Mari @ 26.496655
3 Bangladeshi+Mari @ 26.683673
4 Brahmin_UP+Chuvash @ 26.794254
5 Gujarati+Mari @ 27.074151
6 Chuvash+Kshatriya @ 27.085486
7 Bangladeshi+Chuvash @ 27.241915
8 Kanjar+Mari @ 27.244439
9 Chuvash+Kanjar @ 27.606985
10 Dharkar+Mari @ 27.658244
11 Mari+Uttar_Pradesh @ 27.695378
12 Chuvash+Gujarati @ 27.70266
13 Mari+Punjabi_Jat @ 27.967695
14 Chuvash+Uttar_Pradesh @ 27.984982
15 Mari+North_Kannadi @ 28.009527
16 Chuvash+Dharkar @ 28.063921
17 Kol+Mari @ 28.094145
18 Chuvash+North_Kannadi @ 28.183158
19 Chuvash+Kol @ 28.379488
20 Dusadh+Mari @ 28.588956Since ANE is the only route for Finno-Ugric affinity into South Asia, then admixture is just being extra sensitive when it gives wildly varying estimations of Finno-Ugric (NE-Euro) admixture for people who generally are within 1 to 2% of each other in ANE admixture (going by Davidski's runs) across much of the northern subcontinent from Afghanistan to Bihar.

I wonder how we could measure ANE affinity within West Asian components.

Shaikorth
03-27-2014, 06:38 AM
I wonder how we could measure ANE affinity within West Asian components.

Haven't really looked into that, but gut says the order is West Asian > med > Red Sea/Southwest Asian.

Also regarding ANE percentages and variance, remember that some ANE in admixture tests done so far gets hidden under ENA.

ZephyrousMandaru
03-27-2014, 07:11 AM
Well that shows where the NE-Euro is coming from

Since ANE is the only route for Finno-Ugric affinity into South Asia, then admixture is just being extra sensitive when it gives wildly varying estimations of Finno-Ugric (NE-Euro) admixture for people who generally are within 1 to 2% of each other in ANE admixture (going by Davidski's runs) across much of the northern subcontinent from Afghanistan to Bihar.

I wonder how we could measure ANE affinity within West Asian components.

Maybe you could develop an ADMIXTURE test based on six ancestral components? ANE, Mediterranean, Southwest Asian, European South Asian and East Asian

Arbogan
04-05-2014, 05:16 PM
I'd suggest adding ONGE as another K control(like the original study by lazaridis and reich), Han chinese are not good as sole representatives of ENAs. People with south-asian ancestry are definitely going to show up with inflated ANE, because of how ANE is closer to west-eurasians and modern day south-asians are a a mixture between ENA and west-eursian-like populations. Rather than solely Han-like ENA. While people are going to argue that the reason pathans and south-asians have high ANE, is because of their ancestors likely lived in areas where ANE-like immigrations frequently took place and expanded. I'd still put in ONGE, just to be sure. It's kinda unwise to not use one of the most salient examples of mostly unmixed ENA-like south-asian population. We also know that ANE is high related with the K "West-asian" which in turn is highly related to "ANI" more so than any other K. So to me there must be some underlying connection. Perhaps ANI population movements also carried with them ANE.

Volat
04-05-2014, 05:33 PM
Results
EEF 38.69
WHG 43.33
ANE 17.98

Sein
04-05-2014, 08:42 PM
I'd suggest adding ONGE as another K control(like the original study by lazaridis and reich), Han chinese are not good as sole representatives of ENAs. People with south-asian ancestry are definitely going to show up with inflated ANE, because of how ANE is closer to west-eurasians and modern day south-asians are a a mixture between ENA and west-eursian-like populations. Rather than solely Han-like ENA. While people are going to argue that the reason pathans and south-asians have high ANE, is because of their ancestors likely lived in areas where ANE-like immigrations frequently took place and expanded. I'd still put in ONGE, just to be sure. It's kinda unwise to not use one of the most salient examples of mostly unmixed ENA-like south-asian population. We also know that ANE is high related with the K "West-asian" which in turn is highly related to "ANI" more so than any other K. So to me there must be some underlying connection. Perhaps ANI population movements also carried with them ANE.

