I'll start the ball rolling with a 9-way mixture analysis of 93 European, Near Eastern and Central Asian present-day and ancient populations. The relevant datasheet and R script are available
here and
here.
Below is a simple tree/cluster analysis based on the results, using the freely available Past3 software. Makes perfect sense, I'd say.
It's important to understand that these sorts of tests are basically designed to estimate ancient ancestry proportions, rather than calculate minor admixtures. With that in mind, here are a few observations:
- Karasuk outlier RISE497 (the most eastern Karasuk individual) is surprisingly important for Near Eastern populations
- Nordic LNBA and Sintashta look very similar in terms of overall ancestry proportions, suggesting that they perhaps derive from the same ancestral population
- The effects of postmortem deaminantion or DNA damage appear to be expressed in many of the non-UDG treated ancient samples as minor Sub-Saharan admixture
Can anyone put together a better model for West Eurasians? Also, I'd really like to see a well thought out D-stats/nMonte analysis of South Central Asia.
See also...
Yamnaya = Khvalynsk + extra CHG + maybe something else
D-stats/nMonte open thread #2
228 comments:
«Oldest ‹Older 201 – 228 of 228To simplify things, and increase precision, can someone produce an nMonte datasheet with these outgroups. This would be greatly appreciated:
Mbuti
Yoruba
K14
Samara_HG
Motala_HG
Ami
Han
Eskimo
Selkup
Karitiana
LBK_EN
BedouinB2
Caucasus_HG2
Kharia2
Onge2
Iberia_Mesolithic
Iberia_Chalcolithic
And then all populations included for testing against these outgroups. Please include Karelia_HG, MA1, BedouinB, Onge, and Kharia, so that they can be used as putative mixing populations in the modelling itself.
Thanks in advance.
Using simple math and D-stats, we could have figured out (almost)exactly what Yamnaya was before any Yamnaya genomes were sampled. All we needed Middle Neolithic European genomes. The non-MN proxies all would have impossible results, except for a few, most of whome have similar to EHG and Yamnaya.
This'll be my last post about creating ghost ancestors with D-stats. Check out these results for Srubnaya.
When I model them as Yamnaya+???, All ghost ancestors have them with either impossible results(Extreme distance from EHG and extreme closeness to EEF) or can be fit as a mixture of EEF+Yamnaya, except one ghost ancestor.
With this ghost ancestor, Srubnaya fits as 80% Yamnaya+20% ghost ancestor. It's results are basically identical to heavily WHG admixed MNs. So, if all we had were Yamnaya genomes, using this simple method we could have figured out Srunaya was about 80% Yamnaya and 20% something Sardinia-like.
A few words to the cluster Analysis, especially with regards to the Italian samples: It's somewhat surprising that North Italians and Tuscans cluster with the Greek1 sample, at least that's not what PCAs have hinted at. However, in one of the biggest craniometric cluster analyses of antique crania, with as many measurements as possible, (by I. Schwidetzky) I have seen something very similar:
Classical Etruscans were closest to hellenistic Greeks. Also clustering with them were classical Greeks and the Punic Ibiza (Spanish_Baleares!). The Po plain in the imperial Age was close to them all, but closest to the contemporaneous Greeks and Histria, a Greek colony in Romania. Bologna in the first century was close to the rest of the Po plain, and particularly close to Phanagoria, a Greek colony in southern Russia. Also to Pompeji btw.
So on the whole the ancient craniometric data suggested a Close similarity between Etruscans, Greeks, North Italians and the Baleares, just like in the cluster Analysis based on the D-stats/4mix results.
Now, since the craniometric cluster analysis seems to make sense genetically, it's interesting to look at the other Italian cranial samples used: The area of Rome in the imperial Age resembled the contemporaneous Catalans. -> So possibly the ancient Latins were still quite Iberian-like! Ancient DNA will show if this prediction turns out right, but I would bet it.
The Veneti of the later Iron Age cluster with the southern French Ligurians of the earlier Iron Age. -> The were probably southern French-like.
