PDA

View Full Version : Central Asia input among some Bylany Iron Age people supposed to be a Celtic area



MOESAN
26-08-18, 12:41
Thanks to Eurogenes Davidski here under some opinions (to accept as Bible words or to discuss, as you want) - could be interesting nevertheess...

Saturday, August 25, 2018 Central Asian admixture in Hallstatt Celts (http://eurogenes.blogspot.com/2018/08/central-asian-admixture-in-hallstatt.html)

One of the two Hallstatt_Bylany samples from Damgaard et al. 2018 (https://www.nature.com/articles/s41586-018-0094-2), the individual labeled DA112 (263048 YBP), shows a subtle but clear signal of Central Asian ancestry. As far as I know, this hasn't yet been reported by anyone else, so I'll happily be the first to do it here. To initially explore this issue, here are few D-stats comparing DA112 with DA111, the other Hallstatt_Bylany sample, in regards to their affinity to ancient Central Asian groups:


https://2.bp.blogspot.com/-rE3Sg0RKEJg/W4IR8O0krsI/AAAAAAAAHGw/PiaeQEYfgbc5BXLw8RkxHctUl4Y49d_YwCLcBGAs/s400/Hallstatt_Bylany_DA112_vs_DA111_D-stats.png (https://2.bp.blogspot.com/-rE3Sg0RKEJg/W4IR8O0krsI/AAAAAAAAHGw/PiaeQEYfgbc5BXLw8RkxHctUl4Y49d_YwCLcBGAs/s1600/Hallstatt_Bylany_DA112_vs_DA111_D-stats.png)

Thus, the two significant mixture signals (Z≥3) are produced by Dashti_Kozy_BA, from Bronze Age Tajikistan, and Botai, from Eneolithic Kazakhstan. But a few of the other populations, like Scythian_Pazyryk, are also close to a significant Z-score. Of course, this doesn't mean that the Dashti_Kozy_BA and/or Botai peoples migrated into the Bylany region, in what is now the Czech Republic. It simply shows that DA112 is different from DA111 in a statistically significant way due to ancestry closely related to Dashti_Kozy_BA and Botai.

To further explore this issue I ran a series of mixture models using the G25/nMonte method. Below is an example of one of the models that made good sense and also returned a fairly reasonable statistical fit. Hence, it appears to me that DA112 was in some part, perhaps mostly, of Scytho-Sarmatian origin, with resulting minor admixture from Iron Age Central Asia.


Hallstatt_Bylany_DA112
Scythian_Hungary,79.2
Hallstatt_Bylany_DA111,18.2
Scythian_Pazyryk,2.6
[1] distance%=4.0302
Since it's extremely unlikely that DA112 was the only Hallstatt Celt with this type of genetic structure, then it's reasonable to conclude that at least some Hallstatt populations harbored Scytho-Sarmatian ancestry. Admittedly, this isn't a very surprising conclusion considering the close contacts, as inferred from archaeological data, between the Hallstatt culture and various nomadic groups with assumed origins far to the east of Central Europe. In fact, here's what the Damgaard et al. supplement says about the Bylany burial site:


The anthropologist J. Chochol hypothesized that cremations were of individuals of the local population, whereas the skeletal remains represented immigrant nobility.

Examples of long-distance contact are present in the form of a horse harness (probably Kimery horizon), a mounted stone characteristic of the Carpathian Basin and Black Sea region in Grave 1, and a pin in Grave 18 with bird motifs analogous to those found in the Caucasus.
Nevertheless, it's nice to see archaeogenetics corroborate archeology on yet another issue. For those of you who would like to try running your own G25/nMonte mixture models of DA112, all of the necessary data are available at the links below. If you're new to this, a guide to modeling with the G25/nMonte is available here (http://eurogenes.blogspot.com/2018/02/modeling-genetic-ancestry-with-davidski.html).


Global 25 datasheet (scaled) (https://drive.google.com/file/d/1FSzKKknFGcOgfyA76q9PD7B-n-MJs7L8/view?usp=sharing)

Global 25 datasheet (https://drive.google.com/file/d/1Dj8ysMY4oePcbW5PszuiCnDBJi9ANjtt/view?usp=sharing)

Global 25 pop averages (scaled) (https://drive.google.com/file/d/1bPt1Xqy0HwXpJFcug3vaC3MBqFTMvj5A/view?usp=sharing)

Global 25 pop averages (https://drive.google.com/file/d/1Obeulb6oqo677Fm6dtYAvYYUCRjFqEgz/view?usp=sharing)

See also...