According to Reich et al. (the 2009 paper), Dai versus ASI is the least distant pair in a Papuan vs Dai vs ASI vs Onge tree (although ASI and Onge do seem to form a clade). In other words, ASI is closest to the Dai among all living populations, even though it constitutes a clade with the Onge. The Dai are an ethnic group from China. So East Asian populations can do the trick just fine. And with the Pashtuns, we also have f3 statistics. The HGDP Pashtuns are best fitted as a mix between Druze/Samaritan and Karitiana, using such methods. In other words, one can model the HGDP Pashtuns as a simple mixture between Middle Easterners, and ANE. So, these results seem solid and consistent. In addition, David tried other methods beside ADMIXTURE, and ANE always peaked among South Asians. In addition, the best method he used yielded 36.2% ANE for myself, which is pretty close to the published Native American limit. In fact, using the same method, he got 36% ANE for the Karitiana. In short, I turned out as ANE as the most ANE-rich Native American population, using a method that inferred percentages of ANE for other populations that matched the paper perfectly (I think this is the method David is now using in his Eurogenes Genetic Ancestry Detective work, perhaps he could chime in). He also tried an Iranian sample/participant. They turned out 18% ANE. They were also 6% ENA, and 1% Sub-Saharan African.

Besides what David has done, Everest tried computing ANE percentages for me, and he got 30%. Basically, everything seems to indicate that a population like the HGDP Pashtuns are between 30%-40% ANE. In terms of ENA, I think the average HGDP Pashtun is probably around 10%-15%. Of that, at least 4% will be recent Turkic admixture, and the rest will be ASI. Iranians also seem to be rich in ANE. I think depending on where they originate in Iran, as well as their ethnic background, diverse and heterogeneous Iranian populations will probably range between 10% to 25% ANE. In terms of ENA, most Iranians will probably range between 5%-15%, with some outliers having less than 5% ENA or more than 15% ENA (not sure about Turkmens though). With exception to Azerbaijani and northwestern Iranians, as well as northeastern Iranians, I think most of the ENA admixture for other Iranians (especially southern Iranians) will be predominantly ASI. Also, many Iranians will probably have around 1%-3% Sub-Saharan African admixture, with higher percentages in the south of the country. Although, I think many Iranians from the northwest and northeast will have no real SSA admixture whatsoever, or only less than 1%.

For what it's worth, I think our ANE ancestry provides a good explanation for a figure like this. This is from the supplements of the fineSTRUCTURE paper. "CentralSouthAsians" are just the HGDP Pakistanis. If one would add Iranians, I'm sure they would cluster with Pakistanis and Europeans, not with the Druze and other Middle Easterners. Figures like these always show the same thing, West Eurasians are split between Europeans+South Asians and Middle Easterners. But Iranians are tied with Europeans+South Asians, rather than other Middle Easterners:

1686

Mandoos
04-14-2014, 12:51 PM
Results
EEF 46.70135712
WHG 35.42020689
ANE 17.878436

parasar
04-14-2014, 02:50 PM
I'd suggest adding ONGE as another K control(like the original study by lazaridis and reich), Han chinese are not good as sole representatives of ENAs. People with south-asian ancestry are definitely going to show up with inflated ANE, because of how ANE is closer to west-eurasians and modern day south-asians are a a mixture between ENA and west-eursian-like populations. Rather than solely Han-like ENA. While people are going to argue that the reason pathans and south-asians have high ANE, is because of their ancestors likely lived in areas where ANE-like immigrations frequently took place and expanded. I'd still put in ONGE, just to be sure. It's kinda unwise to not use one of the most salient examples of mostly unmixed ENA-like south-asian population. We also know that ANE is high related with the K "West-asian" which in turn is highly related to "ANI" more so than any other K. So to me there must be some underlying connection. Perhaps ANI population movements also carried with them ANE.


According to Reich et al. (the 2009 paper), Dai versus ASI is the least distant pair in a Papuan vs Dai vs ASI vs Onge tree (although ASI and Onge do seem to form a clade). In other words, ASI is closest to the Dai among all living populations, even though it constitutes a clade with the Onge ...


I don't think ASI can be separated from ANE unless we go far back in time (i.e. in ancient DNA).
Ancestral ASI and ANI split about 2000 (~50000 years) generations back. Soon after (relatively), ASI and the Ongee split about 1700 (42500 years) generations back. The ANE component is not that old (again relatively), and as it developed on the (non-Ongee) ancestral ASI side there should be no ANE in the Ongee who had split a very long time back from ancestral ASI.
The ANE entry into the ANI side occurs about 600 (15000 years) generations back, or 27500 years after the Ongee's split from ancestral ASI.

