search this blog

Sunday, April 22, 2018

Likely Yamnaya incursion(s) into Northwestern Iran


Despite being stratigraphically dated to 5900-5500 BCE (ie. the Chalcolithic period), ancient sample Hajji_Firuz_ChL I2327 from Narasimhan et al. 2018, belongs to Y-haplogroup R1b-Z2103 and shows minor, but unambiguous, Yamnaya-related ancestry on the autosomes. Why is this a problem? Because both R1b-Z2103 and the Yamnaya culture are dated to the Bronze Age, and Yamnaya samples from Kalmykia and Samara regions of what is now western Russia are exceptionally rich in R1b-Z2103.

Thus, pending a successful radiocarbon (C14) dating analysis, it seems unlikely that Hajji_Firuz_ChL I2327 was alive during the Chalcolithic. Rather, it appears that he's partly of Yamnaya or closely related origin and has been wrongly dated. His remains are likely to be from a secondary burial from the Bronze Age that collapsed into the layer below, right into a Chalcolithic bin ossuary burial full of much older bones.

This scenario is strongly corroborated by data from two other ancient individuals from what is now Northwestern Iran:

- Hajji_Firuz_BA I4243 (also from Narasimhan et al. 2018 and from the same site as Hajji_Firuz_ChL I2327) was initially also stratigraphically dated to the Chalcolithic, but is now labeled as a Bronze Age sample after a radiocarbon (C14) analysis of the remains revealed a date of 2465-2286 calBCE. Moreover, this individual packs around 50% Yamnaya-related ancestry.

- Iran_IA F38 (from Broushaki et al. 2016) from an Iron Age burial at Tepe Hasanlu, which is just a few miles from Hajji Firuz, also belongs to Y-haplogroup R1b-Z2103 and harbors some sort of steppe ancestry on the autosomes (see here).

Below is a Principal Component Analysis (PCA) showing how this trio compare in terms of genome-wide ancestry to C14-dated Chalcolithic samples from Hajji Firuz and the nearby Seh Gabi. The relevant datasheet is available here.


Clearly, they're shifted "north" relative to the C14-dated Chalcolithic samples and thus closer to the ancient Eastern Europeans, suggesting that they carry ancestry from north of the Caucasus that was missing, or at least much less pronounced, in the region before the Bronze Age.

I used D-stats of the form D(Outgroup,Pop1)(Pop2,X), in which Pop1 are Eastern European Hunter-Gatherers (EHG) and Pop2 the C14-dated Chalcolithic samples, to test this more directly. And indeed, the D-stats showed that Hajji_Firuz_ChL I2327 and Hajji_Firuz_BA I4243 shared significantly more alleles with EHG than the Chalcolithic samples did (Z≥3).

Mbuti EHG Hajji_Firuz_ChL Hajji_Firuz_ChL_I2327 D 0.0151 Z 3.737
Mbuti EHG Hajji_Firuz_ChL Hajji_Firuz_BA_I4243 D 0.0496 Z 12.072

Mbuti EHG Seh_Gabi_ChL Hajji_Firuz_ChL_I2327 D 0.0188 Z 4.803
Mbuti EHG Seh_Gabi_ChL Hajji_Firuz_BA_I4243 D 0.0531 Z 12.832

To further elucidate this EHG-related signal, I ran a series of mixture models with the qpAdm software. Below are among the statistically most successful models that I could find for Hajji_Firuz_ChL I2327, Hajji_Firuz_BA I4243 and Hajji_Firuz_ChL. Note that each of these models features Yamnaya from Samara as the proxy for the northern, EHG-related ancestry.

Hajji_Firuz_ChL_I2327
Barcin_N 0.087±0.045
Seh_Gabi_ChL 0.768±0.045
Yamnaya_Samara 0.145±0.033
chisq 8.802
tail prob 0.551007
Full output

Hajji_Firuz_ChL
Barcin_N 0.232±0.027
Seh_Gabi_ChL 0.736±0.029
Yamnaya_Samara 0.033±0.018
chisq 4.269
tail prob 0.93439
Full output

Hajji_Firuz_BA_I4243
Barcin_N 0.201±0.038
Seh_Gabi_ChL 0.244±0.042
Yamnaya_Samara 0.554±0.031
chisq 9.209
tail prob 0.512363
Full output

Yamnaya_Samara-related admixture appears in all of the models. But considering the standard errors, the Yamnaya_Samara-related ancestry proportion for Hajji_Firuz_ChL is very close to zero. Moreover, importantly, Hajji_Firuz_ChL can be modeled successfully without any such ancestry, while Hajji_Firuz_ChL I2327 and Hajji_Firuz_BA I4243 can't (refer to the full output files for the details).

I also tested whether ancient samples from what is now Armenia might make better proxies than Yamnaya for the northern ancestry harbored by Hajji_Firuz_ChL I2327. That's a negative (see here and here).

I don't have a clue who these people were who brought Yamnaya or, at least, Yamnaya-like ancestry to the South Caspian region during the Bronze Age and perhaps also the Chalcolithic. It's rather unlikely that they were the early Iranians, who probably arrived in the region from Central Asia during the Late Bronze Age or even Iron Age (for instance, see here). Perhaps they were the Hittites? Indeed, in his book In Search of the Indo-Europeans, archaeologist James Mallory suggested that the ancestors of the Hittites and other Anatolian-speakers entered the Near East via the Caucasus route:

Most arguments for an Indo-European invasion from the northeast concern the appearance of a new burial rite at the end of the fourth and through the third millennium BC. At that time, both north of the Black Sea and the Caucasus, burials on the Russian-Ukrainian steppe were typically placed in an underground shaft and covered with a mound (kurgan in Russian). Before 3000 BC there begin to appear in the territory of the indigenous Transcaucasian (Kuro-Araxes) culture somewhat similar burials such as the royal tomb of Uch-Tepe on the Milska steppe. As tumulus burials are previously unknown in this region, some would explain their appearance by an intrusion of steppe pastoralists who migrated through the Caucasus and subjugated the local Early Bronze Age culture. More importantly, a status burial inserted into a mound at the site of Korucu Tepe in eastern Anatolia has been compared with somewhat similar burials both in the Caucasus and the Russian steppe. The discovery of horse bones on several sites of east Anatolia such as Norsun Tepe and Tepecik are seen to confirm a steppe intrusion since, as mentioned earlier, the horse, long known in the Ukraine and south Russia, is not attested in Anatolia prior to the Bronze Age.

Another option, however, is that they belonged to some other extinct Indo-European group, such as the Gutians (see here). In any case, keep an eye out for more Bronze Age samples from this part of the world. I have a strong feeling that, unlike their Neolithic and Chalcolithic predecessors, they will be rich in steppe ancestry and R1b-Z2103.

See also...

The Hajji Firuz fiasco

Yamnaya: home-grown

Big deal of 2018: Yamnaya not related to Maykop

Ahead of the pack

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...

265 comments:

1 – 200 of 265   Newer›   Newest»
Aram said...

So Yamnayans were light skin and light haired? Because the period in which those BA sample appears is exactly the same period when mentions of namrum people starts in Mesopotamian texts.

Davidski said...

@Aram

Are these two samples light haired? I don't know? But I think correlating my analysis with Mesopotamian texts is a bit out there.

Aram said...

Well it would be amazing if they turn out light skin. But off course it is not very important detail.
What is more important is the corroboration of this data with archaeologic facts. Which I think will support such a incursion.

Anthro Survey said...

Just a thought: I wonder if HF_BA and 2327 could both be ancient North Caucasians and/or Maykop-descended somehow or if what I describe below is merely a coincidence of similar ancestral streams mixing.

Curiously enough, if z2103 Vucedol is assumed to be 50% EEF and 50% steppe-rich newcomer(such a scenario is reasonable given the time-frame), the decoupled steppe-rich, CHG-rich portion plots quite close to HF_BA(~2, scaled). It also plots rather close to the Yamnaya_Ukraine_outlier(~3, scaled). Vucedol horizon is one of the first places in Europe where the Maykop-style shaft-hole axe tradition makes its appearance, so there's also that.

decoupled portion:
Maykop?,0.112067647,0.117556673,0.007205686,0.037324064,-0.020469628,0.018291074,0.006608827,-0.004439565,-0.044461979,-0.037143166,0.001173351,0.006773758,-0.016160194,-0.010930222,0.01445873,-0.000300973,0.007145049,-0.004254535,0.000241226,0.008380783,0.001780054,0.002408011,0.007742323,0.006104807,-0.00379677

Davidski said...

I'm pretty sure I2327 and I4243 are from the same or similar steppe-admixed populations dating to around the same time. The Z2103 isn't a coincidence.

Anthro Survey said...

@Aram

As counter-intuitive as it is, lighter skin and blondism in the Near East--be it Armenia, Syria, NW Iran, or N Caucasus---look to be mostly attributable to some northern strain of ANF-rich groups. So far, unlike with CWC, Pit Grave and Poltavka have yielded results suggestive of black/dark_brown haired and brown eyed populations.

As for Pashtuns and Pamiris? In the Eastern Iranosphere, there is considerable Sintashta/Andronovo admixture. As things stand, Sintashta/Andronovo are CWC derivatives and, hence, share GAC-related ancestry with Bronze Age European groups. Judging by the results, GAC and related EEF cultures in the vicinity seem as the likely culprit for European pigmentation diversity.

vAsiSTha said...

I2327. What are the archaeological evidences of dating of this person? What artefacts have been found along with him?
Afaik, there is no mention of him being a genetic outlier, and no mention of it being an intrusibe burial.

Davidski said...

@Kulkarni

Read what I wrote and try and understand it.

vAsiSTha said...

@davidski i see you giving random knkwledge on archaelogy about the skeleton falling into layer below which you have no clue about.

What i see you doing is manufacturing evidence to suit your theory.

Davidski said...

@Kulkarni

Quit acting crazy or I'll ban you.

For one, I didn't manufacture his Y-chromosome halogroup, which is common on the Bronze Age steppe, but missing from all C14-dated Near Eastern samples from before the Iron Age.

And I didn't manufacture the other Hajji Firuz sample with obvious Yamnaya ancestry that was thought to be Chalcolithic, but turned out to be Bronze Age after a C14 dating. This skeleton fell from the Bronze Age into the Chalcolithic layer. This is a fact.

I told you to read what I wrote and to try and understand it. So why don't you actually do that? How hard can it be to understand what I wrote and to consider it objectively?


vAsiSTha said...

I will consider it objectively once the radiocarbon dates are out.

Davidski said...

I see, so you're going to be biased until then? Why not just cut to the chase and be objective now?

Open Genomes said...

@David

I created a .geno file for DA101. The SNP names are exactly as in the Reich Lab 1240k list. No conversion necessary for you. :D

DA101 has 1171935 SNPs out of the 1198444 non-Y SNPs in 1240k set.

Does it work?

http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA101/DA101.geno

Davidski said...

@Open Genomes

Yes, it works, but it looks like the same files as the last one. So I can't use it.

Rob said...

I agree this individuals produces fits requiring Yamnaya or Afansievo ( I get 12% rather than 17, but whichever).


The description of the samples given:
"The excavation levels are placed in time ordered Phases from “A” to “L” (latest to earliest phases of site occupation). The Hajji Firuz Tepe skeletal materials used in the analysis were excavated in 1968 by Mary Voigt and Robert Dyson (University of Pennsylvania Museum). Burials at the site are all associated 1681 with architectural features and each burial unit contains either multiple individuals in a pit, bin or jar ossuary, or are single floor burials. Individuals within the ossuary context are commingled but, in several cases, the skeletons were in partial positional articulation. Bones from Units K10 (I2327), F10 (I2323, I2328, I4241) and F11 (I4243, I14349, I4351)"

The R1b man is I2327, phase F-G - toward the end of the phases, from unit K10 - a bin ossuary, thus possibly having multiple burials (although it does not describe if this is the case).

Such a derived lineage of Z2013 would be very surprising in the Neolithic.
If from Bronze Age, NW Iran was part of the Kura-Araxes culture.
The problem is C14 requires decent amounts of collagen, whilst aDNA does not.

Ric Hern said...

Didn't Hittites themselves mention that their Ancestors Migrated from the East ? I have read something in that lines years ago so sorry for not providing a link.

The oldest domesticated horse remains South of the Caucasus seems to be in Azerbaijan (Alikemek-Tepesi) and other remains are younger as you move Westwards through Northwestern Iran, Armenia and into Anatolia.

Tracing people are harder when concentrating only on Cultural objects especially if they were traders. So we probably will not find a full Yamnaya Package on this route but they certainly were there.

I wonder if ancestors of Maykop originated in Azerbaijan due to a mixture of Early Yamnaya or earlier related group from the Steppe into Azerbaijan, established trade between the Southern Caucasus and the Steppe (Horses for Women) then migrated back to the Maykop area due to pressure from the South while others migrated Westwards to Anatolia ?

Synome said...

If the evidence for this incursion turns out to be validated by C14 (and I think it's solid) I would lean towards the Gutian hypothesis. Placing a Yamnaya derived population in the Zagros during the Bronze Age is perfect positioning to fit their historic documentation.

The Hittite theory seems slightly less fitting because of the distance from Anatolia and lack of historical mention of any Hittite related peoples in the Zagros area.

AWood said...

Even if the R1b-Z2103 dated slightly earlier than the Bronze age remains, it wouldn't really mean a whole lot. Based on all the other data, it's clear that R1b has nothing to do with the native Mesopotamian cultures. R1b is intrusive from the north.

Onur Dincer said...

Didn't Hittites themselves mention that their Ancestors Migrated from the East ? I have read something in that lines years ago so sorry for not providing a link.