First real foray into Migration Period Europe: the Gepid, Roman, Ostrogoth and others... (http://eurogenes.blogspot.com/2018/03/first-real-foray-into-migration-period.html)


Posted by Davidski (https://www.blogger.com/profile/04637918905430604850) at 11:58:00 PM

MOESAN
29-09-18, 17:28
I was trying to put a thumb-up to my post but I failed??? Does somebody know how to make that? LOL
Seemingly it doesn't motive people too much -
about Bylany and Celts, Davidski wrote:

https://2.bp.blogspot.com/-optx810q0Ww/W6wo_3TtzTI/AAAAAAAAHP0/SBpyoC044r8yEr7xR_E7XidiMxUhIRABACLcBGAs/s450/Prehistoric_Britain_%2526_Ireland_PCA.png (https://2.bp.blogspot.com/-optx810q0Ww/W6wo_3TtzTI/AAAAAAAAHP0/SBpyoC044r8yEr7xR_E7XidiMxUhIRABACLcBGAs/s1600/Prehistoric_Britain_%2526_Ireland_PCA.png)

Perhaps it's not a coincidence that the likely Celtic-speaking Iron Age individuals from present-day England (labeled England_IA) are positioned between these older British and Irish samples and the two ancients from Iron Age burials in present-day Bylany, Czechia, associated with the Hallstatt culture (marked with black stars). That's because the Hallstatt people are generally considered to have been the earliest speakers of Celtic languages.

Hence, what the PCA might be showing is a genetic shift in the British and Irish Isles caused by the arrival of Hallstatt Celts in Northwestern Europe.

Interestingly, the present-day English samples appear to be a mixture of Britain_&_Ireland_BA, England_IA and England_Anglo-Saxon. However, a subset of these samples is also heavily shifted "east" towards one of the Hallstatt individuals and present-day Dutch, suggesting that they harbor extra admixture from continental Europe.

This isn't easy to make out on my plot, because of the clutter, but I can assure you that it's true. Keep in mind that you can plug the datasheet into the PAST program (freely available here (https://folk.uio.no/ohammer/past/index.html)) to have a much closer look at the PCA and even change the color coding.

To check whether England_IA can be modeled as a mixture of Britain_&_Ireland_BA and Hallstatt with formal methods, I ran an analysis with the qpAdm software using all of the publicly available Bronze Age samples from present-day Britain and Ireland. The standard errors are high, likely because Britain_&_Ireland_BA and Hallstatt are closely related, but, overall, we can probably say that the model does limp across the line.


England_IA
Britain_&_Ireland_BA 0.5550.172
Hallstatt 0.4450.172
chisq 18.513
tail prob 0.100973
Full output (https://drive.google.com/file/d/1FC25MBj_K2ewcI0Aiksntyj_DzQ5i5xP/view?usp=sharing)
However, the really important thing about this output is that England_IA cannot be modeled as simply Britain_&_Ireland_BA (the chisq and tail prob are way off). Thus, even though the Hallstatt samples from Bylany don't appear to be ideal proxies for the admixture in England_IA that is lacking in Britain_&_Ireland_BA, the signal they produce does suggest that a closely related population arrived in the British Isles during or after the Bronze Age to give rise to England_IA.



https://2.bp.blogspot.com/-optx810q0Ww/W6wo_3TtzTI/AAAAAAAAHP0/SBpyoC044r8yEr7xR_E7XidiMxUhIRABACLcBGAs/s1600/Prehistoric_Britain_%2526_Ireland_PCA.png
It's true Davidski has a strong tendancy to "manipulate" some data to have a clearer picture, according to him.
This doesn't give any clue about the small "east-asian" input in Bilany, it's true.

hrvclv
29-09-18, 20:17
A previous thread ("Ancient genomes from the Caucasus inc. Maykop" - Angela, May 2018) stated that while both Yamna and Maykop were basically EHG + Caucasus, Yamna had received some extra EEF whereas Maykop had extra Palaeolithic Siberian.

Maybe the DA112 had deep origins in Maykop, while other clans around had more Yamna-like ancestry. Or maybe, even within the same tribe, people had inherited different shares of Yamna and Maykop.

This said, the time lapse between Yamna-Maykop and Halstatt is more than sufficient for plenty of things to happen, even some yet-unknown incursions from the east. Time is precisely the weak point of my hypothesis : I guess that by Halstatt times, Celtic genomes within the same area would have been fairly homogeneous. Anyway, I just thought it might be interesting to recall that Chuan-Chao Wang et al paper.