Dr_McNinja
04-30-2014, 02:46 PM
According to Reich et al. (the 2009 paper), Dai versus ASI is the least distant pair in a Papuan vs Dai vs ASI vs Onge tree (although ASI and Onge do seem to form a clade). In other words, ASI is closest to the Dai among all living populations, even though it constitutes a clade with the Onge. The Dai are an ethnic group from China. So East Asian populations can do the trick just fine. And with the Pashtuns, we also have f3 statistics. The HGDP Pashtuns are best fitted as a mix between Druze/Samaritan and Karitiana, using such methods. In other words, one can model the HGDP Pashtuns as a simple mixture between Middle Easterners, and ANE. So, these results seem solid and consistent. In addition, David tried other methods beside ADMIXTURE, and ANE always peaked among South Asians. In addition, the best method he used yielded 36.2% ANE for myself, which is pretty close to the published Native American limit. In fact, using the same method, he got 36% ANE for the Karitiana. In short, I turned out as ANE as the most ANE-rich Native American population, using a method that inferred percentages of ANE for other populations that matched the paper perfectly (I think this is the method David is now using in his Eurogenes Genetic Ancestry Detective work, perhaps he could chime in). He also tried an Iranian sample/participant. They turned out 18% ANE. They were also 6% ENA, and 1% Sub-Saharan African.

Besides what David has done, Everest tried computing ANE percentages for me, and he got 30%. Basically, everything seems to indicate that a population like the HGDP Pashtuns are between 30%-40% ANE. In terms of ENA, I think the average HGDP Pashtun is probably around 10%-15%. Of that, at least 4% will be recent Turkic admixture, and the rest will be ASI. Iranians also seem to be rich in ANE. I think depending on where they originate in Iran, as well as their ethnic background, diverse and heterogeneous Iranian populations will probably range between 10% to 25% ANE. In terms of ENA, most Iranians will probably range between 5%-15%, with some outliers having less than 5% ENA or more than 15% ENA (not sure about Turkmens though). With exception to Azerbaijani and northwestern Iranians, as well as northeastern Iranians, I think most of the ENA admixture for other Iranians (especially southern Iranians) will be predominantly ASI. Also, many Iranians will probably have around 1%-3% Sub-Saharan African admixture, with higher percentages in the south of the country. Although, I think many Iranians from the northwest and northeast will have no real SSA admixture whatsoever, or only less than 1%.

For what it's worth, I think our ANE ancestry provides a good explanation for a figure like this. This is from the supplements of the fineSTRUCTURE paper. "CentralSouthAsians" are just the HGDP Pakistanis. If one would add Iranians, I'm sure they would cluster with Pakistanis and Europeans, not with the Druze and other Middle Easterners. Figures like these always show the same thing, West Eurasians are split between Europeans+South Asians and Middle Easterners. But Iranians are tied with Europeans+South Asians, rather than other Middle Easterners:

1686That graph is excellent. That makes me wonder, did the research on light skin/hair/eye pigmentation genes reveal how old these are and whether all light skinned Eurasians got them from a common ancestor, and if so, where that ancestor was from? Are those genes in South Asians ancestral (i.e, as old as or older than Europe's) or results of more recent backflow from the west?

Arbogan
04-30-2014, 03:17 PM
Besides what David has done, Everest tried computing ANE percentages for me, and he got 30%. Basically, everything seems to indicate that a population like the HGDP Pashtuns are between 30%-40% ANE. In terms of ENA, I think the average HGDP Pashtun is probably around 10%-15%. Of that, at least 4% will be recent Turkic admixture, and the rest will be ASI. Iranians also seem to be rich in ANE. I think depending on where they originate in Iran, as well as their ethnic background, diverse and heterogeneous Iranian populations will probably range between 10% to 25% ANE. In terms of ENA, most Iranians will probably range between 5%-15%, with some outliers having less than 5% ENA or more than 15% ENA (not sure about Turkmens though). With exception to Azerbaijani and northwestern Iranians, as well as northeastern Iranians, I think most of the ENA admixture for other Iranians (especially southern Iranians) will be predominantly ASI. Also, many Iranians will probably have around 1%-3% Sub-Saharan African admixture, with higher percentages in the south of the country. Although, I think many Iranians from the northwest and northeast will have no real SSA admixture whatsoever, or only less than 1%.

For what it's worth, I think our ANE ancestry provides a good explanation for a figure like this. This is from the supplements of the fineSTRUCTURE paper. "CentralSouthAsians" are just the HGDP Pakistanis. If one would add Iranians, I'm sure they would cluster with Pakistanis and Europeans, not with the Druze and other Middle Easterners. Figures like these always show the same thing, West Eurasians are split between Europeans+South Asians and Middle Easterners. But Iranians are tied with Europeans+South Asians, rather than other Middle Easterners:

1686

That's not quite right. West-iranics are a threeway mix between anatolian farmers, south-west asian herder groups and a very west-eurasian south-central asian group. With some influx of the indo-iranians. So while our ancestry is predominantly from non-semitic affiliated west-eurasian groups,(namely the caucasus-gedrosians), i'd not say we're in overall more affiliated with south-asians and europeans(this might be true of very eastern-iranic groups like pashtuuns and tajiks though). But rather we're somewhere inbetween.