The Marche region of the early imperial Age clusters with a sample from Sicily 300-400 AD and Anatolia 1000-1 BC. So I guess they would be close to East Sicilians, Maltese, Greek2 in the above Cluster Analysis. It makes sense, because the middle Adriatic region of Italy is among the Italian regions with the highest incidence of y-haplogroup J2, it's higher there than in Latium.
What's interesting here: The IE peoples like the linguistically closely related Latins and Veneti seem to be on the branch of the cluster Analysis with lower CHG, together with Iberians etc. The non-IE Etruscans on the other Hand seem to have had already an elevated CHG like modern North Italians, which puts them close to Greeks. From the Marche Region there is at least the North Picene language with unclear affiliation. But at any rate archaeology showed that the east Mediterranean contacts there were much more intense than in Latium. This is directly reflected in the craniometric clustering with Sicily and Anatolia and in the high incidence of J2.
So maybe the late dates inferred for admixture with Near Eastern like populations are rather the result of late mixture within Italy, under Roman rule...
Onur, it is interesting that Afanasievo (9.1%) has even more Caucasus HG than Yamnaya Kalmykia (5.75%) in spite of Afanasievo being in Altai and much farther away from Caucasus. Compare:
Afanasievo: Anatolia Neolithic 0%, Caucasus HG 9.1%, Loschbour WHG 0%
Altai Iron Age: Anatolia Neolithic 5.9%, Caucasus HG 3.55%, Loschbour WHG 0.25%
Azeri Baku: Anatolia Neolithic 41.85%, Caucasus HG 27.15%, Loschbour WHG 0%
Nogai: Anatolia Neolithic 28.45%, Caucasus HG 18.1%, Loschbour WHG 1.35%
Kumyk: Anatolia Neolithic 36.95%, Caucasus HG 26.15%, Loschbour WHG 1%
Turkmen: Anatolia Neolithic 26.25%, Caucasus HG 12.85%, Loschbour WHG 1.15%
Kyrgyz: Anatolia Neolithic 10.7%, Caucasus HG 5.35%, Loschbour WHG 2.5%
Uzbek: Anatolia Neolithic 19.75%, Caucasus HG 11.6%, Loschbour WHG 1.2%
Corded Ware Germany: Anatolia Neolithic 21.9%, Caucasus HG 0%, Loschbour WHG 7.25%
Srubnaya: Anatolia Neolithic 22.2%, Caucasus HG 0%, Loschbour WHG 9.3%
Sintashta: Anatolia Neolithic 25.55%, Caucasus HG 12.5%, Loschbour WHG 11.45%
Andronovo: Anatolia Neolithic 18.35%, Caucasus HG 0%, Loschbour WHG 12.8%
Ukrainian East: Anatolia Neolithic 34.7%, Caucasus HG 0.15%, Loschbour WHG 10.95%
Russian Kargopol: Anatolia Neolithic 27.5%, Caucasus HG 0%, Loschbour WHG 12.05%
Belarusian: Anatolia Neolithic 34.25%, Caucasus HG 0%, Loschbour WHG 12.05%
Lithuanian: Anatolia Neolithic 30.2%, Caucasus HG 0%, Loschbour WHG 17.75%
Czech: Anatolia Neolithic 40.15%, Caucasus HG 0.75%, Loschbour WHG 11%
And by comparison:
Georgian: 42.65% Anatolia Neolithic, Caucasus HG 39.6%, Loschbour WHG 0%, Karasuk outlier 6.4%, Yamnaya Samara 10%
Sintashta is exceptional as it has quite much Caucasus HG and WHG. Otherwise, ancient IEs seem to be quite uniform from Germany to Altai. In general, modern Turkics and IEs in the core areas differ in that Turkics have Caucasus HG and are very low in WHG while IEs in the core areas virtually lack Caucasus HG and are high in WHG.
It would be exciting if Afanasievo yDNA turned to be J2. It would also be very interesting, if we had an autosomal analysis of Xiaohe mummies.
I added Georgian in the list because Georgians also have steppe ancestry (Yamnaya Samara + Karasuk) which is not related to WHG rich IEs, and I would really like to get more ideas about the prehistory of Kartvelian languages.