The oldest recorded place of origin of the Hittite kings is Kussara. Kussara was situated somewhere southeast of Hattusa, their later capital, but was still within the borders of the Anatolian peninsula. The oldest recorded Kussaran kings lived during the 17th century BC (short chronology). We do not know where exactly the Hittite kings or Hittite speakers lived before the 17th century BC.

Michalis Moriopoulos said...

Gutian makes sense to me.

@Aram

The meaning of "namrum" is disputed, and if that's the only evidence to go on that the Gutians were "light," it would be best to let the DNA do the talking, and not the whims of presumptuous Assyriologists from the early 20th century. Even if it did mean light and was a phenotypic reference, it probably referred to skin color, not hair. The Yamnaya (Pit Grave and Poltavka) samples we have are overwhelmingly dark-haired and brown-eyed. Light-haired steppe people (a la Andronovo) seem to only become important later, after significant hybridization with Old Europeans (a la GAC, many of which have been predicted to be blond, surprisingly). If the steppe ancestors of Hittites and Gutians came directly from the tap, we should expect them to look more like Yamnaya than their Proto-Indo-Iranian relatives who came later, no?

All will be revealed sooner or later.

old europe said...

Mike the Jedi

Light-haired steppe people (a la Andronovo) seem to only become important later, after significant hybridization with Old Europeans

Agree…the hybridization not only involved skin and hair color, it involved religion ( cremation rite brought to south central asia) and well… a little thing called…. language!

Anthro

To me R1b L-51 could be a single clan family bottlenecked in europe ( we have not only villabruna but now also Les iboussieres…R1b seems like a paleoeuropean lineage) that jumped on the "remedello package" and rolled over the continent in BB times.
We'll see.

old europe said...



https://www.academia.edu/31562939/RITUAL_FEATURES_OF_BELL_BEAKERS_IN_SUPRA%C5%9AL._THE_OFFERING_TAKING_POSSESSION_OF_THE_LAND_OR_CULTURAL_INTEGRATION

About cremation rite from eastern europe to south central asia here is a study about a big BB settlement in eastern poland 2500/2000 BC ( you should know dave!) I had the chance to visit in 2014 while on pilgrimage in poland (well I wasn't interested in archeology back then). That is the part of europe that picks fair skin and hair along with a documented presence of the cremation rite . Just to show that I don't talk bullshit.

When_in_Rome said...

For those interested, the Anthromadness blog covered the pigmentation of Prehistoric West Eurasians some time ago:

https://docs.google.com/spreadsheets/d/1ial9mdtlpd2aG2XhTgvvWYnxHCfowOO4wv3OqzSv3C0/edit#gid=356778360

For the Steppe population, they seemed to be 100% fixed for the rs1426654 AA variant, indicating lighter skin, and for the rs16891982 they were 21.8% GG (pale), 40.6% CG (light), and 37.5% CC (olive). Based off of this, it would look like at least 2/3 were similar to modern northern Europeans in skin and the rest similar to Caucasian populations.

Hair color was also covered:
About 3/4 were dark haired, 1/4 were light haired.

For the king said...

Interesting. Does Hajji BA work on modern Iranians/Kurds? Modern Iranians/Kurds prefer Sintashta and Andronovo over Yamnaya(Usually) as a steppe source.

It's definitely not an Iranic sample. Sappali teppe outlier strikes me as an early Iranic elite sample (scores 35% Sintashta, compared to 5-10% Sintashta in average sappali samples). Sappali outlier also produces the best fits for modern Iranians.

Parkhai LBA outlier, also Dzharkutan_BA2 can be some sort of early iranics(Post steppe-BMAC mixture). Both score similar steppe levels to the Swat Aryan samples(A bit higher tho, at 20% on average).

By the way Dave and others, check Shahri-Sokhta BA1. The study models them as 6-8% Steppe, not sure if it's real or not. The Shahri Sokhta location does look Indo European influenced, but it's a bit too early(2650BC to 2550 BC). We know the Hajji feruz CHL outlier did score some Yamnaya(1/8 of their ancestry) during the same period in NW Iran, but that sample is way older than SSBA1.

Anthro Survey said...

@OldEurope and Mike
My point exactly re/hybridization w/Old Europe.

@WhenInRome
Check out Genetiker's new format:
https://genetiker.wordpress.com/pigmentation/

@OldEurope
L51 is sister to z2103. Both are downstream of R-L23, which was born no earlier than 5000BC somewhere in eastern Europe by the looks of it. So, even if upstream R1bs were actually born further to the west(e.g.Villabruna), there had to have been a reflux event.

Mongo said...

From the Nature paper The genomic history of southeastern Europe (https://iris.unipa.it/retrieve/handle/10447/273248/529042/nature25778_proof2_new.pdf):

"One version of the steppe hypothesis of Indo-European language origins suggests that Proto-Indo-European languages developed north of the Black and Caspian seas, and that the earliest-known diverging branch, the Anatolian branch, was spread into Asia Minor by the movements of steppe peoples through the Balkan Peninsula during the Copper Age at around 4000 bc. If this were correct, then one way to detect evidence of the spread of Indo-European languages would be the appearance of large amounts of steppe-related ancestry first in the Balkan Peninsula, and later in Anatolia. However, our data provide no evidence for this scenario. Although we find sporadic steppe-related ancestry in Balkan Copper and Bronze Age individuals, this ancestry is rare until the late Bronze Age. Furthermore, although Bronze Age Anatolian individuals have CHG-related ancestry, they do not have the EHG-related ancestry characteristic of all steppe populations sampled to date or the WHG-related ancestry that is ubiquitous in Neolithic southeastern Europe. We caution, however, that at present we only have data from a small number of Bronze Age Anatolian individuals, none of whom are associated with known Indo-European-speaking populations. An alternative hypothesis is that the homeland of Proto-Indo-European languages was in the Caucasus or in Iran. In this scenario, westward population movement contributed to the dispersal of Anatolian languages, and northward movement and mixture with EHG was responsible for the formation of a ‘Late Proto-Indo European’-speaking population associated with the Yamnaya complex. Although this scenario gains plausibility from our results, it remains possible that Indo-European languages were spread through southeastern Europe into Anatolia without large-scale population movement or admixture."

In other words, the Yamnaya complex would represent a late stage of proto-IE after Anatolian had already split away, with the proto-Anatolians moving directly from the Caucasus region (perhaps Maykop?) to Anatolia.

Davidski said...

@Mongo

The three Early Bronze Age Anatolian samples from Mathieson et al. 2018 don't contradict the steppe hypothesis. They support it.

That's because they were buried in jars. This is a rather typical non-Indo-European Near Eastern burial type.

So they were unlikely to have been Indo-European speakers, or even had any contact with Indo-European speakers. Hence, the fact that they lacked steppe ancestry makes good sense.

Open Genomes said...

@David (& @Nick Patterson (Broad)

I got another 1240k SNP file for DA101 which has the following at Eurogenes K13 below. I filtered on QUAL>=8 and mapQ >=37.
I used the bcftools multiallelic caller instead of the consensus caller, which is better for rare alleles.

http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA101/DA101-Qual-8-mapQ-37.geno

This has 1171887 of the 1198444 1240k non-Y SNPs, 97.8%.

Z452295 in Gedmatch, K13:

Population
North_Atlantic 17.51
Baltic 20.78
West_Med -
West_Asian 26.10
East_Med -
Red_Sea -
South_Asian 10.03
East_Asian 5.30
Siberian 15.93
Amerindian 4.27
Oceanian -
Northeast_African -
Sub-Saharan 0.08

The Oracle4:

1 Nogay @ 19.317848
2 Afghan_Turkmen @ 19.457438
3 Tadjik @ 20.146395
4 Uzbeki @ 22.360464
5 Tatar @ 22.720503
6 Aghan_Hazara @ 22.787352
7 Afghan_Tadjik @ 23.861485
8 Uygur @ 26.278913
9 Hazara @ 26.783249
10 Turkmen @ 28.524866

Using 2 populations approximation:
1 50% Tadjik +50% Tatar @ 9.109774

Using 3 populations approximation:
1 50% Afghan_Pashtun +25% Shors +25% Southwest_Finnish @ 7.618892

Using 4 populations approximation:
1 Altaian + Kalash + Southwest_Finnish + Tabassaran @ 5.755090
2 Altaian + Finnish + Kalash + Tabassaran @ 5.790455
3 Burusho + North_Swedish + Shors + Tabassaran @ 5.998634
4 Hakas + Kalash + North_Swedish + Tabassaran @ 6.022688
5 Burusho + Hakas + North_Swedish + Tabassaran @ 6.024228
6 Kalash + Southwest_Finnish + Tabassaran + Tuvinian @ 6.099697
7 Hakas + Kalash + Southwest_Finnish + Tabassaran @ 6.099822
8 Burusho + Shors + Southwest_Finnish + Tabassaran @ 6.103725
9 Altaian + Kalash + North_Swedish + Tabassaran @ 6.151460
10 Finnish + Kalash + Tabassaran + Tuvinian @ 6.242604

The mean reference alleles is 1.49122. Is that somewhat high?

I think that the K13 and the Oracle4 looks pretty acceptable.

The North Atlantic may be a big high, so this may be reference bias, but otherwise, it looks a lot what we'd expect to see from a Tien Shan Hun, someone of Altaic Proto-Turkic and Uralic ancstry admixed South Central Asian.

If that doesn't work, I can try some other values to lessen the reference bias, like a read depth>=2.

Davidski said...

@Open Genomes

Can you post a 23andMe zip file of these new genotypes?

Davidski said...

@All

So apparently there's no R1a in the ancient Indus Valley samples. No surprise.

The Genetic History of Indians: Are We What We Think We Are?

Open Genomes said...

@David

This file is the 1240k SNPs in 23andMe format, with as many rsIDs as possible instead of Reich Lab names:

http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA101/genome_DA101-1240k-rsIDs-minQual-8-minMapQ-37.zip

This is the same set of 1240k SNPs in 23andMe format but with the exact Reich Lab names:
http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA101/genome_DA101-1240k-Reich_Lab_names-minQual-8-minMapQ-37.zip

And this file is the "consolidated" SNP array SNPs that you would find from 23andMe and the Illumina Axiom chip (FF and Ancestry) more suitable for Gedmatch, but will fewer SNPs:
http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA101/genome_DA101-consolidated_SNP_arrays-minQual-8-minMapQ-37.zip

Whatever makes it easier for you to convert. What's your preferred format? Do you just import 23andMe formatted files into Plink? Or do you eventually convert it to ANCESTRYMAP .geno format?

Onur Dincer said...

@Open Genomes

Do you have the GEDmatch IDs for the samples DA203, DA204, DA222, DA230 and DA124?

Davidski said...

@Open Genomes

This looks much cleaner. I can try to run it in the G25 later today. But if you can tighten up the quality scores some more, that'd be great, because there are lot of markers to work with.

K7 DA101
Ancient_North_Eurasian 36.17
Basal-rich 18.62
East_Eurasian 20.46
Oceanian 0.36
Southeast_Asian 0.19
Sub-Saharan 1.18
Villabruna-related 23.02

Btw, I've got things set up here to convert commercial data files very quickly into PLINK and eigenstrat formats, so it's easy for me to work with these sorts of zip files.

Arza said...

I've averaged semi-randomly 7 out of 10 results from "4 populations approximation":

Hun_predicted,0.0957,-0.0152,0.0116,0.0359,-0.0346,0.0125,0.0072,0.0057,-0.0196,-0.0186,-0.0043,-0.0003,-0.0018,-0.0093,0.0112,0.0045,-0.0045,0,0.002,0.0035,-0.0072,0.0001,0.0003,0.0031,-0.001

Nearest sample is Molaly_MLBA:I3769 (3.362%, scaled)

Open Genomes said...

@Onur

This is what's in Gedmatch so far:
https://docs.google.com/spreadsheets/d/1Rja6ZyjQrz3UK7_HTagkzczoOFjwcXOGdNOm9FC3Rik/edit?usp=sharing

As you see, I'm trying to debug the files right now with David's help, so when I get something working that's "clean enough", I can redo these and do others.

Open Genomes said...

@David

Does your 23andMe format conversion work for any generic 23andMe format file, or does it rely on specific SNP names? Because if you can use just 23andMe format files, I'll run the final version for 1240k with the Reich Lab names and then that saves yet another step and makes it 100% compatible with the Reich Lab tools and datasets.

I'll try different parameters now like read depth >=2, and a higher base quality score, and I'll run that for the commercial SNP array / Gedmatch SNP set and post it below. Then when we settle on something, I can run whatever version you want for the 1240k set.

Davidski said...

@Open Genomes

It works even without actual SNP numbers, just with b37 positions.

But it's easier for me if the SNPs in your files have the same IDs as the Reich Lab 1240k SNPs.

Onur Dincer said...

@Open Genomes

Thank you for your efforts! According to the leaked info, DA203, DA204, DA222 and DA230 are from medieval Turkic populations (Karakhanids and Karluks), not sure about DA124 though due to the lack of info on ethnicity.

See: https://anthrogenica.com/showthread.php?13821-Siberian-aDNA-and-Turkic-Iranic-and-Uralic-populations&p=383122&viewfull=1#post383122

Unknown said...

@ Open Genomes

I wouldn’t pay much attention to mapping quality scores since those are driven by number of base mismatches to reference, and since aDNA is very fragmented andthe reads are short, they can often be mappedto multiple sites because of repeats in the reference. Thus an aligner can reporta high score (few base mismatches) when in fact the read has been mapped to the wrong site, and conversly report a low score because of variation between the subject and the reference, or sequencing errors. Studies have shown the reported mapping scores to be inaccurate.