Sein
04-30-2014, 06:12 PM
That's not quite right. West-iranics are a threeway mix between anatolian farmers, south-west asian herder groups and a very west-eurasian south-central asian group. With some influx of the indo-iranians. So while our ancestry is predominantly from non-semitic affiliated west-eurasian groups,(namely the caucasus-gedrosians), i'd not say we're in overall more affiliated with south-asians and europeans(this might be true of very eastern-iranic groups like pashtuuns and tajiks though). But rather we're somewhere inbetween.

In this regard, please refer to Figure S13 from the supplementary materials of Hellenthal et al., as it is very illustrative. It has Iranians, and the samples do cluster with South Asians and Europeans, to the exclusion of Southwest Asians. When analyzing West Eurasians, (by which I mean Europeans, South Asians, and Southwest Asians), Southwest Asians are an outgroup to Europeans and South Asians. But West Asians and Caucasians consistently cluster with Europeans and South Asians. In a cladistic sense, Europeans, West Asians, Caucasians, and South Asians are part of a single broad clade within the larger West Eurasian grouping.

Also, genetic distance is a separate matter from how populations relate to each other as lineages. For example, the HGDP Pashtuns have a significantly lower Fst distance to Bedouins than they do to certain South Indian populations. Yet, they do not constitute a clade with Bedouins, but group earlier with the genetically more distant South Indians. Another example would be WHG and ANE. WHG is closer to EEF than ANE, but it is still a clade with ANE.

Arbogan
04-30-2014, 10:07 PM
In this regard, please refer to Figure S13 from the supplementary materials of Hellenthal et al., as it is very illustrative. It has Iranians, and the samples do cluster with South Asians and Europeans, to the exclusion of Southwest Asians. When analyzing West Eurasians, (by which I mean Europeans, South Asians, and Southwest Asians), Southwest Asians are an outgroup to Europeans and South Asians. But West Asians and Caucasians consistently cluster with Europeans and South Asians. In a cladistic sense, Europeans, West Asians, Caucasians, and South Asians are part of a single broad clade within the larger West Eurasian grouping.

Also, genetic distance is a separate matter from how populations relate to each other as lineages. For example, the HGDP Pashtuns have a significantly lower Fst distance to Bedouins than they do to certain South Indian populations. Yet, they do not constitute a clade with Bedouins, but group earlier with the genetically more distant South Indians. Another example would be WHG and ANE. WHG is closer to EEF than ANE, but it is still a clade with ANE.

iranians form a closer clade with tamils? Please show me. That seems implausible. Had you said pathans, or sindhis that would be different.

Sein
04-30-2014, 10:14 PM
iranians form a closer clade with tamils? Please show me. That seems implausible. Had you said pathans, or sindhis that would be different.

I've provided the reference. Find it in here:
https://www.sciencemag.org/content/suppl/2014/02/12/343.6172.747.DC1/Hellenthal.SM.pdf

Look at the Bedouins, Jordanians, Syrians, and Palestinians, and look at the Iranians.

Arbogan
04-30-2014, 10:33 PM
I've provided the reference. Find it in here:
https://www.sciencemag.org/content/suppl/2014/02/12/343.6172.747.DC1/Hellenthal.SM.pdf

Look at the Bedouins, Jordanians, Syrians, and Palestinians, and look at the Iranians.

they weren't using behar iranians... were they.

Sein
04-30-2014, 10:40 PM
They used the Behar Iranians. 15 Behar Iranians, the majority, do actually form a closer clade to their Indian samples (which are from South India. So, in the big genetic picture, Iranians are actually closer to Tamils than they are to Palestinians or Syrians) than to Southwest Asians. They are clustering with Turkish samples and populations from the Caucasus.

3 Behar Iranians are different from the rest, and cluster with Syrians/Jordanians.

I always thought that results like these were highly unusual, until I started seeing work on aDNA. It all makes much more sense now, especially if we look at things in terms of "derived" and "basal" Eurasian.

Arbogan
05-02-2014, 12:38 PM
They used the Behar Iranians. 15 Behar Iranians, the majority, do actually form a closer clade to their Indian samples (which are from South India. So, in the big genetic picture, Iranians are actually closer to Tamils than they are to Palestinians or Syrians) than to Southwest Asians. They are clustering with Turkish samples and populations from the Caucasus.