Afanasievo is derived from a Yamnaya population genetically intermediate between Yamnaya Kalmykia and Yamnaya Samara.
So Afanasievo males probably belonged to R1b, although they may have been R1a if, like Corded Ware, they actually originated from an as yet unsampled Yamnaya group or closely related steppe population that carried R1a.
If they come out R1a, this will also suggest that they're the ancestors of the R1a-rich Tarim Basin mummies.
Archeology shows that Afanasievo was a typical western steppe pastoralist population like Yamnaya, and now we know that they were closely related to Late Neolithic/Bronze Age Central and Eastern Europeans. So they were in all likelihood Indo-European speakers.
@Krefter
One thing about CHG is that it behaves a bit differently in D-Stats and in Admixture. Kotias has an underwhelming behaviour in D-stats, only showing high'ish affinity to Georgians, Yamnaya to a lesser degree some North Europeans that got direct Yamnaya input. On the other hand, in admixture the CHG component is very strong from India to SE Europe.
So one possibility is that Kotias itself is not very relevant except for Georgia (and surroundings) and Yamnaya. But when we get another Kotias-like (bus not exactly the same) genome, it will show a stronger affinity in Dstats to what Admixture shows as CHG-rich populations. With that theoretical genome, many populations (especially from Asia, but probably also SE europe) will probably get less Anatolia Neolithic and more CHG in Dstats+nMonte.
Re: ghost populations, it would be much more difficult to figure out EHG and CHG instead of Yamnaya as a single population. So things can be more complicated. In any case, we have an idea about the possibilities of how unsampled areas could look like by using modern populations, but that doesn't solve the problem of what was there at one time and what came at another time. In Europe you're figuring out Yamnaya by having WHG and EEF already. In Asia we don't have Mesolithic and Early neolithic samples to do the same. There is no magic, we can only speculate.
Unrelated, and just an "out of curiosity" test. Trying to fit MA1 as WHG + something S/SE Asian shows this:
MA1
"Dravidian_India" 59.2
"Loschbour_WHG" 32.5
"Dai" 5.25
"Ami" 3.05
"Onge" 0
distance=0.04353
The MA1 ancestry in India must be very old. It would be interesting to get UP samples from anywhere between India and Tajikistan to see what they looked like.
@Davidski
1) So what happened to the new evidence I already posted - on your blog-post from Match 1st?
2) Well known professors of archeology and history have claimed that an early-mesolithic arrival of the uralic language is the ONLY sensible way to explain WHEN the uralic language spread north to the Gulf of Finland.
So what is it they did not know, that you know, to make their conclusions "senseless"?
3) Considering highly professional linguists like professor Alinei and Wiik as "fringe elements" are somewhat arrogant though.
4) Regarding your patience with "off-topic rants" I'll say it's impressive.
Thus it is somewhat surprising that you choose to exclude specific resumes of new facts, highly relevant to solve the genetc riddles of mesolithic Eurasia from the discussion. Being honest is as important as bring login, og scirnce is what you AIM to maintain.
@Frank
The academic discussion about the origin and evolution of the Uralian languages are more than a century already. Over the last three decades I have got to follow the discourse from various angles. Thus I am well aquianted with the works of mr. Montanus and his predesessors within the field. Whether considered fringe or not.
As you may note - the question is still far from settled:
http://www.sgr.fi/sust/sust242.html
"Afanasievo is derived from a Yamnaya population genetically intermediate between Yamnaya Kalmykia and Yamnaya Samara."
Yamnaya Kalmykia: Anatolia Neolithic 0.4, Atayal 0, Caucasus HG 5.75, Karasuk outlier 2, Loschbour WHG 0, Yamnaya Samara 90
Yamnaya Samara: Yamnaya Samara 100
Afanasievo: Anatolia Neolithic 0, Atayal 0, Caucasus HG 9.1, Karasuk outlier 3.8, Loschbour WHG 0, Yamnaya Samara 87.1
Afanasievo Caucasus HG percentage 9.1. is not intermediate between Yamnaya Kalmykia Caucasus HG percentage which is 5.75 and Yamnaya Samara Caucasus HG percentage which is 0.