With regards to biases you’ll never get rid of them. It’s something that’s inherent in this field. Hg38 is a vast improvement over hg19, but is tricky to use properly.

For the past year I’ve been using the ATLAS ( Analysis Tools for Low-coverage and Ancient Samples) pipeline and have been happy with the results. Using it you can also account for post-mortem damage
https://bitbucket.org/phaentu/atlas/wiki/Home

Open Genomes said...

@David

Here are the DA101 23andMe files for read depth=1 (no filter), map quality=37, and base qualities from 8 to 15, and the same for read depth=2.

http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA101/

The filter used is listed in a comment in the second line of the file.

I think that a read depth=2 cuts down the number of SNPs considerably. You can decide on the break-even point of the base quality scores vs. the total SNPs and then tell me. Then I'll make the 1240k 23andMe and .geno files with those parameters too (using the Reich Lab names).

Open Genomes said...

@Kurd Dgk

Thank you for the link to ATLAS. It certainly would be very helpful for working with WGS.

However, what I'm doing here is a special "limited" case, because the analysis we're doing right now (unfortunately) is based only on discrete set of SNPs.

1. While the mtDNA is based on a complete alignment, the read depths are high, and the two rCRS and RSRS references are close enough to the actual sequence to work with just a regular variant caller. However, I'll try ATLAS for the mtDNA as well.

2. What I'm trying to do with the Y is to call *all* known / named SNPs, and report the allele depths along with the ancestral and derived states in a spreadsheet that anyone can analyze themselves. I had to adjust the bcftools parameters to make certain that all alleles were called in all cases. Determining whether a sample is derived for a particular SNP / haplogroup is very much a "judgement" call that has to be done heruristically. That's because at the haplogroup we're dealing with a known sequence, practically down to the base, so we have to do is check a few specific SNPs at that clade level and immediately below. I also have a tool to create the same records for "novel" SNPs (with a depth filter, 2 works well). I've actually seen a case where a single short read that wasn't of very good quality had a derived SNP phylogenetically in the right place.

I think it's just better to give everyone a summary of the data (the alleles and allele counts) and let people figure it out on their own. (It's not like when certain people "highlight a SNP in bold on a webpage and then claim it's derived. ;) ) Also, as we know, many Y SNPs aren't on the ISOGG Tree (really!) and YFull doesn't show SNPs that only appear in one sample - even when that sample is clearly in a subclade. (That's because often "subclades" in ISOGG are completely wrong.) There are plenty of :"private" trees that list all the SNPs for each sample, and those people can look at the "novel SNPs" table and maybe see something that others wouldn't notice.

3. For the autosomal, because of the analysis that we've been doing which is based on a defined set of SNPs (the Reich Lab 1240k list or even the commercial SNP arrays / "Gedmatch template" list), it's useful again to just slice the BAM by these lists and just call them.

Of course, this approach of just capturing a relatively random set of SNPs (not the Affymetrix Human Origins array but the additional Illumina SNPs) completely ignores haplotyping and in fact almost makes that impossible. As you know, the Copenhagen Lab tries to do WGSs without a capture array, the a recent Veeramah (2018) study used a specially designed "neutralome" capture array, if I read it correctly. (A kind of "reverse exome" chip.)

Sooner or later, a wide-ranging or WGS capture array going to be the standard for aDNA instead of just SNPs, and that's when something like ATLAS will really come in handy. (In the meantime, rumor has it that the Reich Lab is designing a Y capture array just like the have for the mtDNA.)

We've actually been experimentally realgining the Roman soldier FN_2 to see if we can get more aligned reads out of the WGS. We'll try ATLAS on FN_2.

Open Genomes said...

@David,

On Kurd Dgk's advice, I also generated 23andMe commercial SNP array files without any map quality filter, with base qualities from 9 to 15 and no read depth filter:

http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA101/

Using a calculator to validate the accuracy and completeness of these SNP calls is of course a great idea. I'll also try to upload all of these to Gedmatch and see if we can figure out the "right balance" between the number of SNPs and bad calls / damage using K15 and Oracle4. So far though, K15 Oracle4 made a lot of sense even for the unfiltered files, although he distances were quite high. We'll see.

I think that this Eurasian Steppe set has some very important individuals in it:

https://docs.google.com/spreadsheets/d/1Rja6ZyjQrz3UK7_HTagkzczoOFjwcXOGdNOm9FC3Rik/edit?usp=sharing

1. An R-U106* - from the Steppe?
2. An R-P312* (perhaps from Hungary)
3. An E-V22 who really does look to be South Central Asian
4. 3 Haplogroup O samples, O-PK4, O-F4062, and O-F871, who may be our first aDNA East Asian WGSs
5. C-Y11990 and C-Y4541 and a series of Haplogroup Qs who may be representative of the Siberian Neolithic
6. An N-L1025 who would be representative of the western Uralics (and also the Rurikids and Gedeminids)
7. J-Y13534 (Timurids / Barlas?) who may be representative of the BMAC culture, if they're early enough.

I haven't even done the mtDNA for any of them, except two.

I would say that as interesting as all of these are,the O East Asians will fill in a big gap in our calculations. So far we've just been guessing about where Neolithic East Asians (North Chinese and Koreans) fit.

Davidski said...

@Open Genomes

The potential Uralic individual belonging to N-L1025 would be nice to have.

Open Genomes said...

@David, maybe he's a Gedeminid. It seems that they study considers Hungary and Poland part of the "Eurasian Steppe" (which it is) there's clearly a Late Neolithic G-PF3378 who I think seems to be from Hungary.

I'll download that one then and then you can decide what parameters I should use to filter the SNPs.

Davidski said...

@Open Genomes

Pretty sure Poland isn't part of the Eurasian steppe, but the Hungarian Plain or Carpathian Basin is sometimes considered an extension of it.

By the way, these files produced less than 1% Sub-Saharan admixture, so these sorts of parameters should be fine.

DA101-consolidated_SNP_arrays-minQual-9-minMapQ-37.txt

DA101-consolidated_SNP_arrays-minQual-15-minMapQ-37.txt

Karl_K said...

@Davidski

Wow. Rai and Shinde have seriously changed their tune in the last year.

“How do I say it? See, I am a nationalist,” Rai says over the phone. “People will be upset. But that’s how it is. All the studies are showing that people came here from elsewhere.”

Ric Hern said...

I think it will surely be a problem for the linguistic hypothesis that Hittite resembles the oldest form of Indo-European if they only spread via the Caucasus into Anatolia at +-2400 BCE.

Wouldn't the oldest or older form of Indo-European then be closest to people who spread earlier from the epicentre ? We see that European migrants during the 1400s and 1600s retained a lot of the earlier language forms of their countries of origin than the people who stayed put in that countries.

We see Steppe related Bell Beaker and Corded Ware peoples already in Central Europe prior to 2400 BCE. and even Steppe related ancestry arriving in Ireland at the time when the Hittite Culture was just an Archaeological embryo.

Karl_K said...

@Ric

"Steppe related ancestry arriving in Ireland at the time when the Hittite Culture was just an Archaeological embryo."

Right. But many here will just tell you that the languages of all of Europe, and especially the British Isles were totally replaced more recently than that by people with nearly identical genetics, and minor genetic influence. That is the fall back argument.

EastPole said...

This is interesting:

https://twitter.com/amwkim/status/988532420831010817


https://en.wikipedia.org/wiki/Volga_River#Nomenclature

But Rava is also a Slavic river name:

https://en.wikipedia.org/wiki/Rawa_(river)

It is also cognate with Laba (Elbe) l:r, b:v.
So it can be a CWC river name.

Ric Hern said...

@ Karl_K

My argument against this fall back argument is this:

Irelands Neolithic population was replaced by Central European related Steppe/Farmer admixtures. They did not really mix significantly with Neolithic Irish. This happened within +-400 years from Central Europe to Ireland.

Now compare this with the Split between Dutch, Flemish and Afrikaans Languages. Afrikaans have been separated from Dutch for nearly 400 years but still Afrikaans people can understand Dutch and Flemish people fairly well. You can almost just call them dialects of each other.

So why would nearly identical genetic people have spoken something totally different ?

Open Genomes said...

@David

The break-even point is a base quality of >=10, with no other filtering.
There's no spurious noise and the percentages hardly change with any additional filtering.

Here's the base quality >=10 DA101 23andMe format .zip file for the 1240k SNPs with the Reich Lab names:

http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA101/genome_DA101-1240k.zip

1070742 out of maximum of 1198444 SNPs, 89.34%.

You should be able to put that directly in Global25.

K13 results for DA101, base quality >=10, for the "Gedmatch" SNPs:

Kit Number: Z239204 Elapsed Time: 10.74 seconds

Population
North_Atlantic 17.84
Baltic 20.67
West_Med -
West_Asian 26.73
East_Med -
Red_Sea -
South_Asian 9.88
East_Asian 4.20
Siberian 16.50
Amerindian 4.17
Oceanian -
Northeast_African -
Sub-Saharan -

Oracle4:

Using 1 population approximation:
1 Nogay @ 19.474754
2 Afghan_Turkmen @ 19.772781
3 Tadjik @ 20.165670
4 Uzbeki @ 23.036974
5 Tatar @ 23.060909
6 Aghan_Hazara @ 23.450630
7 Afghan_Tadjik @ 24.060871
8 Uygur @ 27.242842
9 Hazara @ 27.496092
10 Turkmen @ 28.649939
11 Tabassaran @ 29.153820
12 Afghan_Pashtun @ 29.166658
13 Chuvash @ 30.268799
14 Kabardin @ 30.772602
15 Chechen @ 30.788715
16 Lezgin @ 31.560823
17 Kumyk @ 31.789127
18 Mari @ 32.297440
19 Balkar @ 32.650303
20 Kazakh @ 34.121143

Using 2 populations approximation:
1 50% Tadjik +50% Tatar @ 9.549594

Using 3 populations approximation:
1 50% Afghan_Pashtun +25% Shors +25% Southwest_Finnish @ 8.114545

Using 4 populations approximation:

1 Altaian + Kalash + Southwest_Finnish + Tabassaran @ 6.334741
2 Kalash + Ket + Swedish + Tabassaran @ 6.346260
3 Kalash + Ket + North_Swedish + Tabassaran @ 6.373364
4 Hakas + Kalash + North_Swedish + Tabassaran @ 6.406739
5 Altaian + Finnish + Kalash + Tabassaran @ 6.431836
6 Balochi + Ket + La_Brana-1 + Tabassaran @ 6.479008
7 Kalash + Ket + Norwegian + Tabassaran @ 6.486293
8 Kalash + Selkup + Swedish + Tabassaran @ 6.541377
9 Hakas + Kalash + Southwest_Finnish + Tabassaran @ 6.548597
10 Burusho + North_Swedish + Shors + Tabassaran @ 6.557090
11 Kalash + Norwegian + Selkup + Tabassaran @ 6.596027
12 Altaian + Kalash + North_Swedish + Tabassaran @ 6.625070
13 Kalash + North_Swedish + Shors + Tabassaran @ 6.627977
14 Kalash + Southwest_Finnish + Tabassaran + Tuvinian @ 6.660076
15 Kalash + Shors + Southwest_Finnish + Tabassaran @ 6.665052
16 Burusho + Ket + Swedish + Tabassaran @ 6.694208
17 Burusho + Hakas + North_Swedish + Tabassaran @ 6.703238
18 Burusho + Selkup + Swedish + Tabassaran @ 6.707608
19 Burusho + Norwegian + Selkup + Tabassaran @ 6.708549
20 Burusho + Shors + Southwest_Finnish + Tabassaran @ 6.723035

Interesting he has Altaian, Khakas (Hakas) who are "Tadars", Kets, and Southwest Finnish show up. This East Hun clearly an Altaic Siberian with some Uralic admixture whose ancestors admixed with a Kalash + Tabassaran CHG / Iranian Neolitic like population when they moved south.

I'm downloading DA171, the N-L1025 right now.

Rob said...
This comment has been removed by the author.
postneo said...

@david
Haji Firuz and other r1b from the same region don’t cluster together on the pca. Why is that?

Karl_K said...

@postneo

The age of the samples from that region are likely quite vast, and there could have been a lot of migration of various peoples. If they can't be definitely linked to each other by carbon dating or archaeology, then there is no reason they should be linked genetically.

Jijnasu said...

Dr.Rai and Dr.Shinde have handled this pretty maturely. Hope online OITists will do so too. While some may considered the Indo-European question nearly closed, this is probably just the begining. A simplistic equation of steppe (though IE likely expanded from or nearthe steppe) with Indo-European languages doesn't seem to explain everything. It also overlooks complex processes of interaction and assimilation. The most important question as to why these languages were so successful hasn't been explained yet. Hopefully as we gain access to more data these problems will eventually be better explained.

Karl_K said...

@ Jijnasu

I agree. It is very encouraging to see scientists of the region embrace the facts unearthed by their own work, despite their original hypotheses.

Anonymous said...

@Anthro Survey
GAC probably picked up its light eye/skin/hair phenotypes from the WHG, SHG and EHG. We know blonde hair is from Afontova Gora, we know blue eyes are from WHG and we know that the ANF lacked a light skin gene (although they had the other).
And even then, the ANF already had some share of WHG/EHG mix before entering Europe, so it justifies their outliers.

So, for Corded Ware, we know they didn't mix with GAC (although Ukraine_Eneolithic probably did), so they got their light stuff from Narva, the "original source", and so probably every Steppe_MLBA guys.

Sexual selection did the rest.

Karl_K said...