3 Behar Iranians are different from the rest, and cluster with Syrians/Jordanians.

I always thought that results like these were highly unusual, until I started seeing work on aDNA. It all makes much more sense now, especially if we look at things in terms of "derived" and "basal" Eurasian.
it's odd indeed. I attribute it to iranians generally greater drift towards central asia (with their general richness of the "gedrosia" which is a designation for the genetic signatures of neolithic south-central asians). But I mean tamils? Shouldn't they technically be equally distant as bedouins? Because besides the south-central asian affinity of west-iranics, there is nothing driving them to be closer to tamils, and in large.... arabized syrians, aswell as caucausians and turks and all the others, do have a greater genetic commonality. Even if syrians are drifted towards saudi arabia, turks and caucasians towards europe, and iranics towards south-central asia. I don't see them as neccesarily seperate peoples. Besides the sub-clades also place druze in the same clade as turks, iranians and caucasians . So you might be right about basal eurasian. Where would pre-islamic south-west asians be placed, that would be a good question. Besides... are phylogenetics better for overall ancestral propositions, than admixture tests? Do we go for greater genetic commonality or genetic origin?

parasar
05-04-2014, 04:22 AM
I've provided the reference. Find it in here:
https://www.sciencemag.org/content/suppl/2014/02/12/343.6172.747.DC1/Hellenthal.SM.pdf



In the Figure S13: FineSTRUCTURE tree, the geographically distant (by land) American Indians and Papuans/Melanesians cluster together.

Sein
05-04-2014, 05:09 AM
In the Figure S13: FineSTRUCTURE tree, the geographically distant (by land) American Indians and Papuans/Melanesians cluster together.

This is actually quite interesting. On PCA plots, Papuans/Melanesians cluster towards West Eurasians, in comparison to other ENA populations. I think there is a chance that interesting dynamics are at play here, but we need ancient DNA. Despite being 40%-50% ANE, Native Americans are often the most extreme of all ENA populations on PCA plots. Something similar could be happening to the Papuans/Melanesians, especially since they cluster much closer to West Eurasians.

Shaikorth
05-04-2014, 08:03 AM
This is actually quite interesting. On PCA plots, Papuans/Melanesians cluster towards West Eurasians, in comparison to other ENA populations. I think there is a chance that interesting dynamics are at play here, but we need ancient DNA. Despite being 40%-50% ANE, Native Americans are often the most extreme of all ENA populations on PCA plots. Something similar could be happening to the Papuans/Melanesians, especially since they cluster much closer to West Eurasians.

Native Americans and Oceanians are usually not the furthest from West Eurasians specifically on global PCA's because that dimension gets hijacked by two easily distinguishable populations: the Sardinian isolate and Han Chinese which is probably the result of a recent expansion. Basques will take the place of Sardinians if they are missing. This is slightly misleading as a marker of ENA. We know from the Lazaridis paper that Han are worse ENA proxies than Onge, Taiwanese aborigines and, it seems safe to say, She, and may even have some ANE admixture because ENA keeps showing up in almost everyone in admixture-based attempts to separate ANE and ENA using Han reference. The alternative to ANE-like ancestry in Han is that almost everyone in Eurasia has East Asian ancestry. Eurogenes. IBS test also shows Han sharing more with MA-1 than Oceanians or She.

Native Americans are however the most distant from Africans in the global PCA's, which is basically correct and shows also by using IBS and other methods.

FINESTRUCTURE results change with linked and unlinked modes and sampling - both sample sizes and selected reference populations - is still important which makes Hellenthal et al. clustering problematic for some populations. The Native American/Oceanian link appears in many FINESTRUCTURE tests, but it would be useful to know whether they are linked or unlinked, later according to the creators shows deeper ancestry and more overall similarity while the former is for more recent genealogy. My own guess is that ENA's cluster together due to their shared low haplotype diversity, and East Asians and Siberians are in their own branch due to their ancestors having taken part in some shared population expansion. More recently mixed Uygurs and Hazaras join them because they don't have other places to go.


http://www.maths.bris.ac.uk/~madjl/finestructure/Clear_WorldCopymap1a.png

East Asian-specific population history may also affect shared drift and explain the discrepancy in MA-1's IBS sharing and shared drift results regarding East Eurasians. All Siberians and East Asians, but also South Asians (https://docs.google.com/file/d/0B9o3EYTdM8lQUFdqNmpfYjE3dFE/edit?pli=1) go up in relative similarity to MA-1 when IBS is used compared to shared drift, with Kets, who were the only Siberians to share more drift with MA-1 than all Europeans, now rivaling Karitiana, Hazaras and Yakuts getting more similarity than Lithuanians and Han getting more similarity than Spanish.