teal CHG component in Yamna and Afanasievo arrived during Khvalynsk period

Angela said:
It doesn't matter, Bicicleur, the percentages are wrong. It's a flawed analysis, as Fire-Haired also pointed out. Admixture alone can't be used, you have to look at formal statistics as well. Plus, even in terms of admixture the percentages are off from those in the academic papers.
Click to expand...

K = 14 means 14 eigenvectors are calculated, which are specific to this dataset
it makes perfect sense to compare genomes within this dataset

however, if you run K = 14 on another dataset another 14 eigenvectors will be calculated
therefore it doesn't make sense to compare genomes from 2 different datasets on which K =14 has been calculated
that is the mistake many make
that is why they think K admixture don't work
 
bicicleur said:
and here you can see what the teal actualy means in this table (at the top of this image) :

View attachment 8222

attachment.php
Click to expand...
Why are you showing Armenians and the Semites (Jews + Assyrians)??? Modern Armenians & Semitic Assyrians are not really a good proxy for 'teal'. Both groups are NOT native to the Iranian Plateau. Original Armenians are native to Southern Caucasus and Semites are native to the Levant. Assyrian/Akkadians ancestors came from the Levant. Aramaic and Ugaritic ancestral Semitic tribes of the Assyrians came from the Levant. Their Akkadian ancestors came from the Arabistan.


" The region of origin of the reconstructed Proto-Semitic language, ancestral to historical and modern Semitic languages in the Middle East, is still uncertain and much debated. A 2009 Bayesian analysis identified an origin for Semitic languages in the Levant around 3750 BC with a later single introduction of Ge'ez from what is now South Arabia into the Horn of Africa around 800 BC, with a slightly earlier introduction into parts of North Africa and southern Spain with the founding of Phoenician colonies such as ancient Carthage in the ninth century BC and Cádiz in the tenth century BC.[1][2][3] The earliest records of Semitic languages are from 30th century BCE Mesopotamia.

Other theories include origins in the Arabian Peninsula or North Africa. "

https://en.wikipedia.org/wiki/Ancient_Semitic-speaking_peoples
 
Angela said:
It doesn't matter, Bicicleur, the percentages are wrong. It's a flawed analysis, as Fire-Haired also pointed out. Admixture alone can't be used, you have to look at formal statistics as well. Plus, even in terms of admixture the percentages are off from those in the academic papers.
Click to expand...
Possibly you are right, and I too take these amateurish admixture runs rather lightly. However, I find it more logical to see Europe after IE bronze age invasion more ANE (or is it EHG - dark blue) than typically more WHG - black. There is always a question where this surge of WHG comes from in modern Europeans?
We don't need to do much of mental equilibristics to explain extra EHG/ANE after Bronze Age, to agree with huge invasion of IE from Yamnaya. Having said that, I'm still digesting this K14 runs.
 
Goga said:
Why are you showing Armenians and the Semites (Jews + Assyrians)??? Modern Armenians & Semitic Assyrians are not really a good proxy for 'teal'. Both groups are NOT native to the Iranian Plateau. Original Armenians are native to Southern Caucasus and Semites are native to the Levant. Assyrian/Akkadians ancestors came from the Levant. Aramaic and Ugaritic ancestral Semitic tribes of the Assyrians came from the Levant. Their Akkadian ancestors came from the Arabistan.


" The region of origin of the reconstructed Proto-Semitic language, ancestral to historical and modern Semitic languages in the Middle East, is still uncertain and much debated. A 2009 Bayesian analysis identified an origin for Semitic languages in the Levant around 3750 BC with a later single introduction of Ge'ez from what is now South Arabia into the Horn of Africa around 800 BC, with a slightly earlier introduction into parts of North Africa and southern Spain with the founding of Phoenician colonies such as ancient Carthage in the ninth century BC and Cádiz in the tenth century BC.[1][2][3] The earliest records of Semitic languages are from 30th century BCE Mesopotamia.