About hair color genetics,

There is a new study out (mostly about Europeans) that found 124 genes that account for only 21% of hair color variability.

So... I would be very careful trying to determine the hair color of ancient samples.

https://www.nature.com/articles/s41588-018-0100-5

Aram said...

Anthro Survey & Mike the Jedi

I agree that hybridisation with EEF was important for the appearence of more lighter phenotypes in Steppe. Imho thisvHJ-BA in most likelihood do has some ANF/EEF because it is shifted slightly toward 'west'.

I am generally not interested in phenotypes but this particular case is interesting to test some theories about namru.

Ric Hern said...

@ Karl_K

"The age of the samples from that region are likely quite vast, and there could have been a lot of migration of various peoples. If they can't be definitely linked to each other by carbon dating or archaeology, then there is no reason they should be linked genetically."

Maybe this points to some kind of trading outposts of Steppe people through the centuries ? Maybe Steppe people wanted to cut out the Middleman (Maykop) ? Maybe some unified to form the Gutians and others migrated to Anatolia.

Aram said...

Those folks also could be the erectors of Hakkari kurganic stellaes in SE Turkey.

----

The earliest stelae are in the style of bas relief while the latest ones are in a linear style. They were made during a period from the fifteenth century BC to the eleventh century BC in Hakkari. Stelae with this type of relief are not common in the ancient Near East however there are many close parallels between these and those produced by a variety of peoples from the Eurasian steppes between the third millennium BC and the eleventh century AD.[6]

https://en.m.wikipedia.org/wiki/Hakkâri#Hakkari_kurgan_stelae

Archaelog said...

Some excerpts from the article at http://www.openthemagazine.com/article/science/the-genetic-history-of-indians-are-we-what-we-think-we-are

"In the past, genetic studies have thrown up somewhat confusing results. Those focused on Mitochondrial DNA, which is transmitted only from mother to daughter, had suggested very little external infusion into the Indian gene pool over the past 12,500 years. But now that researchers have also traced the descent over millennia of the Y-chromosome, which is transmitted from father to son, they have found that about 17.5 per cent of Indian male lineage belongs to the haplogroup R1a, which is found today across Europe and Central and South Asia, leading some scientists to suggest that the movement into India was probably of male migrants, which would explain why Mitochondrial DNA studies did not show such a pattern of gene dispersal."

Apart from the obvious mistake of saying that mtDNA is transmitted only from mother to daughter, they have not stopped saying the migration was solely a male one. Obviously the migration included both males and females since this was not a military expedition or campaign like the case of Huns or Turks. This was a large scale migration and settlement which could produce a language and cultural shift. Y chromosome haplotypes can undergo founder effects unlike the female equivalent. That is why Y lineages in India show variance from the Bronze Age while mtDNA doesn't.

Open Genomes said...

@David

DA171, Y N-L1025 mtDNA H2a1

1240k 23andMe file:

http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA171/genome_DA171-1240k.zip

125905 SNPs out of 1198444, 10.5%

Gedmatch Z713550

K13:

Population
North_Atlantic 25.73
Baltic 60.13
West_Med 1.61
West_Asian -
East_Med -
Red_Sea 4.50
South_Asian 3.23
East_Asian -
Siberian 2.03
Amerindian -
Oceanian -
Northeast_African -
Sub-Saharan 2.76

Oracle 4:

Using 1 population approximation:
1 Lithuanian @ 11.114440
2 Estonian @ 14.567318
3 East_Finnish @ 14.657040
4 Belorussian @ 15.033764
5 Estonian_Polish @ 15.330928
6 Erzya @ 15.987166
7 Kargopol_Russian @ 16.076241
8 Russian_Smolensk @ 16.181431
9 Finnish @ 17.082684
10 Southwest_Russian @ 18.008085
11 Ukrainian_Belgorod @ 18.419487
12 Polish @ 19.588999
13 Ukrainian @ 22.129522
14 Southwest_Finnish @ 22.689356
15 Ukrainian_Lviv @ 23.297407
16 South_Polish @ 23.523617
17 La_Brana-1 @ 23.806120
18 Chuvash @ 29.127838
19 Croatian @ 30.396633
20 North_Swedish @ 31.556002

Using 2 populations approximation:
1 50% Lithuanian +50% Lithuanian @ 11.114440


Using 3 populations approximation:
1 50% Lithuanian +25% Lithuanian +25% Lithuanian @ 11.114440


Using 4 populations approximation:
1 Lithuanian + Lithuanian + Lithuanian + Lithuanian @ 11.114440
2 East_Finnish + Lithuanian + Lithuanian + Lithuanian @ 11.315384
3 Erzya + Lithuanian + Lithuanian + Lithuanian @ 11.508182
4 Estonian + Lithuanian + Lithuanian + Lithuanian @ 11.761994
5 Kargopol_Russian + Lithuanian + Lithuanian + Lithuanian @ 11.880618
6 East_Finnish + Erzya + Lithuanian + Lithuanian @ 11.926003
7 Belorussian + Lithuanian + Lithuanian + Lithuanian @ 11.948862
8 Estonian_Polish + Lithuanian + Lithuanian + Lithuanian @ 12.026574
9 Finnish + Lithuanian + Lithuanian + Lithuanian @ 12.027840
10 East_Finnish + East_Finnish + Lithuanian + Lithuanian @ 12.036226
11 Erzya + Estonian + Lithuanian + Lithuanian @ 12.087805
12 East_Finnish + Estonian + Lithuanian + Lithuanian @ 12.205635
13 Belorussian + East_Finnish + Lithuanian + Lithuanian @ 12.236085
14 Lithuanian + Lithuanian + Lithuanian + Russian_Smolensk @ 12.249960
15 East_Finnish + Estonian_Polish + Lithuanian + Lithuanian @ 12.296120
16 East_Finnish + Kargopol_Russian + Lithuanian + Lithuanian @ 12.368361
17 Belorussian + Erzya + Lithuanian + Lithuanian @ 12.402226
18 Erzya + Finnish + Lithuanian + Lithuanian @ 12.447273
19 East_Finnish + Lithuanian + Lithuanian + Russian_Smolensk @ 12.491582
20 Erzya + Estonian_Polish + Lithuanian + Lithuanian @ 12.540218

Surprise! Lithuanian. Not. I wonder when he's from? After all, he matches the Grand Dukes of Lithuania ... ;)

Davidski said...

@All

The Hun from the Tien Shan is now in the Global25 datasheets. Very interesting, but not unexpected, results...

Hunnic_Tien_Shan:DA101

Scythian_ZevakinoChilikta,44
Dali_MLBA,43.2
Sarazm_Eneolithic,8.8
Shahr_I_Sokhta_BA3,4

[1] distance%=2.7356

https://drive.google.com/file/d/1FSzKKknFGcOgfyA76q9PD7B-n-MJs7L8/view?usp=sharing

vsva34 said...

Open genoms: Can you check y snp calls i4550 Q Zvejnieki sample from mathieson 2018 if he is ancestral to nordic Q?

Synome said...

@Davidski

Fascinating!

I always imagined the Huns to be more of an East Scythian/"Mongola" type mixture. I wonder how much substructure we can find in the Huns, who were after all a steppe confederation? Maybe the only common ancestry element is the Scythian, and cultural conglomeration did the rest.

Arza said...

Mordovian
Lithuanian 71.5%
Hunnic_Tien_Shan:DA101 28.5%

Distance 1.986%

Archaelog said...

I wonder where the Turkic cultural element comes from in the Tien Shan Hun sample. Either from the Dali element, in which case a Proto-Turkic language could have been spoken in that region. Or it could come from the Scythian part that had already been Turkicized.

Synome said...

I reviewed some of the findings on Y-HGs from Xiongnu remains. There seems to be a high incidence of Q. I believe there is a strong connection between the Xiongnu and Huns.

That, together with this evidence for Dali_EBA ancestry in a Hunnic individual suggests to me that we need to now take seriously that there may have been a large Siberian contribution to the Xiongnu/Huns, and this would lend support to the theory that at least some of them spoke a Siberian language, maybe even Yeniseian.

mzp1 said...

Rai said he was an Invasionists from the beginning and he is talking about a lack of R1A from a sample size of ONE.

velvetgunther said...

@mzp1
"Rai said he was an Invasionists from the beginning and he is talking about a lack of R1A from a sample size of ONE."

A sample size of one from Rakhi Garhi. Surely they have other samples from India.

bellbeakerblogger said...

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0195491
Estimating genetic kin relationship in prehistoric populations

Karl_K said...

@Salden

Read the recent interviews. Of the 4 promising samples, they only got usable DNA from one petrous bone.

Karl_K said...

And, Shinde is an author on the new pre-print. Read his recent interviews, he emphasizes that there were IVC people going out to other locations, but agrees with the direction of language spread. He also believes that modern Indian Culture was developed during the mixing of people, but does not rule out that the physical location could have been within India.

Tesmos said...

@Open Genomes

Could you also checkout DA119 or/and DA111?

velvetgunther said...

@Karl_K
"...He also believes that modern Indian Culture was developed during the mixing of people, but does not rule out that the physical location could have been within India."

Does anybody dispute that? I don't think anybody claims that modern Indian culture as we know today was derived lock, stock and barrel from the Steppe.

Karl_K said...

@velvetgunther

I wasn't suggesting anyone disagreed with that ever.

I was just mentioning where he has shifted his focus when speaking, which is away from out-of-india topics.

Slumbery said...

@Synome

The fans of the theory that the Huns actually spoke ancient Hungarian (apparently supported by Eupedia) can get an orgasm from such a comment. :)

@Open Genomes
There are people who argued here that Basques are a real problem for accepting R1b as an IE marker. If that were true then Lithuanians (and the Slavified Baltics east of them) would be a deadly poison for the N1c as the core Uralic marker theory. Just saying.

Sanuj said...

@Karl_K

Can you point me to any of Dr. Shinde's interview regarding his view on direction of language flow?

"Shinde does not like the word ‘migration’. “It is better to say movement,” he says, implying a two-way pattern. “Everyone back then was moving to and fro. Some people were moving here and some were moving out. There was contact, yes. There was trade. But local people were involved in the development of several things. So I am not very sure of the interpretation.” - Yep agreed.

Also, Dr. Rai mentions that R1a is missing in the one Rakhigarhi sample. He also said in the earlier interview that the sample matches NI Brahmin - which i would construe as showing Steppe-like ancestry.

mzp1 said...

Yeah, that is what is being claimed. All Indo-Aryan culture is derived from Vedic, which as you are so certain about, was just a Steppe Culture imposed on South Asians.

According to you India is more influenced by Steppe Culture than any other IE society, given it has more elements of IE culture than any other society.

Aryan Migration Theory posits a civilizational shift the size and nature of which is unmatched by anything in recorded history. You guys need to understand it is a highly speculative and far-fetched position to begin with.

We need real hard evidence to entertain such an extreme scenario. Pointing out one genome that is R1A negative is not anywhere near good enough.

Kristiina said...

@ Slumbery

N1c has not been detected in Kivutkalns and it is possible that it arrived in Lithuania only after 300 BC. If you suggest that this apparently late N1c brought the Lithuanian language to Lithuania, what do you think that Kivutkalns R1a1 spoke? Personally, I do not find it problematic that N1c changed to Lithuanian language in Lithuania. In any case, at the moment we have very little evidence to claim that N1c-L550 spoke Proto-Baltic. Of course, we need to know the context of this new N1c sample.

Jijnasu said...

@sanuj @mzp
You guys need to read what that article has to say first. Dr. Shinde being an archaeologist is merely emphasizing from his field of expertise that the evolution of late bronze age cultures in India involved several complex process including evolution and assimilation but in no way rules out movement. Dr. Rai whom you've been quoting seems to be fairly certain of significant migration from central asia (despite being a longtime critic of the Idea). The migration from central asia is pretty much established. The only argument for OIT now is to say that language has nothing to do with genetics, as Talageri has. But I think that's a very weak argument. BTW there are several examples of language change even in densely populated regions and the harappans were a steadily declining culture in the proposed timeframe so not impossible

Aniasi said...

@Davidski

How should we understand the SIS and Sarazm proportions for the Hun? Does this indicate South Asian or related ancestry?

velvetgunther said...

@Karl_K
In the face of overwhelming evidence, I think Rai and Shinde have done the wise thing. If they at all want any stake in future research, they must have thought that it's best to go ahead and publish the data and accept the findings, rather than sit on the data and remain isolated while the rest of the world moves on. I don't know if they are being honourable, as somebody here suggested, it's more likely they just want to cover their *sses. Back in 2013, Shinde had said (and I can't find any reference right now but if I have to, I will) that his sole aim in carrying out this research was to PROVE that there has been a demographic continuity from the Indus Valley till today. There was also an interesting quote from Rai in this article, on the lines of "I am a nationalist..." Why does that even matter? Can one not be a nationalist unless one can prove that there has been demographic continuity for the last 8000 years? If that was the case, then there would not have been any European or North American nationalists, which is clearly not the case!

@ mzp1
I don't know know who you are referring to as 'you', but I hope it's not me as I never said any of those things.

Jijnasu said...

@velvetgunther
The archaeological position has always been that there was population continuity in india and that any language change only involved diffusion. Also 'nationalist' means something quite different in the Indian context. A lot of Indian scientists would describe themselves as nationalists as opposed to the social science academics who are often associated with activism often seen as bordering on being disruptive. Questioning their intentions without proof smacks of racism

Sanuj said...

