A new dataset has appeared online from a yet to be published paper titled
Cosmopolitanism in the depths of Barbaricum evidenced by archaeogenomic data from the Late Iron Age Goth community of the Masłomęcz group.
Most of these Gothic samples are clearly of Scandinavian origin, and very similar to present-day Swedes. Overall, however, they create a somewhat heterogeneous cluster that also overlaps with present-day Poles thanks to the presence of a few Balto-Slavic-related and possibly Roman-related individuals.
The Principal Component Analysis (PCA) plots below were produced with the excellent
Vahaduo G25 Global Views tool using the data
here.
Their Y-haplogroups more or less reflect the PCA results:
PL046 R-YP6228
PL048 I-PH833
PL049 I-A11537
PL052 R-Y48961
PL059 I-PH833
PL062 I-S15301
PL065 I-Y294193
PL066 R-FGC2555
PL067 R-S7759
PL070 I-CTS10028
PL071 I-BY316
PL076 I-S9318
PL082 I-Z2041
PL085 J-Z38241
PL086 I-FT29339
See also...
Early Slavs from Tribal Period Poland
Wielbark Goths were overwhelmingly of Scandinavian origin
High-resolution stuff
642 comments:
«Oldest ‹Older 601 – 642 of 642Mbuti
Iran_GanjDareh_N
Kotias_M
Barcin_N
Natufian
Iberomarusian
Samara_H
Tyumen_N
Luxemburg_M
China Xiaohe
Shamanka_EN
Mongolia_North_N
Ong
@Shomu it means that buried inside a NW Pontic Kurgan ~3900bc was a migrant from somewhere around that region.
I ran F3 just now
(DinkhaTepe_BA_IA_2, Jordan_LBA,Steppe_C atacomb)
I am getting Z = - 7.40
Replaced Catacomb with Sintashta and Z drops to -1.20.
This means DinkhaTepe2 has a Jordan_LBA + Catacomb like base that later mixed with BMAC like people?
HajjiFiruz_BA -1.25
Tavshut is insignificant.
https://ibb.co/0ynKKT5w
If proto-celtic has no shared inherited maritime vocabulary how could celtic from the west possibly be true? They did not develop a common specific terminology for sailing and navigation. It came independently when central european Celts reached the insulars. Celtic from the west theory seems foolish and driven by bias. Dna does not support it either
@EthanR
Very interesting 🤔
DinkhaTepe2
P-value: 0.76 (Excellent fit)
Chi-squared: 6.62
Degrees of Freedom (dof): 10
BMAC (SappaliTepe_BA): 52.0% (±4.8%)
Levant (Jordan_LBA): 37.9% (±3.3%)
Steppe (Catacomb): 10.1% (±3.1%)
Final model with F3 backing.
At hasanlu though something weird is happening.
F3 analysis gives high negative z for Jordan_LBA-Catacomb and Jordan_LBA-BMAC but positive value for BMAC and catacomb.
How to read it?
Brother, what are you trying to prove ❓
I am getting high negative Z values F3 for Katelai_IA and SaiduSharif_H for Iran_ShahrISokhta_BA2 and Sintashta but high positive for Udegram_IA. Any clue what could be happening?
Burusho.DG | P: 0.561 | Chisq: 7.74 | dof: 9 | ShahrISokhta: 67.2% | Sintashta: 20.1% | Russia_Yenisei_LBA_Karasuk.SG: 12.6%
Yenesian and Burushaski links, anybody 🧐
Birlik_Tasmola 600-200bce also works.
@ rozbójnik
As with everything, the context & 'philosophy' of theories being formulated needs to be understood. At the time, west European archaeologists were very much into 'local evolution', de-emphasizing migrations, with the view that BB evolved from local Atlantic Megalithic and in turn Celtic deveoped from that. ADNA disproved that, although for the first few years the pendulum swung too far in the other direction due to poor interpretation; esp. when it came to northern & central Europe (...for some reason).