As for dates for Afanasievo, according to Wikipedia "Conventional archaeological understanding tended to date at around 2000–2500 BC. However radiocarbon gave dates as early as 3705 BC on wooden tools and 2874 BC on human remains. The earliest of these dates have now been rejected, giving a date of around 3300 BC for the start of the culture."
Rob emphasized that the actual commencement of Yamnaya is 3200 BC, so it looks like Afanasievo started even earlier in spite of being farther in Asia.
Davidski, I may agree with you when we get Afanasievo yDNA and an autosomal analysis of Xiaohe and we have a subclade analysis of Xiaohe R1a1.
@batman
There are two threads on Uralics at the links below. Feel free to post your evidence there.
http://eurogenes.blogspot.com.au/2015/09/uralic-genes.html
http://eurogenes.blogspot.com.au/2015/10/reconstructing-genetic-history-of.html
If these threads are too dead for you, then please wait until the topic of Uralic origins comes up again. This should happen soon, when more ancient DNA from Northeastern Europe is published.
@Kristiina
Keep in mind that these nMonte results might be skewed a little by deamination.
As far as I can tell based on all my analyses, Afanasievo is basically identical to Yamnaya Kalmykia, and it's hard to tell which is more southern, but it's probably Yamnaya Kalmykia.
To sum up, do you mean this:
Afanasievo 3300 BC, Caucasus HG 10% + EHG, Proto-IE
Southern Yamnaya 3200 BC, Caucasus HG 5.75% + EHG, Proto-Hittite
Fatyanovo 3200 BC, Caucasus HG 0?, IE (xHittite, Tocharian)
Finnish Corded Ware 3200 BC, Caucasus HG 0?, IE (xHittite, Tocharian)
Western Corded Ware 2900 BC, Caucasus HG 0%, IE (xHittite, Tocharian, Armenian)
Sintashta 2100 BC, Caucasus HG 12%, Indo-Iranian
Srubna 1800 BC, Caucasus HG 0%, Indo-Iranian
Andronovo Fedorovo 1500 BC, Caucasus HG 0%, Iranian
Armenian MBA 1500 BC, Caucasus HG 34%, Armenian
or this
Afanasievo 3300 BC, Caucasus HG 10% + EHG, Proto-IE
Southern Yamnaya 3200 BC, Caucasus HG 5.75% + EHG, Proto-IE
Fatyanovo 3200 BC, Caucasus HG 0?, Proto-IE
Finnish Corded Ware 3200 BC, Caucasus HG 0?, Proto-IE
Western Corded Ware 2900 BC, Caucasus HG 0%, IE (xHittite, Tocharian, Armenian)
Sintashta 2100 BC, Caucasus HG 12%, Indo-Iranian
Srubna 1800 BC, Caucasus HG 0%, Indo-Iranian
Andronovo Fedorovo 1500 BC, Caucasus HG 0%, Iranian
Armenian MBA 1500 BC, Caucasus HG 34%, Armenian
More evidence of Steppe migration to South Asia. I used the method I described above(finding ghost ancestors) on Tajikistan and Gujarat.
I assumed they have ancestry from people like Kharia. I then tried to find who their non-Kharia ancestors are. Here's the makeup of their non-Kharia ancestors, when I model it as Steppe/EEF/WHG/CHG/Northern West Asia(Caucasus, Iran, Iraq).
GujaratiA's West Eurasian side: 53.5% Yamnaya, 24.75% Anatolia Neolithic, 18.7% CHG, 3.7% WHG @ D=0.008472.
Tajik_Ishkashim's West Eurasian side: 68.45% Yamnaya, 26.1% Anatoilia_Neolithic, 5.45% CHG @ D: 0.013808.
For GujaratiaA the ghost ancestor's results only make sense when they're modelled as 40-50% Kharia and for Tajik_Ishkashim when they're modeled as 20-40% Kharia. In all the other models, the ghost ancestor is unrealistically distant from Kharia or has clear Kharia-affinity.