@jijnasu

Archaeologists do have a great say in studies of cultures where we don't have literary evidence, so Dr. Shinde's views matter a lot. I am also not arguing against any movement, of course people have always been moving. But, imposing a linguistic theory based on moderate admixtures, and cherry picking some group to show elevated admixture is not a strong argument. I could also argue that Brahmins were elites, they would mix with exotic ancestries more than others.

That Steppe ancestry is THE IE marker in India is a pre-conceived conclusion that people make. There have been a lot of people coming in from C Asia to India, but IE has always been here, how is it that only "this" particular movement brings IE?

"BTW there are several examples of language change even in densely populated regions and the harappans were a steadily declining culture in the proposed timeframe so not impossible"
Declining, as in already having Iron workings in 1800 BC https://www.cambridge.org/core/journals/antiquity/article/origins-of-iron-working-in-india-new-evidence-from-the-central-ganga-plain-and-the-eastern-vindhyas/3986B90D94F333149BE7861683609BE9
Steppe traders, overcame all that, even though they were in a big monority, spread from C Asia to Bangladesh and remotest mountains of Nepal, without an elite domination, without any evidence of Uralic in their language even though they were in the same vicinity for many centuries. I would say that you are drawing a very simplistic picture to just wish away the questions.

Also, Davidski has been saying that R1a is an EHG marker, but because WSHG and EHG have a common source Neolithic modeling cannot disambiguate the two in this qpAdm setup, and we now know that WSHG was already there in Indus Periphery.

velvetgunther said...

@Jijnasu
I agree with you. I would like to draw a difference between Indian nationalism and Hindutva nationalism, and I know the difference when I see it. I personally find nothing wrong with healthy nationalism, but Hindutva nationalism isn't it.

mzp1 said...

Steppe Scythian Saka invade India and become 'Indianized' completely, adopting Buddhism and Sanksrit.

Steppe Mongols Invade Central Asia and become Persian speakers and Islamized.

Steppe Scythian Parni free parts of 'Iran' from Greek rule, take up local Western Iranian Language and become Zoroastrian.

Skythians and Huns invade Europe and become assimilated.

vs

Primitive Steppe nomads invade India and Central Asia, an extremely advanced and wealthy civilization, and leave no trace of any earlier history, language or culture from the Gangetic plain all the way to the shores of the Caspian Sea.

The only Steppe language imposed outside of the Steppe would be in Turkey, and I dont know how populated this region was at the time.

But I suppose it is possible because the IVC people didnt know war, they were just peaceful workers ready to be exploited by the incoming 'warlike' steppe nomads. I mean the IVC peopled likes toys and games, chess for instance...

https://www.openart.in/wp-content/uploads/2016/09/i_board.jpg

But wait, isnt Chess a game about war...

Vedic Aryans and IEs in general enjoyed gambling. The IVC didnt have any of that did they...

http://downloads.bbc.co.uk/rmhttp/schools/primaryhistory/images/indus_valley/what_they_did_for_us/i_dice.jpg

and oh yeah, these guys didnt have spoked-wheels or chariots, obviously the greatest civilization on Earth at the time, with great cities and even toy models of carts for children to play with, couldnt come up with the technological ingenuity to place wooden spokes inplace of a solid wooden wheel. Nope, only basic steppe nomads were able to make that giant leap.

Like I said, the Aryan Migration Theory is way, way out there. Genetics is one thing, but we need a much more complete argument to take it seriously.

Jijnasu said...

@sanuj
While the how isn't clear yet, if as everything so far seems to indicate r1a reached India in the 2nd millenium BCE, the men who carried it were pretty dominant - considering the fact that it occurs at high frequencies in most northern caste groups (30% even in 'lower caste' chamars) so why not their language as well? Famillies being patrilocal and patrilineal, understandable that it was the mens' language and culture that would dominate. You could as you argue that this was the movement of some unknown ethnic group, assimilated into an already indo-aryan population but that seems like a poor explanation. Since you brought up brahmins, study of the veda was supposed to commence early at the age of 7 and involve 12 years of study, besides the observance of several taboos. Do you think a group of migrants could have supplanted a large part of this priestly population easily and without affecting the oral transmission of these hymns?

Jijnasu said...

The sumerians were the founders of some of the first cities in the world. They didn't think of using spoked chariots either. Also their languages eventually went extinct. How do you suppose that happened

mzp1 said...

Sanuj,

We dont have enough data yet to determine conclusively the genetics of the IVC in terms of Steppe DNA or R1A.

Why are you trying to make arguments without sufficient data, it just looks like you are trying to push a-priori positions.

mzp1 said...

"The sumerians were the founders of some of the first cities in the world. They didn't think of using spoked chariots either. Also their languages eventually went extinct. How do you suppose that happened"

I dont know, why dont you tell me and make your point. Were they overthrown by Scythians and end up Iranian speakers? If not what is your point?

Jijnasu said...

All I'm saying is languages and cultures even 'great' ones have been known to get replaced. We don't know everything yet but evidence points to some version of the steppe theory atleast for late PIE now.

mzp1 said...

What's 'Late' PIE?

Samuel Andrews said...

Imo, people will exaggerate the mixed and exotic origin of Indians. Few have even 20% real European-sintashta ancestry. Many have %10 or less.

The Aryan invasion happened 3,500yo. If you go back to the neolithic most of their ancestors lived in India or somewhere near by.

Ancient DNA doesnt support myths of millions of years of continuum but it doesnt show a history defined by constant foreign migration and mixing.

Palacista said...

@mzp1
in 1500 BCE the IVC was long dead, there were people living in the area of the IVC but the urban civilisation was no more. The IE invaders were not primitive, they had a considerable cultural weight to go with their military prowess and technology.

As for leaving no trace of the IVC civilisation, there was nothing left to wipe out, on the other hand the economy and culture of the rural population must have contributed to the make up of the culture that was created in iron age India, a culture that was created in India from internal and external sources.

Sanuj said...

@Palacista

There is no evidence of any invasion. Stop using evidence of a putative movement to extrapolate into an invasion without evidence.

Iron age started in 1800 BC or before. This is radiocarbon dated proven evidence - https://www.cambridge.org/core/journals/antiquity/article/origins-of-iron-working-in-india-new-evidence-from-the-central-ganga-plain-and-the-eastern-vindhyas/3986B90D94F333149BE7861683609BE9
By 1500 BC bronze age people from Steppe would be easy meat. Also, as i have said before Vedic Samhita is Bronze Age, but we already have Iron works in 1800 BC. This is real solid evidence, not some presumptive linguistic theory.

Matt said...

@Davidski, quite interesting re: Tien Shan Hunnic sample.

I made some quick graphics, firstly placing the samples on PC1 and PC2 compared to other ancients, then some neighbour joining and distance comparisons: https://imgur.com/a/kMejaQE

This sample looks to be similar to the Scythian from Aldybel, and Alakul Andronovo sample in West-East Eurasian related content. But in terms of the different type of West Eurasian content he/she is more genetically southern/Turan than either by the look of it.

Slumbery said...

@Kristiina

The conditional phrasing of my statement was intentional, I am not against the idea that N1c is a mostly specific Uralic marker. But just for the sake of argument:

"N1c has not been detected in Kivutkalns and it is possible that it arrived in Lithuania only after 300 BC."

"Possible" is not a very strong phrasing here. Of course it is possible. I do not think it is likely however.

"If you suggest that this apparently late N1c brought the Lithuanian language to Lithuania, what do you think that Kivutkalns R1a1 spoke? "

I do not believe in haplotype specific languages. Correlation is possible and happens, especially at a time close to the birth of a haplotype, but I do not see why an ancient group that is mostly N1c could not speak the exact same language as another roughly contemporary ancient group that is mostly R1a1. Or conversely, I do not see why two N1c dominated groups that are roughly contemporary and lived significantly after the first spread of N1c could not speak completely different languages.

Lithuanians have higher N1c ratio than Latvians (who are heavily mixed with Uralics, including very recent assimilation) and Estonians. If that could happen from a recent (after 300 BC) founder effect without effecting the language whatsoever, then anything can happen and haplotypes come out as a very unreliable marker for the language of ancient groups (at least haplotypes alone, without other information).

EastPole said...

@mzp1
“Primitive Steppe nomads invade India and Central Asia, an extremely advanced and wealthy civilization, and leave no trace of any earlier history, language or culture from the Gangetic plain all the way to the shores of the Caspian Sea.”

I guess you are referring to a theory as presented on this picture:

https://2.bp.blogspot.com/-NgS98pXc5xQ/WsBG3hBeptI/AAAAAAAAGmk/lPfc18ty8FMEhIdRIyoKK2O9qWEDH4RCQCLcBGAs/s1600/Narasimhan_et_al_Fig_4_Tale.png

And after Narasimhan et al. publication that theory has been repeated in many articles for example in “The Economist”:

“The wider study not only confirms that “Aryans” (geneticists avoid the term) probably migrated from the steppes around the Volga and Don rivers to both India and Europe at around the same time.”

https://www.economist.com/sites/default/files/imagecache/640-width/images/print-edition/20180407_ASM994.png

https://www.economist.com/news/asia/21740048-aryans-did-not-come-india-they-conquered-it-new-study-squelches-treasured-theory-about

Or this:

http://www.sciencemag.org/news/2018/04/south-asians-are-descended-mix-farmers-herders-and-hunter-gatherers-ancient-dna-reveals

Or this:

http://www.sciencemag.org/news/2018/04/did-ancient-mesopotamians-get-high-near-eastern-rituals-may-have-included-opium?utm_source=sciencemagazine&utm_medium=facebook-text&utm_campaign=gethigh-19011


I agree with you and think that above theory is not right. It was not the migrations of primitive Yamnaya nomads to India and to Greece which resulted in some cultural changes. It was not Yamnaya culture which influenced India and Greece. People who influenced India and Greece were a mixture of nomads from Sredny Stog II and highly advanced rural civilizations of Central-Eastern Europe like Tripolye, GAC and TRB.
There are many reasons to believe that their mixed cultures, languages, religions were quite advanced and could influence other advanced cultures like IVC or Minoan civilization.

Anonymous said...

Frankly, I think they have one good sample and four partial, from reading the article. Fits with what Nick Patterson said here.

So, 4 x mtDNA, 1 x Y-DNA + autosomal. I predict Y-DNA L, solely on the gossip that has dripped though.

mzp1 said...

@Eastpole

Why are quoting the Economist when we have already discussed the paper here in greater detail. Just go back and look at the last 100 comments on the first post.

So we should just accept it because the Economist wrote that?

I don't think migrationists can win the debate on a fair footing and hence often point to the opinion of others to justify their positions. But The economist isn't here to help you.

I've had people come up to me and matter-of-factly mention the Aryan Invasion when I talked about the IVC. Clearly they had read about it somewhere. I despatched them pretty easily, showing them their position could not be sustained when challenged face-to-face.

Anonymous said...

@mzp1
You keep talking about how impossible it was for the Steppe Aryans to change so drastically the Culture and Language of mainland India, but then you willingly forgets to cite the biggest examples of this happenning outside of IE:
1. Turkic Invasions of Central Asia. Almost a perfect mimicry of the IE.
2. Arab Invasions of MENA. Some giat-herders from Arabia some day organized and conquered everything in their sight, changing the language and religion of everybody.
3. Afro-Asiatic expansions. Herders from North Africa bypassed the Sahara in the Green Period, changing the culture and language of Chad.
4. Bantu expansions. Bantu pastoralists expanded East and South in Africa, changing everything as well.
5. Iberian conquering of America. Unlike NA, they went there and changed the language and culture of natives.
6. Persia. The Steppe guys conquered the urbanized Elam and changed culture and language again.
7. Hungarian conquest. A handful of Magyars conquered a region and imposed culture and language.

You see, your arguments are bogus.

Karl_K said...

@namedguest

And importantly, we see the genetically steppe-related people expand at the same time across a huge mass of Eurasia, and bringing IE language to many locations.

The IE languages couldn't have originated in South Asia without several even more unlikely events occuring.

Bob Floy said...

@mzp

Karl and namedguest are right, you know.

elliv said...

@mzp1
If we skip all BS and look and the basics.
1. There are clear relations between languages and religion in Europe and India.
2. There are clear DNA evidence of East European migration to deep south central Asia 1500 BC.
3. Those migrants have Y-HGs that are popular in north India.
4. There are no evidence of Out-of-India movement.

If you really are into science you should rethink your position and join the rest in trying to figure out the details. Else your position will be just faith-based and end up like another creationist ideology.

Matt said...

I don't think the Bantu Expansion is classically thought of as an expansion of pastoralists. More of an Austronesian style agriculturalist expansion.

The expansion of Latin is another example of a large language expansion from a proportionately small originating group.

It'll be interesting to see how much the expansion of Arabic was a demic expansion. It would not be too surprising if it was quite a bit so.

old europe said...



Elliv J

Agree on what you said....apart from opposing science and creationism....science has nothing to say about the ultimate origins of things . Science can explain the essence and how things transform and change but cannot explain why things exist. creationism is in a religious book as the Bible but from a strict rational and philosophical point of view has the same dignity as other explanation and by the way is the more rational response to the question of why things and above all human beings exist.

mzp1 said...

@nameguest

Not really...

1. Turks I have already mentioned in my comment. They were mostly islamiced and Persianized, their language only remains in Turkey.

2. Only had an impact in the ME. Islam further spread y Persians. Arabic isn't spoken outside the ME.

3-4 Way smaller scales here.

5 Iberians were far more advanced than native south Americans.

6 Persians were not 'Steppe guys' then. Elam not comparable to IVC + Central Asia wrt size and geography.

7 Don't know about this so can't comment.

Sanuj said...