Anyhow, there have since been alternative proposal about 'Celtic from the Middle' (https://www.cambridge.org/core/journals/cambridge-archaeological-journal/article/an-alternative-to-celtic-from-the-east-and-celtic-from-the-west/4F186F087DD3BE66D535102484F8E8C3),
whilst recent aDNA papers going back to some form of Urnfield model (e.g Patterson, McColl).
@rozbójnik
That raises the question — who lived on the Atlantic shores of Western Europe simultaneously with the landlocked Proto-Celts? Perhaps other Indo-European tribes with similar genetics, their Indo-European languages/dialects now extinct.
It's starting to crystallize for me that the number of extinct Indo-European languages/dialects is perhaps much bigger than we think. You have the Atlantic shores of France, Belgium and Southern Netherlands, you have the Dutch Bell Beakers who are seemingly not Germanic based on their Y-DNA, you have the Battle Axe culture whose Y-DNA almost doesn't exist outside Scandinavia/Finland, you have pre-Celtic British Isles, pre-Celtic Indo-Europeans in the Iberian peninsula, and that's just those off the top of my head.
@shomu
Now that I have access to Qpadm, F2,F3,F4...I am trying to learn them and improve my interpretation and analysis of several populations which was previously limited to G25 and PCA.
Take for example Burusho. F4 shows affinity for Khovsgol and Northern sources but G25 always favor Tibetian and southern source. On qpadm inclusion and exclusion of certain pops can favor one model over the other even to the extent of going against what F3 and F4 might point to.
So trying to learn and make sense to perform better analysis which eventually would lead to better interpretation.
A couple of Bros pointed me to the recent Bulan-Koba / Turkic paper
https://genarchivist.net/showthread.php?tid=1498&page=2
How does the mass of R1a-Z93 and even J2a support the Slab Grave ancestry hypothesis proposed by (presumably) Turkish internet posters ?
We get they have a lot of northeast asian related ancestry, but where is this from specifically, and why is it almost exclusively Female mediated? (Obviously there might be some relevant Y-hg C2 clades)
@Rob
Ask these Turks why only R1b-L73 and the M478 subclade, we know is older in the Baltic and eastern Europe and probably (so I think) derived from Villabruna 14000 Years ago and relatives of his refugium?
@Rob
There's an interesting story behind that.
The Turkic expansion absorbed masses of Indo-Iranian and other non-Turkic speakers.
These people were basically forced to flee or join the Turkic expansion, because otherwise they were killed.
Once the Indo-Iranians switched to Turkic it became important for them to become as Turkic as possible, and that actually involved taking Turkic brides from the east with Asian features to cement their link to the Turkic homeland.
Yes, the socio-political dynamics determine which language prevails and takes hold in a population and not which Y haplo they carry or how much autosomal ancestry their gene pool has from X population.
Hence an I male in Anatolia might had been speaking a non-Indo european Hattian while someone with J in same region was speaking Hittite.
I made some change to the right pops for Burusho model. I introduced Taiwan_Hanben_IA.AG as a reference pops and model with TibetanPlateau_Zongri.AG has p: 0.4 vs Kazakhstan_Birlik_EIA_Tasmola.AG has p: 0.7.
Is Taiwan_Hanben_IA.AG properly anchoring the EA ancestry in Burusho by properly differentiating between Northern EA vs Tibetian related EA?
@ Davidski
Yep Y-hg R1a-Z93 were originally introduced to the region due to what we broadly classify as 'Indo-Iranians', and post-Neolithic/ post -Corded Ware language expansion & acquisition was more complex than expanding male groups.
I think my initial comment relates more to the Bulan Koby period samples, The paper states that the Pazyryk & other Scythian groups collapsed due to 'the military campaigns of the Xiongnu Empire in the 2nd c. BCE, when the Bulan-Koby archaeological culture in the Mountainous Altai emerged' and that the 'Bulan-Koby exhibits major cultural influences from the Xiongnu and later steppe polities,'. However, the BK group is dominated by R1a-YP1542, which is a Tasmola linage, and their genomic profile shows a slight western shift in their qpAdm models (e.g. additional 'Sarmatian' ancestry). It is therefore hard to rationalise this often quited historical explanation with the aDNA evidence at hand.