Other theories include origins in the Arabian Peninsula or North Africa. "

https://en.wikipedia.org/wiki/Ancient_Semitic-speaking_peoples
Click to expand...
Goga, look at this huge WHG in Iranian Zoroastrians. Where did it come from?
https://genetiker.wordpress.com/2016/10/07/k-14-admixture-analysis-of-lapita-genomes/
 
LeBrok said:
Goga, look at this huge WHG in Iranian Zoroastrians. Where did it come from?
https://genetiker.wordpress.com/2016/10/07/k-14-admixture-analysis-of-lapita-genomes/
Click to expand...
Yeah, that's absurd much. I don't know where it is from? What do you think? This doesn't make any sense, maybe they made an error. But if is true, I don't think it is from the Steppes, because in the Steppes there is not much of that dark blue auDNA, almost the same amount as those Zoroastrian samples. Don't forget that at one point ALL people in Iran were 'Zoroastrians' and there is a continuations of the Iranian from Bronze age to nowadays. If there was a migration from the Steppes, there would be also much much more EHG (other blue) in those Zoroastrian samples and much less of that dark blue..


Look at the Levant N (neolithic), some of those samples have also very much of that dark blue component.
 
That has been actually known for long time now but somehow some bloggers delibaretly ingored this. Even in the Haak papers you could see CHG like admixture in some of the Samara Eneolithic samples. And as Maciamo pointed out above and as I too had said this. The Leyla Tepe is probably a transition region. As far as I remember there was a study some times back that said archeologically there is connection between Leyla Tepe- Maykop and some older cultures in North/Northwest Iran. So the source is ultimately somewhere there. It is no coincidence that the oldest Kurgans are found in the Leyla Tepe culture closely followed by the Samara Eneolithic.

However I disagree that the original R1a and R1b people were EHG. I believe the original R1 people were something part ancestral to both EHG and Iran_Neo/CHG. Something ANE like (not necessary ANE itself).
 
bicicleur said:
K = 14 means 14 eigenvectors are calculated, which are specific to this dataset
it makes perfect sense to compare genomes within this dataset

however, if you run K = 14 on another dataset another 14 eigenvectors will be calculated
therefore it doesn't make sense to compare genomes from 2 different datasets on which K =14 has been calculated
that is the mistake many make
that is why they think K admixture don't work
Click to expand...

More accurately, the figure should refer to the number of subpopulations (if the usual nomenclature applies to this particular set). ADMIXTURE uses a relatively simple optimization algorithm applied directly to the SNPs you chose to feed it with.

The question is if global analyses like the one in the original post are informative at all. For example, if, say, Khvalynsk is closest to the ancestral blue component, does it really make sense to model significantly older Samara hunter gatherer as partly derived from a Khvalynsk-like population? The problem would be even more pronounced with the Karelian individuals, of course. This approach tends to produce hypothetical populations where they really aren't needed to arrive at a relatively complete understanding of the population dynamics in a given time and space.
 
In GEDmatch Multiple Kit Analysis, Karelian EHG share a lot of segments with Samara EHG, and both also share with Khvalynsk.

And Khvalynsk shares a lot with Indo-Iranian cultures. These cultures (Sintashta, Srubna, Potapovka) share with Ancient Iranians:

http://www.eupedia.com/forum/thread...Yamna-ancestry?p=494942&viewfull=1#post494942

Check this comparison of segment sharing between 17 ancient samples (two from ancient Iran/Kurdistan, 15 from the Steppe):

GEDmatch comparison of ancient Iranians and ancient Steppe

Vdxgqn6.png
 
Nothing new. There is some Steppe ancestry in West Asia (Iran, Kurdistan, etc) due to some backmigration of the EASTERN Iranic speaking Scythians from the Steppes. Ancient Iranians Iranized the Steppes. Many thousands of years later some Iranized Scythinas migrated back into the Iranian Plateau. It has been documented.

But it doesn't mean that ancient Original Iranians are from the Steppes. Only 'Iranized' East Iranic speaking Scythians were from the Steppes. But those Scythians were MULTI inter-racial/ethnic people...
 
MarkoZ said:
More accurately, the figure should refer to the number of subpopulations (if the usual nomenclature applies to this particular set). ADMIXTURE uses a relatively simple optimization algorithm applied directly to the SNPs you chose to feed it with.