@ Chad, thanks for the MA-1 stats, they've allowed me to include MA-1 in some nMonte runs (still have a couple of regression predicted stats in there, for Hungary_HG to Onge2 and Karelia_HG to Onge2, but they look basically right, so I don't think they cause any problems).
Using MA1 as well as the set I used upthread plus Itelmen (Anatolia_Neolithic, BedouinB, Caucasus_HG, Dai, Esan_Nigeria, Hungary_HG, Karelia_HG, Loschbour_WHG, Masai_Kinyawa, Nganasan, Onge, Ulchi, Ust_Ishim), I found a choppy contribution of MA1 to South Asia vs Siberians e.g. :
Pathan:
Caucasus_HG 28.3, BedouinB 16.45, Karelia_HG 14.35, Anatolia_Neolithic 13.1, Itelmen 10.9, Onge 5.25, Dai 4.4, Ust_Ishim 2.9, Nganasan 1.35, Esan_Nigeria 1.2, Loschbour_WHG 0.95, MA1 0.8, Yakut 0.05, Hungary_HG 0, Ulchi 0
Dravidian India:
Ulchi 22.95, Caucasus_HG 19.65, Onge 14.15, BedouinB 12.75, Yakut 6.3, Ust_Ishim 6.05, Karelia_HG 6, Anatolia_Neolithic 4.05, Esan_Nigeria 3.7, Dai 3, Hungary_HG 1.4, Itelmen 0, Loschbour_WHG 0, Nganasan 0, MA1 0
While:
GujaratiC:
Caucasus_HG 23.9, MA1 20.95, Dai 19.7, BedouinB 12.9, Onge 8.3, Anatolia_Neolithic 7.4, Loschbour_WHG 2.85, Esan_Nigeria 2.1, Ust_Ishim 1.9, Hungary_HG 0, Itelmen 0, Karelia_HG 0, Nganasan 0, Ulchi 0, Yakut 0
Kalash:
Caucasus_HG 32.4, BedouinB 14.15, MA1 13.35, Anatolia_Neolithic 11.9, Dai 10.8, Karelia_HG 10, Onge 3.65, Itelmen 2.5, Esan_Nigeria 0.8, Ust_Ishim 0.45, Hungary_HG 0, Loschbour_WHG 0, Nganasan 0, Ulchi 0, Yakut 0
The Kalash and GujaC look pretty much like I'd have expected them to (except a lack of Karelia_HG in GujaC), while Pathan and Dravidian India not really... Whatever it is makes these populations choose Karelia vs MA1 vs Siberians must be quite finely balanced.
The differences from Alberto's fit must relate to his use of the post-Yamnaya steppe and Dravidian India, and not using Karelia and this somehow gives rise to the higher MA1 share.
I did also find that Bronze Age Armenia picked up some Itelmen as well, for some reason.
Bronze Age Armenia: Caucasus_HG 37.55, Anatolia_Neolithic 34.9, BedouinB 12.35, Karelia_HG 8.25, Itelmen 5.75.
and Georgian, but less so:
Georgian: Caucasus_HG 39.7, Anatolia_Neolithic 37.6, BedouinB 12.2, Karelia_HG 5.5, Itelmen 4.15, Onge 0.85
(But then this is probably part of the phenomenon of low level apparent ENA in West Eurasia?).
The MA1 stats fit where they look like they should in PCA space - http://i.imgur.com/GR48D7m.png (or http://i.imgur.com/qX6yCAr.png for just West Eurasian ancients+Ust Ishim) although closer than they would be if there was somehow an MA1 column and row together as well.
@Alberto and Matt,
You got to take out East Asians and Africans, when modelling South Asians. You should only do that if there are relevant East Asian and African outgroups. David's D-stat's outgroups are good at displaying European diversity and it also pretty good at displaying West Asian diversity, but certainly not for Africa and East Asia.