"No evidence of Out of India movement"

http://eurogenes.blogspot.com/2015/10/mitochondrial-dna-from-maykop-wolfgang.html
And
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0073682
And
Even this Narsimhan et al paper is talking about Indian migrants in BMAC and Iran. So much for the "No evidence of Out of India movement".

Combined with the recent genetic findings and the opinion of the lead archaeologist of a cultural continuity and no evidence of any invasion type scenario, and looking at it in light of an early Iron age in Indian subcontinent we can safely say there was no turn in population in the first half of 2nd Millennium, nor a change in language.

Now, as per scientific approach, we should look for more credible theories for spread of IE languages beginning with looking in a greater deal on what was happening South of Caucasus and Hittites and Mitanni Indo-Aryans.

mzp1 said...

@Elli J,

2. What evidence? We dont have good enough samples from South Asia. For all we know steppe dna + r1a could be post 1200 bc in South Asia which is too late. This study only weekens your position. Steppe guys need to invade and do a lot of work in a very short period of time. Both in South Asia and Central Asia.

4. No? What about the IVC artefacts found all over central Asia. Indian Mtdna in ancient south caucuses. Indians in Bmac etc.

R1A could of spread out of India for all we know.

These are some weak arguments being proposed here.

vAsiSTha said...

But no mention of the collapse of the 'macho alpha males in horse drawn chariots sacking ivc cities' theory lol.

mzp1 said...

Maybe they were flying chariots like in IE myths, that would be how they 'bypassed' the BMAC before invading India.

Davidski said...

@Sanuj

He also said in the earlier interview that the sample matches NI Brahmin - which i would construe as showing Steppe-like ancestry.

The Indus Valley samples will be practically the same as the Indus Periphery samples.

No Steppe-like ancestry.

Anonymous said...

@mzp1 you purposely ignore the fact that Indo-Aryans were cremated before the Buddhist era. You purposely ignore that these burials were not Indo-Aryans. You purposely ignore the appearance of the steppe impurity of about 1200 BC, which was not in the previous era.
It is your authority de facto prohibited the publication of data on Rakhigarhi (IVC).
Because these burials around 1200-800BC were not Indo-Aryans as were the aborigine, it is not surprising that there is no R1a. You wrote yourself that you did not understand anything in the study, so your empty words mean nothing, just an empty polemic.

Bob Floy said...

@mr. kulkarni
Just keep telling yourself that anyone who accepts AIT in any form just has macho conquest fantasies, that's real good.

@mzp
Supernord is right, you know.

Anthro Survey said...

@namedguest

*Ultimately*, the relevant alleles to explain modern-day variation in Europe may arisen in some WHG-related groups in the Balkans, Anatolia or even Levant, that's true. However, it is through a mechanism involving basal-rich ANF and EEFs that these alleles were ampliefied(further amplified during the Bronze Age expansions). This occured through a series of founder effects, cultimnating in some rather blond NE European EEFs, no doubt.

Mesolithic European WHGs were quite brunette. Not one group--including Narva, btw---attains significant blondism rates. Ukranian WHGs, like Narva, also come out pretty brunette.

If you look on Genetiker's Pigmentation spreads, we already begin to see significant blondism in Hungarian and Greek EEFs, otoh. Starcevo were no strangers to it, either. Virtually no WHGs even come close. So, it's easy to envision a small founding population of Danubian EEFs multiply into one with even higher rates of blondism than their predecessors as they colonized NE Europe.

Might they have encountered blond Narva-like EEFs? Absolutely, but that would have only added to their existing kit.

Perhaps GAC was not important in their genesis, but CWC was still a hybrid of some other NE EEF groups.

mzp1 said...

Supernord is rambling nonesense about a handful of samples from Swat that tell us diddly squat about the overall genetic makeup of a large, densely populated, and likely genetically structured region.

vAsiSTha said...

@bob floy
The author of the blog surely harbored those fantasies. maybe still does. ask him.

Anthro Survey said...

@Matt

Re/Arabic---Agree, but methinks the degree to which the spread of Arabic was accompanied by a demic shift is vastly overstated by Arabists. The same holds for Latin and especially New Persian. Plus, demic shift mediated by bedouin migrants in some regions post-dated Arabization so that's a confound to look out for.

Btw, been doing some modeling lately and it seems my earlier runs overestimated the amount of Anatolia_BA-like ancestry in France and various Iberian regions. That's because they were based on the erroneous assumption of an Iberia_Chl_average-like base across the board. When, for example, I give Catalans wiggle room by including Iberia_Chl and some WHG-poor EEFs like Remedello or Barcin itself, they don't get more than 5% Anatolia_BA. French lack it altogether.

This near eastern signal might translate into roughly ~15% S. Italian ancestry *assuming* it was the exclusive source. That's unlikely, however, given the Roman-age flow of Anatolian "Greeks" to Mediterranean Europe.

Anonymous said...

mzp1 rambling nonsense you carry. You are a layman who knows nothing and nothing to understand. This see all, fact. You know nothing about the Swat valley, absolutely, as well as any question.
Unfortunately for you, is the fact that it was not in IVC nor any steppe component nor R1a. After this article have nothing to hide and article about IVC is likely to be published.

Karl_K said...

@Mr. Kulkarni

"The author of the blog surely harbored those fantasies. maybe still does. ask him."

I personally don't care much about the macho conquest idea. However, there is no real evidence against this idea. The steppe nomads very likely did have a society where being macho was a big deal. Many signs point to this. Probably, they just rarely had a reason to actually fight anyone, so they didn't.

In the case of South Asia, they probably just moved into a new successful niche, during a hard time for the local people, and local people wanted to be a part of the new successful economic situation. Or something like that.

Open Genomes said...

R1b-P312 and R1b-U106 on the Steppe? Go figure it out!

DA111 R1b-P312 mtDNA: H6a1a Gedmatch: Z373539

1240k .zip:
http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA111/genome_DA111-1240k.zip
SNPs read: 595279 percent covered: 49.67

Were there Celts on the steppe, or this Hungary?

DA119 R1b-U106 mtDNA:U4b3 Gedmatch: Z873665

1240k .zip:
http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA119/genome_DA119-1240k.zip

SNPs read: 541006 percent covered: 45.14

A steppe Germanic might be a Goth or a Varangian.

Sanuj said...

"Or something like that."

Sums it all up pretty well.

Karl_K said...

I'd say so. The fact is that it happened. How or why? We do not really know.

Onur Dincer said...

@Open Genomes

Those two genomes (DA111 and DA119) have autosomal profiles just as would be expected based on their Y-DNA haplogroups. The former is indeed likely Celtic and the latter is likely Germanic, both of them are probably from Hungary, Romania or Ukraine.

Could you also check out DA203, DA204, DA222, DA230 and DA124?

mzp1 said...

"The fact is that it happened"

All the comments I wrote above were to make the point that you cannot justifiably make these statements as things stand.

I just find it sad how some people are so desperate to peddle this invasion theory when they have no arguments to back it up and, more importantly, it has nothing to do with them.

Are your ancestors North Indian? Does the dating of the Vedas matter to you in any way?

I don't think so.

Would it make you feel good if India was as externally influenced as Europe was? From the day Europeans went to India they've been feeling insecure about their origins.

Now you want me to take your word for it.

Come back when you have a real argument.

Onur Dincer said...

@mzp1

Would it make you feel good if India was as externally influenced as Europe was? From the day Europeans went to India they've been feeling insecure about their origins.

Nothing could be further from the truth. Phenotypic variation and diversity in South Asia are too high and they vary in a way approaching Europeans or broadly West Eurasians in one extreme, approaching Andamanese or similar-looking peoples in another extreme, and approaching East Eurasians in yet another extreme. It is obvious even just based on the looks that South Asia received migrations from different parts of the world in different times and under different circumstances. You do not see so much native phenotypic variation and diversity in most parts of Europe.

Onur Dincer said...

Also, those phenotypic extremes in South Asia are highly correlated with geography and caste.

Karl_K said...

@mzp

I really don't see what personal feelings, or where someone is living, or their own personal genetics has to do with this discussion. Try to look at things more objectively.

mzp1 said...

I don't know how this is even relevant but the phenotypes don't have much to do with steppe ancestry. It is a largely native development based on geography.

20% steppe ancestry won't make a Gujarati look like a pashtun or kalash. I know enough mixed Indian/Brits to be able to see that.

Onur Dincer said...

@mzp1

I don't know how this is even relevant but the phenotypes don't have much to do with steppe ancestry. It is a largely native development based on geography.

Phenotypes always correlate with genetics, and South Asia is no exception. The phenotypic situation in South Asia closely parallels the genetic situation there.

20% steppe ancestry won't make a Gujarati look like a pashtun or kalash. I know enough mixed Indian/Brits to be able to see that.

Gujaratis have significatly lower steppe ancestry and overall West Eurasian ancestry than Pashtuns and the Kalash, so they certainly do not look the same as them. As for the looks of recent South Asian-European mixes, they too are highly correlated with ancestry proportions/genetics.

mzp1 said...

@karl_K

"The fact is that it happened"

Is a subjective opinion expressed as objective fact. Your the one confusing personal bias with objectivity. Don't attempt to turn that around on me.

Rob said...

@ Blogger Open Genomes said...
R1b-P312 and R1b-U106 on the Steppe? Go figure it out!

Nice work. I doubt the first is 'Celtic', There were no "Celts" in the Hun period in west Steppe or Hungary. But certainly makes sense Chernyakov culture groups to have had R1b and some became Huns/ recruited as such. Several finds og "Euro-Mongoloid' individuals around.

Anthro Survey said...

Yes, it's true that Gujarati resemble neither Pashtun nor Kalash, but the *average* Gujarati doesn't have nearly 20% Steppe_MLBA. Brahmins might,yes, but they're what, maybe 7% of the population? Furthermore there are additional confounds because Indians have both AASI(20-40%) and diverged subcontinent-specific Iran_N ancestry. Pashtuns can be modeled as a mix of Iran_N and Iran_Chl-like, otoh, in their non-steppe_MLBA portion. So, even if we assume their Iran_N layer to have been phenotypically similar to that which colonized India, Iran_Chl-like layer surely had its own phenotypic implications.

Anthro Survey said...

We should also remember that phenotypic variation can take on a life of its own, given enough time, making ancestral contributions harder to guess. So, even barring what I've stated above, the relative dearth of people resembling Pamiris among India's upper caste needn't imply they don't have steppe_MLBA-related ancestry.

Onur Dincer said...

@Rob

I doubt the first is 'Celtic', There were no "Celts" in the Hun period in west Steppe or Hungary. But certainly makes sense Chernyakov culture groups to have had R1b and some became Huns/ recruited as such.

We do not know from which period they are. They may be from an earlier period than the period Huns showed up in Europe.

Several finds og "Euro-Mongoloid' individuals around.

Yes, but those two R1b guys are completely European genetically, and even western European specifically, they have very Celto-Germanic genetics.

Onur Dincer said...

@Anthro Survey

We should also remember that phenotypic variation can take on a life of its own, given enough time, making ancestral contributions harder to guess. So, even barring what I've stated above, the relative dearth of people resembling Pamiris among India's upper caste needn't imply they don't have steppe_MLBA-related ancestry.

Yes, but let's not forget that the steppe ancestry found among upper caste Indians never reach the levels found among Pamiris.

Jijnasu said...

@anthrosurvey
Besides the influences of the unique Iran Neolithic and AASI ancestry, selection likely influenced phenotypic variation in India as well. The derived allele for SLC24A5 occurs in several castes at frequencies higher than their estimated west european ancestry while in a few groups such as some non brahmin tamil castes it occurs at lower than expected frequencies.

mzp1 said...

I meant Gujaratis + 20 % steppe will not look like a pashtun.

Just look at photos of 50/50 indian/white children. They are far from pashtun/kalash phenotypes.

Google Indian white mixed.

Modern euros are lighter than steppe and these are 50% European.

Samuel Andrews said...

@Anthro Survey,

mtDNA wise, Globular Amphora looks like a farmer ancestor of eastern Europeans and probable Balto Slavs in bronze age Latvia. However, Hungary Neolithic looks like a genetic dead end. Based only on mtDNA, I have serious doubts that the Danbien Neolithic route contributed much ancestry to anybody.

Jijnasu said...

that's because pashtuns aren't a mixture of modern Indians and modern europeans?

Onur Dincer said...

@mzp1

Steppe ancestry is not everything. The biggest genetic difference between Brits and Pashtuns/Kalash is found in their non-steppe ancestries, so that is where most of the sources of their phenotypic differences should be sought.

Rob said...

@ Dave, or anyone
Based on this when do you envisage the IA migration/ colonisation/ invasion occurred & how ?

Anthro Survey said...

@Jinjasu

Yes, I agree re/selection. We have to remember that this steppe_MLBA admixture even took place some 2500-3000 years ago. Plenty of room for selection and drift to occur in the endogamous Indian castes.

@Onur Dincer

That's true, but Pashtun steppe_MLBA levels are lower and more comparable to upper caste indians depending on the subpopulation. This also holds for the Kalash.Nonetheless, both of these still have considerable overlap with Pamiris, unlike upper caste North Indians.

My point is that even if what I wrote above re/confounds wasn't the case, we should still be careful about making precise judgments about relative amounts of steppe_MLBA. Just because a sub-population appears to be more steppic to our eye than another, doesn't mean it actually is and vice versa. Stong phenotype-ancestry correlations certainly do exist across geographical space, of course, and can be useful hints of ancestral structure if used with caution.

Anthro Survey said...

@Mzp and Onur Dincer

I wouldn't expect those half Indian/half British children to resemble Pashtuns anymore than I would expect Mexicans or Colombians to strongly resemble Hazara(beyond some basic Mongoloid affinities).