The question is if global analyses like the one in the original post are informative at all. For example, if, say, Khvalynsk is closest to the ancestral blue component, does it really make sense to model significantly older Samara hunter gatherer as partly derived from a Khvalynsk-like population? The problem would be even more pronounced with the Karelian individuals, of course. This approach tends to produce hypothetical populations where they really aren't needed to arrive at a relatively complete understanding of the population dynamics in a given time and space.
Click to expand...

indeed K = 14 supposes the whole dataset can be explained by 14 ancestral components

I checked th whole list and it seems to work for this dataset

But on the other hand, you could ask the same question about e.g. EEF, is the Stutgart genome realy the best representative? I don't think so. The stutgart genome is not ancestral, maybe the recently found Boncuklu genomes are a better fit. It is almost certain there is already some WHG admixture in the Stutgart genome.
 
Goga said:
Many thousands of years later some Iranized Scythinas migrated back into the Iranian Plateau.
Click to expand...

Actually Iron Age and Medieval Kurds/Iranians do not share any segments >3 cM with IA Scythian. They share with Sintashta, Srubnaya, Potapovka. BTW, only high quality samples are useful for these comparisons (I used only samples with >400,000 SNPs).
 
Tomenable said:
Actually Iron Age and Medieval Kurds/Iranians do not share any segments >3 cM with IA Scythian. They share with Sintashta, Srubnaya, Potapovka. BTW, only high quality samples are useful for these comparisons (I used only samples with >400,000 SNPs).
Click to expand...
double post
 
T637158 & M381564 are Iron Age Iranian (sample F38) and Medieval Iranian (sample I1955).

They are from Western Iran (Iranian part of Kurdistan).
 
bicicleur said:
indeed K = 14 supposes the whole dataset can be explained by 14 ancestral components

I checked th whole list and it seems to work for this dataset

But on the other hand, you could ask the same question about e.g. EEF, is the Stutgart genome realy the best representative? I don't think so. The stutgart genome is not ancestral, maybe the recently found Boncuklu genomes are a better fit. It is almost certain there is already some WHG admixture in the Stutgart genome.
Click to expand...

Arguments about perceived 'purity' are misplaced. Stuttgart is useful because he's in the right place at the right time to tell us something about the peopling of central Europe. If you want to further break up his ancestry that's quite another matter. You could do the same thing with the Karelian genomes.
 
MarkoZ said:
Arguments about perceived 'purity' are misplaced. Stuttgart is useful because he's in the right place at the right time to tell us something about the peopling of central Europe. If you want to further break up his ancestry that's quite another matter. You could do the same thing with the Karelian genomes.
Click to expand...

Stuttgart should be considered EEF with WHG admixture.
The same bias as in a well-defined K admixture.
 
bicicleur said:
Stuttgart should be considered EEF with WHG admixture.
The same bias as in a well-defined K admixture.
Click to expand...

I guess we're getting a little off-topic. The fact remains that Stuttgart was an actual individual and in all likelihood a good representative of the larger Central European Neolithic.
 
Tomenable said:
Kotias shares more segments with Steppe than IranN does; so I guess Kotias is a more likely source of CHG in Yamna:

GEDmatch comparison of Kotias CHG, Iran Neo and Steppe

wZf8yxO.png
Click to expand...
I don't know, it there is something fishy here. How come the best fit for Iran and CHG is Potapovka (almost in Siberia). Also Khvaliynsk, the R1b from Samara) is only at 34 and 16? If he was a fresh immigrant from South Caucasus, as we supposed in other thread, the numbers should have been in high hundreds, but they are not. See the Khvalinsk to Potapovka is at 240.
And Scythian IA is so little related to anyone as he have come from the Moon (almost). Kotias and Iran samples are over 10ky old while others are from 6 to 3 ky old. Also we have to keep in mind that there was bigger relationship steppe HG to CHG due to sharing some WHG/EHG genes.
 
You must log in or register to reply here.

This thread has been viewed 24875 times.

Back
Top