Some fits for Siberians without only Dai and Onge as ENA, but with MA1:
Yakut - Dai 77.1, MA1 18.5, Karelia_HG 4.05, Anatolia_Neolithic 0.35 (distance% = 1.2803%)
Ulchi - Dai 91.1, MA1 8.9 (distance% = 1.6771%)
Nganasan - Dai 80.75, MA1 15.3, Karelia_HG 3.95 (distance% = 1.8578)
Itelmen - Dai 77.7, Karelia_HG 14.3, MA1 - 8 (distance% = 3.3039)
Seemingly random preferences of Karelia_HG and MA-1. Could be an outgroup selection issue, or it could support the idea that modern populations have ANE from a group equally related to MA-1 and Karelia_HG. Then again, RK pointed out that the double outgroup (Primate X SSA Y) method of D-stat similarity has issues and that Nick Patterson agrees, which is why they use a single outgroup. Dunno how large of an effect that would cause.
Sein,
Try this one. Hopefully it doesn't have any errors.
https://drive.google.com/file/d/0B9o3EYTdM8lQNTdxcm1xbUhXVXc/view?usp=sharing
Alberto
On that previous sheet the Atayal percentages make no sense, this is why I'm saying that is noise cause by the high MA1 related ancestry in India. And also yes some Europeans pick up dravidian but the model is generaly worst (isn't that imply that there is a shared ancestry but not a direct one?). I mean, Estonian is 3,75 while lithuanians are 0, Maltese 0 while sicilians have 4,35, Aragon has 0 while Andalucia has 2,9.... How do you explain that?!
Probably not enough East Asian outgroups in that particular sheet to estimate East Asian ancestry correctly.
But I can tell you that the elevated East Asia in Estonians versus Lithuanians does make sense.
To some degree they capture the same thing.
Only Atayal in rough order
West Sicilian, Albanian, Estonian, Hungarian, Bulgarian, Ukrainian_East, French, some Iberians, Bergamo, Sardinian, Czech, English.
Only Karasuk in rough order:
Kumyk, Chechen, Finnish, Iranian Jew, Georgian Jew, Lezgin, Abkhasian, Iraqi Jew, Syrian, Lebanese, Armenians, Bedouins, Georgians, Maltese, West Ukrainian, Croatian, some Spanish, Lithuanian, Norwegian
Only Ulchi
Yemenite Jew, Scottish, Basque, Icelandic
Two or three of the above
Uzbek, Kyrgyz, Turkmen, Nogai, Turkish, Iranian, Kargopol Russian, Mordovian, North Ossetian, Adygei, Saudi, Sephardi, Ashkenazi, Druze, Palestinian, Jordanian, Greek, EastSicilian.
Now, West Ukrainians having 5% Karasuk and East Ukrainians having 3% Atayal doesn't make sense if we assume they come from different sources, so same thing is probably represented there, whatever it is. What's interesting is why some populations get more than one of the components.
This latest sheet has a lot of East Eurasian outgroups. Might be better for East Asian ancestry.
https://drive.google.com/file/d/0B9o3EYTdM8lQNTdxcm1xbUhXVXc/view?usp=sharing
@Davidski
In my comment to your post of Match 1st I did actually make a summary of numerous new data, referring the new, archaeological timeline that links the late paleolithic Hamburg/Bromme/Pertunian-culture to the early mesolithic Ahrensburg/Lyngby/Swidrien-cultures.
Thus we have a well documented match, in both time and space, rhrough ALL the critical climate-phases of the late Pleistocene Fact by fact.
Combined with the results from KO14 and MA1 this is now giving us a chronology that pinpoint a genological continuity between the paleolithic European and the present Eurasians.
According to Eske Willerslev K14 is "most close to the present Danes, Swedes, Finns and Russians". Consequently we can't exclude that a cultural and linguistic continuation within Europe at large and northern Europe specifically.
Since 1992 this possible continuation have been rised by various researchers and described by solid academic works - like that of professor Alinei and his collegues.
Since the results of K14, MA14 and other paleolitic/mesolithic samples it's not possible to debunk these claims as "fringe" or "non-sensical". Whatever our previous comprehensions, opinions or preferences may have been.