Strong phenotypic correlations will exist b/ween autosomal components so long as the "path" by which these components founds its way into a population's genome is the same or at least similar. So, it's not just a matter of non-shared components, but this plays a role too, obviously.

Onur Dincer said...

@Anthro Survey

Totally agree with your points.

Anthro Survey said...

@Sam

Yes, but where did GAC EEFs come from, then?

Rob said...

@ named guessed

“1. Turkic Invasions of Central Asia. Almost a perfect mimicry of the IE.”

How so ?
The steppe hypothesis envisions a Big Bang scenario from a sacred well spring
The Turkic expansions were layered- first Hunnno- Bulgar, then Turkic proper with GokTurks than Others still, with Kipchak riding on its use by western Mongols
And Turkic isn’t spoken in Europe, apart from minorities left over from a modern empire expansion


7) “Hungarian conquest..””
Please elaborate . What about it?

Davidski said...

@All

DA111 & DA119 PCA plot

https://drive.google.com/file/d/13nXi-uJhsmxF-l6ydoDq376b7FOOT3Rc/view?usp=sharing

DA111 & DA119 PCA coordinates

https://drive.google.com/file/d/1TR8rn58q1tE09VMIw6vSGs7FUhif-fRg/view?usp=sharing

Jijnasu said...

@anthrosurvey
Very true, also some supposesly 'steppic traits' in south asia may have more to do with iran neolithic ancestry than the steppe in some groups. Upper caste non-brahmin dravidians particularly from the west coast can be fairly light-skinned and occassionally even light-eyed. yet the dravidian is often stereotyped as dark. Phenotype is a very rough indicator of ancestry.

mzp1 said...

The point I was making is that if 50% British doesn't massively shift Indo Aryans (let's say to match the fairness of pashtuns) then 10%-20% is not taking you from 'average' Punjabi to kalash.

The reason for the phenotype diversity is due to selection and time-depth, not external admixture.

Anyway I am happy with your last comment.

Anthro Survey said...

@Jijnasu

It's an interesting trend I've noticed that on a few occasions. The actress Aishwarya Rai fits this bill and she hails from the area(I'm assuming she's local?).

Will have to check, but, as I recall, Gujarati mainly just get Iran_N-related ancestry, without Iran_Chl. I'm assuming this is also true for people further south in Kerala, etc., who we don't have in G25(actually, never seen ADMIXTURE output for them either). If that's the case, then, yeah, it's deff a local development stemming from an Iran_N stream of ancestry. Sindhi do get Iran_Chl, though.

Bob Floy said...

@mr. kulkarni
"The author of the blog surely harbored those fantasies. maybe still does. ask him."

You clearly have some fantasies of your own.

Davidski said...

@Rob

Based on this when do you envisage the IA migration/colonisation/invasion occurred & how?

The finding of practically unadmixed Eastern European steppe individuals dating to the Bronze Age in the Ferghana Valley, and widespread admixture from such people in present-day South Asians, especially Indo-European speakers and Indian upper castes, doesn't leave too many options.

Obviously, there was a migration of Indo-European speakers from the steppe through the IAMC during the Bronze Age into South Central Asia, and then dispersals throughout the region and eventually deeper into South Asia.

On the doorstep of India

I'll leave it to others to work out all of the details over the next decade or so.

Onur Dincer said...

@mzp1

The reason for the phenotype diversity is due to selection and time-depth, not external admixture.

It is both, not just internal factors, but a combination of internal and external factors.

Jijnasu said...

She would be an example yes. Many of these groups are often understudied with most research focussing on Brahmins, dalits and tribals

mzp1 said...

@Davidski

"@ Dave, or anyone
Based on this when do you envisage the IA migration/ colonisation/ invasion occurred & how ?"

I'm not gonna go into yet another version of the migration theory here and now. I will just say that Rob asked you a fair question and instead of acknowledging it you diverted into something else.

A simple "I dont know" would have sufficed.

Anshuman said...

Ask them what actually constitutes steppe ancestry..Which they keep talking about.which haplogroup would represent it..which specific autosomal mutation represents it.

Onur Dincer said...

@Davidski

DA111 & DA119 PCA plot

https://drive.google.com/file/d/13nXi-uJhsmxF-l6ydoDq376b7FOOT3Rc/view?usp=sharing


Thanks. It visualizes their Celto-Germanic genetic affinities pretty well.

mzp1 said...

@onur

It is mostly internal as the the phenotypes are more closely explained by geography/caste than steppe admixture.

It would look like phenotype diversification happened on top of steppe admixture according to geography. That would imply an earlier structuring and possible earlier steppe admixture.

VerdianoBR said...

@mzp
What are you talking about: "Turks I have already mentioned in my comment. They were mostly islamiced and Persianized, their language only remains in Turkey."? In Central Asia? Really? Until the Slavic expansion to Siberia with the modern Russian Empire, virtually all the Eurasian steppes had become Turkic-speaking, and even today in Central Asia only most (most - not all) of Aghanistan and Tajikistan remains Iranian-speaking. Uzbekistan, Turkmenistan, Kazakhstan, Kyrghyzstan, you name it... overwhelmingly speaking Turkic languages and still with a heavy pre-Turkic local ancestry. Ditto for Azerbaijan and even significant parts of Iran itself. You're really misinformed if you really think - not just try to hide it because it clearly doesn't fit your petty wishful thinking - that Turks were mostly "islamiced and persianized" - actually Turks adopted Islam willingly in almost all places to dominate the conquered peoples better, imparted the locals much of their steppe culture and were extremely successful in imposing their language onto more populous and advanced peoples. Get over it.

And they weren't the only examples of such "backward tribal areas > more advanced urban civilizations" movement and language imposition in documented, historic times. Magyars, Aztecs, Arabs, even in some ways Germans and Slavs managed exactly that in cultures that were apparently way beyond their capabilities. Don't try to fool yourself making things up.

Open Genomes said...

@David @Onur

DA222 is a very high coverage Central Asian genome rumored to be either Karluk or Karakhanid.

DA222 Y J-Z7760 mtDNA A25 Gedmatch: Z342597

1240k 23andMe format:
http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA222/genome_DA222-1240k.zip
SNPs read: 1170454 percent covered: 97.66%

K13 Oracle4:

Using 1 population approximation:
1 Afghan_Turkmen @ 11.099735
2 Uzbeki @ 12.520078
3 Hazara @ 13.258211
4 Aghan_Hazara @ 15.147772
5 Kazakh @ 18.063383
6 Uygur @ 18.389980
7 Kirgiz @ 19.512558
8 Shors @ 20.301899
9 Hakas @ 21.949875
10 Nogay @ 22.251713
11 Altaian @ 26.367954
12 Afghan_Tadjik @ 27.097479
13 Tadjik @ 29.241249
14 Turkmen @ 31.082531
15 Mongolian @ 32.371464
16 Tuvinian @ 35.570374
17 Tatar @ 36.216877
18 Ket @ 37.541454
19 Kabardin @ 38.023266
20 Afghan_Pashtun @ 38.358185

Using 2 populations approximation:
1 50% Afghan_Tadjik +50% Hakas @ 6.861816

Using 3 populations approximation:
1 50% Altaian +25% Kalash +25% Nogay @ 5.856803

("Hakas" = Khakas, aka "Tadar")

Using 4 populations approximation:
1 Hakas + Kabardin + Kalash + Oroqen @ 4.348867
2 Chechen + Hakas + Kalash + Oroqen @ 4.507372
3 Hakas + Kalash + North_Ossetian + Oroqen @ 4.561662
4 Burusho + Chechen + Hakas + Oroqen @ 4.755246
5 Balochi + Mongolian + Nogay + Selkup @ 4.793129
6 Balkar + Hakas + Kalash + Oroqen @ 4.814628
7 Hakas + Kalash + Lezgin + Oroqen @ 4.876982
8 Balochi + Hakas + Hakas + Kazakh @ 4.919081
9 Brahui + Hakas + Hakas + Kazakh @ 4.946056
10 Balochi + Hakas + Nogay + Tuvinian @ 4.946147
11 Altaian + Chechen + Kalash + Oroqen @ 4.948618
12 Kabardin + Kalash + Oroqen + Shors @ 4.954937
13 Buryat + Hakas + Kalash + Nogay @ 4.955349
14 Burusho + Hakas + Lezgin + Oroqen @ 4.956622
15 Brahui + Mongolian + Nogay + Selkup @ 4.962278
16 Hakas + Kalash + Oroqen + Tabassaran @ 4.972881
17 Brahui + Hakas + Kazakh + Shors @ 4.981367
18 Burusho + Hakas + Oroqen + Tabassaran @ 5.006890
19 Kabardin + Kalash + Tuvinian + Tuvinian @ 5.014536
20 Balochi + Hakas + Kazakh + Shors @ 5.022939

Anshuman said...

Do these guys even get that their theories which had oscilaltion between R1a,R1b as steppe marker not so long ago was giving low yields..That too does not help as only one sample from Swat is R1a and that too in 500-300 bc time when Greeks had entered there

mzp1 said...

"And they weren't the only examples of such "backward tribal areas > more advanced urban civilizations"

Because they aren't at all.

"Uzbekistan, Turkmenistan, Kazakhstan, Kyrghyzstan, you name it"

Kazakhstan? Kyrgstan? That's the best you can do for settled advanced urban civilizations. Those people all look Turkic to me.

So tell us, where is this great Turkic religion and literature this great civilization produced once it controlled such large resources?

VerdianoBR said...

@Mzp Are you really trying to make us believe that Uzbekistan, Turkmenistan, Azerbaijan and Turkey had no urban civilizations by the Middle Ages? And that after Turkification all those countries had no cultural (including literary) development at all throughout the centuries? And are you really trying to discuss about population genetics by judging that people "look Turkic to me" even if you don't even have any quaint idea of how the early Proto-Turks really looked like? (and no, sorry, but even modern Turkic Central Asians look very different from Altaian Turks, those closest to the most probable Urheimat of Proto-Turkic) Oh, dear, then you're not just deluding yourself. You're just ignorant and apparently proud about that, indeed. Learn more, that's all you need.

Karl_K said...

@Anshuman

"Ask them what actually constitutes steppe ancestry..Which they keep talking about.which haplogroup would represent it..which specific autosomal mutation represents it."

It is not a specific mutation or haplogroup. It is allele frequency at many thousands of SNP positions on all of the autosomal chromosomes.

You can easily get this information out of the data by comparison of genomes from different source if you really enjoy looking at very long lists.

Onur Dincer said...

@mzp1

It is mostly internal as the the phenotypes are more closely explained by geography/caste than steppe admixture.

It would look like phenotype diversification happened on top of steppe admixture according to geography. That would imply an earlier structuring and possible earlier steppe admixture.


You are again restricting the subject to the steppe migation. The IA steppe migration was not the only migration to South Asia, there were many migrations from many different sources and times, thus many external factors contributing to the South Asian genetic and phenotypic diversity and variation.

Rob said...

@ Onur

“We do not know from which period they are. They may be from an earlier period than the period Huns showed up in Europe.”

But surely not halstatt or la Tene. Otherwise it’s not within the scope of “Turkic migrations” topic

Anthro Survey said...

The Hunnic "empire" had vassals deep in Central Europe at one time, so West Euro affinity together w/R1b-l51 in some Dark Age steppe samples is interesting but not anomalous.

Rob said...

@ AnthroS
I still prefer to look to the most obvious answer- the chernyakov culture, instead of Central Europe and ancient celts

Anthro Survey said...

@Verdiano

We can't treat Uzbekistan as a singular unit because it isn't. There's much to be said about the contrast between the urban southeastern portion and the less densely populated Karalpakstan. Prior to the Soviet Era, this region was either majority Tajik or they constituted a significant minority. Uzbekification and/or ethnic cleansing altered the demographic balance there. Besides, the cultural contrast between sedentary Turkics living there and between nomadic, yurt-dwelling ones in the more desolate regions was pretty stark. Even Turkic-speaking regions further north around Tashkent and south Kazakstan have more of a "Sart" vibe to them than Turkic.

It's true that Turks weren't forcibly converted to Islam, but were rather evangelized by Samanid missionaries. When they entered what were to become Persianate lands and came to constitute the de-facto military caste of these states, they didn't exactly force their "steppe culture" on anyone, though. In fact, cultural exchange was undeniably pretty lopsided with Turks becoming quite Persianized with time. This is especially true for Turkic populations residing in urban areas. There was quite a bit of stigma associated with nomadic Turkic populations in the Seljuk and Ottoman states alike, in fact.

Going back to those Uzbeks---they're pretty darn Persianized in terms of both customs and language. So much so, that in Afghanistan, where the Tajik-Uzbek bitterness is less pronounced, these groups are very much allied.

Anthro Survey said...

@Rob

Oh, I doubt these samples belong to ancient Celts. We have to remember, though, that *Celt-like* groups probably existed around Czechia and Hungary(Late Rome's fringes, basically) circa when the Hunnic confederation made its mark.

Onur Dincer said...

@Open Genomes

DA222 is a very high coverage Central Asian genome rumored to be either Karluk or Karakhanid.

DA222 Y J-Z7760 mtDNA A25 Gedmatch: Z342597

1240k 23andMe format:
http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA222/genome_DA222-1240k.zip
SNPs read: 1170454 percent covered: 97.66%


Thanks! That Karluk sample has a 50-50 proportion of East Eurasian-West Eurasian ancestries, the exact proportion I was expecting to see among Karluks and Oghuz/Turcomans, just like in modern Uyghurs. The only mistake in my prediction was that I expected them to differ from modern Uyghurs more in their West Eurasian sub-components than in their East Eurasian sub-components, the opposite is the case it seems.