Angela
29-09-18, 21:17
One of the reasons I like the new Lazaridis et al paper is because they removed the human part of the analysis as much as possible. When they couldn't they went to excruciating lengths to ensure accuracy. People should read the methodology section. It's a wonder.

PCAs can be easily manipulated by the populations you choose to include, the number of samples in your populations, etc. Even more sophisticated statistical tools can be manipulated by choosing certain specific populations.

The gullibility of a lot of people in this hobby amazes me. (present company excepted). It's taken years for the head of "Nutrition Science" at Cornell to fall. That's a disgrace. Population genetics can be so easily misused that these "findings" have to be checked as soon as possible.

If you want to be believed, and that applies to anyone, publish every detail of your methodology and all your samples, as someone like Lazaridis or Reich does. Then we'll see. It's all about honest methodology and replication. If the information to replicate the experiment is not produced then I don't pay much attention.

As to the "meat" of the issue, I don't see anything generally implausible about it.

MOESAN
29-09-18, 23:47
A previous thread ("Ancient genomes from the Caucasus inc. Maykop" - Angela, May 2018) stated that while both Yamna and Maykop were basically EHG + Caucasus, Yamna had received some extra EEF whereas Maykop had extra Palaeolithic Siberian.

Maybe the DA112 had deep origins in Maykop, while other clans around had more Yamna-like ancestry. Or maybe, even within the same tribe, people had inherited different shares of Yamna and Maykop.

This said, the time lapse between Yamna-Maykop and Halstatt is more than sufficient for plenty of things to happen, even some yet-unknown incursions from the east. Time is precisely the weak point of my hypothesis : I guess that by Halstatt times, Celtic genomes within the same area would have been fairly homogeneous. Anyway, I just thought it might be interesting to recall that Chuan-Chao Wang et al paper.

I don't know where people found Maykop have extra 'paleosiberian' and Yamna extra 'EEF', I did not see this in Wang, or it has to be very light- maybe you confused genuine Maykop (with kind of EEF or ANF) with Steppe-Maykop which has almost no EEF and has some light 'paleosiberian' (just enough, I think Yamna Kalmykia had surely this last element in it) - all the way the admixtures of Wang seem a bit strange if we consider that his green colour marks pure 'CHG' -
But Maykop, in my mind, has very little weight in the Celt prehistory - concerning this thread and Bylany, based it's true upon only two DNAs, and other Celts, I think they had contacts with some Scythians (celtic art is presumed to be influenced by scythic art); dates of Bylany (around 600 BC) make it possible too - here it could be an outsider,or rather the descendant of an ancient outsider, from Steppes, and perhaps Scythian -
I understand that coming from Eurogenes, based on a so little sample, this thread is not so interesting - maybe some new discoveries from other sources about the same topic will awake a new interest for the question - let's this thread have a rest for a while

halfalp
30-09-18, 11:24
One of the reasons I like the new Lazaridis et al paper is because they removed the human part of the analysis as much as possible. When they couldn't they went to excruciating lengths to ensure accuracy. People should read the methodology section. It's a wonder.

PCAs can be easily manipulated by the populations you choose to include, the number of samples in your populations, etc. Even more sophisticated statistical tools can be manipulated by choosing certain specific populations.

The gullibility of a lot of people in this hobby amazes me. (present company excepted). It's taken years for the head of "Nutrition Science" at Cornell to fall. That's a disgrace. Population genetics can be so easily misused that these "findings" have to be checked as soon as possible.

If you want to be believed, and that applies to anyone, publish every detail of your methodology and all your samples, as someone like Lazaridis or Reich does. Then we'll see. It's all about honest methodology and replication. If the information to replicate the experiment is not produced then I don't pay much attention.

As to the "meat" of the issue, I don't see anything generally implausible about it.

You might be right, but if this was the attention, i have to say i think Davidski is very honest in the way he deals with the samples. I looked at his methodology as well as some of Anthrogenica users like Kurd, and they use very different and genuine pop source for their calculator, at least they try to brainstorm things, but it doesn't mean their results are more accurate than the professional ones. What i have noticed with Davidski, a lot of people tend to hate him and respect him at the same level. They might be blame him to be too narrow to other hypothesis, but not about the methodology he use for his results. Davidski is probably a very passionate guy and he will continue his work, whatever the new hypothesis or results are telling us.