@Davidski
In my comment to your post of Match 1st I did actually make a summary of numerous new data, referring the new, archaeological timeline that links the late paleolithic Hamburg/Bromme/Pertunian-culture to the early mesolithic Ahrensburg/Lyngby/Swidrien-cultures.
Thus we have a well documented match, in both time and space, rhrough ALL the critical climate-phases of the late Pleistocene Fact by fact.
Combined with the results from KO14 and MA1 this is now giving us a chronology that pinpoint a genological continuity between the paleolithic European and the present Eurasians.
According to Eske Willerslev K14 is "most close to the present Danes, Swedes, Finns and Russians". Consequently we can't exclude that a cultural and linguistic continuation within Europe at large and northern Europe specifically.
Since 1992 this possible continuation have been rised by various researchers and described by solid academic works - like that of professor Alinei and his collegues.
Since the results of K14, MA14 and other paleolitic/mesolithic samples it's not possible to debunk these claims as "fringe" or "non-sensical". Whatever our previous comprehensions, opinions or preferences may have been.
New D-stats sheet with Motala_HG samples included in the rows and columns.
https://drive.google.com/file/d/0B9o3EYTdM8lQUzh3T0EwMHdSeUU/view?usp=sharing
And another one...
https://drive.google.com/file/d/0B9o3EYTdM8lQRmRlZzdvOExDaDQ/view?usp=sharing
@Davidsky,
I'm really relieved to see that some North Caucasians can in fact score around 20% CHG. On both my mother's and my father's side, I'm a Kebertei Circassian with very old family records, and I scored only 24.12 Caucasus_HG according to the recently updated puntDNAL K10. Some people told me that I was supposed to score much higher CHG, but I scored around 50% WHG instead. I somehow get it given the extreme fairness of all my extended family, but I'm disappointed since I was expecting to see more relevant oracles. I think CHG is a main component for South Caucasians and some North Caucasians, rather than Circassians who originate from the Northwesternmost part of the Caucasus region. My top oracles often indicate I'm supposedly closer to West and Northwest Europeans than East Europeans while some calculators tell I have good amount of Baltic, too. I am confused. My 23andMe results show that over 60% of my ancestry is defined as "Nonspecific European", "Nonspecific North European", and "Nonspecific South European". This is very disappointing. What surprises me is that Adyghe in Russia score less than 30% Yamnaya, according to your spreadsheet. In some calculators, it's even less. Though our ancestal homeland is in present-day Russia, more than 90% percent of total Circassian population in the world live outside Russia today. As a matter of fact, there are very few sub-branches left in Russia. That's why I believe that not all Adyghe in Russia truly represent the genetic profile of early Adyghe people because most of them have "recent" ancestry from other North Caucasians, while in diaspora they don't due to isolation. Despite this, the data in current genetic calculators almost exclusively stem from Russia. I don't know if it's same in yours. So Davidsky, I would to know what you think about this issue overall.
It depends on the model used, but keep in mind that it's difficult to estimate CHG ancestry with ADMIXTURE-based calculators.
In my D-stats/nMonte analysis Adygei score 29.5% extra CHG above what they have in the Yamnaya, Karasuk and maybe other components. So their overall CHG score should be about 40%.
https://docs.google.com/spreadsheets/d/1zSErfrAWLuyT_PW03BlOCuHYVDGhXjpHrGfkDuuSoJM/edit?usp=sharing
This seems accurate when compared to other analyses, like PCA. So if you're Adygei or from a similar ethnic group that's what you should be scoring, otherwise you'd plot somewhere far away from North Caucasians, and I doubt that would be reflective of reality.
Davidski
I am a very late with my answer, but maybe you still read it.
"But I do think it's often possible to work out whether specific language groups were spoken in particular places at certain times, and genetics can be used to help make such judgements."
I agree with this statement of course. But your arguments often imply that those specific language groups where also the only ones spoken in a region / archaeological horizon. Based on historical examples of material cultures with diverse language setup it is very possible (I'd even dare to say more likely) that entities like Corded Ware, Yamnaya or Andronovo were populated by linguistically diverse populations. This is not in odds with the idea that they had important role in the spread of Indo-European.
Post a Comment