Here are the Harappa World component proportions of the Karluk sample:

S-Indian 2.68
Baloch 18.90
Caucasian 13.52
NE-Euro 12.26
SE-Asian -
Siberian 24.29
NE-Asian 21.54
Papuan -
American 1.65
Beringian 2.98
Mediterranean 1.54
SW-Asian -
San -
E-African -
Pygmy -
W-African 0.64

And here are average Harappa World component proportions of the modern Uyghur reference samples:

S-Indian 5.18
Baloch 16.48
Caucasian 13.02
NE-Euro 12.03
SE-Asian 2.07
Siberian 14.63
NE-Asian 29.73
Papuan 0.48
American 1.48
Beringian 1.68
Mediterranean 1.88
SW-Asian 1.30
San 0.03
E-African 0.02
Pygmy -
W-African -

Davidski said...

@Open Genomes

Can you pull down the highest quality Egyptian mummy and Canaanite genomes and process them in the same way? If they work I can add them to the G25 datasheets.

https://www.ebi.ac.uk/ena/data/view/PRJEB15464

https://www.ebi.ac.uk/ena/data/view/PRJEB21330

Anthro Survey said...

@Open Genomes

Can you also do this for the Guanche?

Open Genomes said...

@David

DA228 a first glance appears to be our first Chinese ancient genome.

DA228 Y: O-F714 mtDNA: A15c Gedmatch: Z845678

1240k in 23andMe format:

http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA228/genome_DA228-1240k.zip
SNPs read: 1003635 percent covered: 83.74%

However, at K15 Oracle4, he looks very different than a typical Han Chinese.
He looks part Turic, part Uralic, and part Tungusic, with European affinity.
Who could he be? Xiongnu?

Using 1 population approximation:
1 Tatar @ 20.783182
2 Uygur @ 21.934755
3 Afghan_Turkmen @ 22.491089
4 Uzbeki @ 23.141068
5 Kazakh @ 23.616861
6 Aghan_Hazara @ 26.093121
7 Nogay @ 26.403139
8 Hazara @ 26.677250
9 Chuvash @ 27.743759
10 Kirgiz @ 27.754847
11 Shors @ 28.772167
12 Mari @ 30.045675
13 Moldavian @ 31.095219
14 Hakas @ 31.216671
15 Serbian @ 32.070538
16 Romanian @ 32.494301
17 Hungarian @ 32.840744
18 Croatian @ 33.251427
19 Bulgarian @ 33.463650
20 Ukrainian_Lviv @ 34.734238

Using 2 populations approximation:
1 50% Hungarian +50% Kirgiz @ 7.607682

Using 3 populations approximation:
1 50% Kirgiz +25% La_Brana-1 +25% Romanian @ 5.538270

Using 4 populations approximation:
1 Chuvash + Nogay + Norwegian + Xibo @ 2.149634
2 Nogay + North_Swedish + Tatar + Xibo @ 2.223074
3 Hezhen + Nogay + North_Swedish + Tatar @ 2.234707
4 Chuvash + Nogay + North_Dutch + Xibo @ 2.274684
5 Chuvash + Hezhen + Nogay + Norwegian @ 2.296829
6 Chuvash + Nogay + Swedish + Xibo @ 2.335683
7 Chuvash + Danish + Nogay + Xibo @ 2.382205
8 Chuvash + Nogay + Orcadian + Xibo @ 2.389293
9 Chuvash + Hezhen + Nogay + North_Dutch @ 2.395756
10 Japanese + Mari + Nogay + Norwegian @ 2.397129
11 Japanese + Mari + Nogay + North_Dutch @ 2.439818
12 Chuvash + Danish + Hezhen + Nogay @ 2.505448
13 Chuvash + Hezhen + Nogay + Swedish @ 2.541106
14 Chuvash + Hezhen + Nogay + Orcadian @ 2.547318
15 Chuvash + Nogay + North_German + Xibo @ 2.571594
16 Chuvash + Irish + Nogay + Xibo @ 2.637760
17 Chuvash + Hezhen + Nogay + North_German @ 2.649568
18 Japanese + Mari + Nogay + Orcadian @ 2.666461
19 Chuvash + Nogay + Southeast_English + Xibo @ 2.673505
20 Danish + Japanese + Mari + Nogay @ 2.682477

Onur Dincer said...

@Rob

Read the abstract of the Eurasian steppe paper, it says: "In order to understand the genetic history of the Eurasian steppe populations, we have sequenced 137 ancient genomes (~1X average coverage) spanning a 4000 years time series." So technically it is possible that they sampled some La Tene or Hallstatt era samples. Also they do not start with the Hunnic era, which is clear from this statement in the abstract: "We find evidence of a highly dynamic population history; the Iranian-speaking Scythians that dominated the Eurasian steppe throughout the Iron Age (~1 millennium BCE to common era) emerged following admixture between Late Bronze Age herders of western Eurasian descent and East Asian hunter-gatherers."

The full abstract:

"The Eurasian steppe, stretching about 8000 kilometres from Hungary and Romania in the west to Mongolia and western China in the east, is culturally among the most dynamic areas in the world. In the past four millennia, it has been variously dominated by Iranian-, Turkic- and Mongolic-speaking groups, and its temperate grasslands have been a crossroad for extensive movements of peoples, goods, and ideas between Europe, Siberia, South and East Asia. In order to understand the genetic history of the Eurasian steppe populations, we have sequenced 137 ancient genomes (~1X average coverage) spanning a 4000 years time series. We also genotyped 502 individuals from 16 contemporary self-reported ethnicities. We find evidence of a highly dynamic population history; the Iranian-speaking Scythians that dominated the Eurasian steppe throughout the Iron Age (~1 millennium BCE to common era) emerged following admixture between Late Bronze Age herders of western Eurasian descent and East Asian hunter-gatherers. The steppe nomads later further admixed with Turkic-speaking groups of East Asian ancestry that spread westward across the steppe in multiple waves: firstly, the Xiongnu confederations that emerged in Mongolia around the 3nd/2nd century BC; secondly, the Huns (4-5th century CE), infected with plague basal to the Justinian Y. pestis strain that destabilized the eastern Roman Empire in the 6th century CE; and thirdly during various short term dynasties, including the Mongol Empire of Genghis Khan and his descendants. These recent historical events transformed the Eurasian steppe populations from being Indo-European speakers of largely western Eurasian ancestry to the present-day Turkic-speaking groups, primarily of East Asian ancestry".

Onur Dincer said...

@Open Genomes

DA228 a first glance appears to be our first Chinese ancient genome.

DA228 Y: O-F714 mtDNA: A15c Gedmatch: Z845678

1240k in 23andMe format:

http://www.open-genomes.org/genomes/Eurasian%20Steppe/DA228/genome_DA228-1240k.zip
SNPs read: 1003635 percent covered: 83.74%

However, at K15 Oracle4, he looks very different than a typical Han Chinese.
He looks part Turic, part Uralic, and part Tungusic, with European affinity.
Who could he be? Xiongnu?


He is too North European-admixed to be Central Asian Turkic. He is probably Uralic or Uralic-admixed or from some ancient population whose genetics we do not know.

Rob said...

@ Onur

Fair enough. Eagerly awaiting the full paper.

Onur Dincer said...

@Rob

Fair enough. Eagerly awaiting the full paper.

You won't have to wait long. It will be published by mid May.

postneo said...

Haji Firuz alleged chalc r1b is quite shifted from the Bronze Age r1b. They cannot be from the same time frame

postneo said...

There is “west Eurasian” ancestry in India but none of it is from a single pulse. For e.g such components in andhra communities are quite different from from Karnataka Gujarat etc. In many cases there are separate disparate west Eurasian like signals within the same endogamous group probably from different unrelated ages and events. The single mlba steppe pulse is not going to hold.

Davidski said...

@postneo

Haji Firuz alleged chalc r1b is quite shifted from the Bronze Age r1b. They cannot be from the same time frame

You mean Hajji_Firuz_I2327 and Hajji_Firuz_BA_I4243.

Yeah, they can be from the same period, and even from the same population. They just have very different ratios of basically the same components.

Obviously, Hajji_Firuz_BA_I4243 might be a female with a father from the steppe, while Hajji_Firuz_I2327 has more distant steppe ancestry.

The single mlba steppe pulse is not going to hold.

It works just fine for practically all Indo-Aryans. Many Dravidians can't be modeled with Steppe_MLBA, but that's not exactly a problem for the Kurgan hypothesis or AIT, now is it?

You're either not very good at this or very lazy.

Matt said...

@AnthroSurvey, interesting re: Catalans and Iberia_Chl. I think there might be some hints of a relatively WHG poor EEF source, more Remedello like, involved in post-Roman populations in Western Europe, and Italy could be a source for that (based on Remedello).

...

Using the G25 data for DA111 and DA119: https://imgur.com/a/LUM13JK
NJ trees and then plots.

DA119 looks close to the averages for Beaker Central Europe and present day Austrians, while DA111 closest to the average for French and Beaker Southern France.

(In the case of DA111 rather than SW European this may more plausibly represent SE European EEF-rich ancestry later disrupted by other population movements? Or maybe SW European links are plausible?).

On the margins DA111 may be slightly more related to EEF and possibly East Asians than the Beaker Central Europe average is .

Ric Hern said...

http://eurogenes.blogspot.co.za/2018/01/the-case-of-chalcolithic-fortresses-in.html?m=1

Can this evidence of two different occupational layers point to a slow Steppe migration/movement all along the Northern Caucasus lowlands Southeastwards to Azerbaijan and from there fanning out into Armenia and Northwestern Iran ?

Archaelog said...

@Jijnasu There were multiple incursions of North Indian communities into the South. First the Jains from 300 BCE, then the Brahmins in several waves. To find the source of genetic variation in communities along the west coast you mentioned, we probably have to look no further than the effect of these migrations.

Southern Karnataka was a Jain centre with large settlements lasting to 200-300 AD when they were finally overwhelmed by the revival of Brahmanical religion. Brahmin settlements in
the south existed from the Sangam Age but the migration of North Indian Brahmans into the southe steadily increased when the north came under attack by Huns and Turks.

There were also some trade related settlements from the Persian gulf region but these communities are culturally/religiously distinct from the majority population. But even taking all this into account, the amount of "aberrant" phenotypical variation in South India, as some might put it, is quite minuscule.

MaxT said...

"he looks very different than a typical Han Chinese.
He looks part Turic, part Uralic, and part Tungusic, with European affinity. Who could he be? Xiongnu?"


What time period is this sample from? I'm guessing hes Scythian.

Any idea if we will get Shang Dynasty aDNA from China? I'm certain earlier steppe Indo-Europeans were there before Scythians, and influenced Shang Dynasty in many ways, especially their materiel culture.

Christopher Beckwith on Shang Dynasty says "Indo-Europeans may have been responsible for the foundation of the Shang Dynasty" we also see appearance of chariots for the first time in China under Shang.

Jijnasu said...

@chetan
There is a quite a bit of phenotypic variation in South India as well, there's a NW-SE cline, and one along the caste hierarchy as well. Its pretty unlikely that this entered south India during the during the historical period. Non-Brahmin (though perhaps brahminical in terms of religion) upper-caste dravidian speakers are different from brahmins from the same region. Despite having simillar amounts of aasi the steppe ancestry is much lower while the amount of iran neolithic ancestry is higher. Brahmin migrations were multiple, small and rather late to have any major impact on south Indian society genetically.

Nirjhar007 said...

Dr. Rai infoms ,to not take the Open Magazine interview seriously, as the interviewer unfortunately didn't understand many crucial points of what was said to him ..

Davidski said...

@Nirjhar

I doubt that anyone is taking the said interview seriously. Although I gotta admit, it wasn't quite as entertaining as that schlock in the Jagran, after which I almost wet my pants from laughing so hard.

Indian smoke and mirrors

Nirjhar007 said...

That one was genuine.

Davidski said...

Well that's very reassuring.

Archaelog said...

@Jijnasu "Non-Brahmin (though perhaps brahminical in terms of religion) upper-caste dravidian speakers are different from brahmins from the same region."

Yes I agree that some of the variation may date from Indus times along a Iran_N - AASI cline. But the effect of the said Jain-Brahman migrations are not to be discounted either.

Nirjhar007 said...

You can stay assured or not assured, it is your choice. But I suspect you are getting nervous about those Anatolian genome bud...

Anonymous said...

@Nirjhar007

I think that the interviewer did not understand only one thing that the paper will be published in the near future. It is very likely that it was meant to be published in the distant uncertain future (until infinity).

Anonymous said...

@ Nirjhar007

What exactly did the interviewer not get? And how does that change the point of view presented in the article?

Anonymous said...

@supernord

From what the article states, and from what Nick Patterson said here, they have great trouble getting good samples. Maybe they should go back to the lab and try the newer techniques that have proven successful in the South-East Asian paper. I don't know if that was available to them.

Davidski said...

@All

DA222 and DA228 Global25 coordinates are here...

https://drive.google.com/file/d/19vzQtW8ursu4p6WQbONpSKV6thYhSmO_/view?usp=sharing

DA222 is also now in the Global25 datasheets, labeled as Karluk_Medieval:DA222.

https://drive.google.com/file/d/1FSzKKknFGcOgfyA76q9PD7B-n-MJs7L8

Anonymous said...

@epoch2013

The results are ready for them, as I understood completely. Here is only to publish they threaten already as year, but not publish. So it's their new "to be published soon" means nothing more than the past times.

«Oldest ‹Older   1 – 200 of 265   Newer› Newest»