I think the Reich group might clarify this in their paper, and suggest that BK is not actually Turkic whilst Shaz Turks have BK ancestry. But the latter apparently have additional Mongol ancestry, which might be associated with the appearance of certain Y-DNA C2 lineages (we''d need an expert on that to clarify).
The Bulgars and Huns in Europe (Lir-Turkic) are strongly associated with R1a-Z93 and some Q1. So it seems that the 'Turkification' of Inner Asian Iranians was almost entirely female-mediated, or came from a group in Mongolia which just happened to be heavily R1a-Z93 and deficient in Y-hg C.
@ Gio
''Ask these Turks why only R1b-L73 and the M478 subclade, we know is older in the Baltic and eastern Europe and probably (so I think) derived from Villabruna 14000 Years ago and relatives of his refugium?''
I don't think the issue of Villabruna refugium is directly relevant to Turkic expansions. Anyhow, R1 makes a funny circuit- R1b-L754 arrives to Europe from Siberia in the Ice Age, its ancestors then spread back East. R1b-M73 and R1b-PH155 feature in various Altaic groups, the latter is probably Central Asian anyway. BTW did you know I was the first (or second) to properly sequence Y-hg R1b-PH155, from a modern Italian man ?
Wherever Turkic speakers arrive, they will easily displace local languages and assimilate other languages. There isn't even necessarily any genetic contribution. The Kumyks are an example.
@Rob
"BTW did you know I was the first (or second) to properly sequence Y-hg R1b-PH155, from a modern Italian man?"
I studied that haplogoup more than 15 Years ago, and through the few STRs it was difficult to understand all, but after I convinced myself that perhaps the R-PH155 could be the only one old in Asia and not migrated from Europe, even though the last documentations in the aDNA weren't older than the Others like R-M73, certainly older in Europe, thus I left open the possibility that also R-PH155 could be from Europe. About the samples I studied I never excluded that they could have come in more recent times through the Turkish expansion, but evidently all these hgs went and came in the Siberian corridor.
The explanation of Davidski seems to me reliable.
@Rob
Thanks I read your link. Good read. And which model for celtic do you prefer currently?
@Radiosource
Yes there were many IE dialects that went extinct or became substrates. Temematic maybe a substrate in Balto-Slavic. Nordwestblock probably extinct. Atlantic bronze age dialect probably extinct or a substrate in celtic
Look at this map
https://ibb.co/7tLGfYvZ
Turkic is a state of mind. Signed by the Gagauz and the Yakut.
@Rob
Bulan-Koby, despite being conquered by the Xiongnu, likely didn't adopt a Turkic tongue. The Turks really arrived in the the mid 5th century and as can be seen in the ancestry results became more prevalent during the Turkic khaganate which had it's beginnings in the Altai region. A similar story happens with the Tagar culture and the subsequent Tashtyk culture. They persist until the 4th century and then get displaced by the Kyrgyz as can be seen with the arrival of the Chaatas culture which showcase many typical Turkic traits.
And the Altai is more relevant for Common Turks, while Oghuric Turks were already roaming the European steppes as can be seen with the Huns.
There are going to be exceptional cases at the periphery of an expansion, but at a macro-scale language expansions have associated genetic impacts. However to me it seems that, although Turkic has highlevels of
Slab Grave ancestry, their ethnogensis occurred in NW Mongolia/ Altai region during the post-Xiongnu period.
bulk genomic ancestry =/= nuanced ethnogenesis
@ Gio
''Others like R-M73, certainly older in Europe''
Yes, with a distinct sub-lineage L1432 moving east toward Botai and Omsk districts in Siberia c 4000 BC.
Yet in 2026 one can still read claims that R1b-M73 is a recent arrival to Europe from Lake Baykal (where it’s completely missing in the aDNA record), notably the “usual suspects” Troll Sloppers from FraudArchiver
@ Norfern
I agree with much of what you say, however will begin with this
''Bulan-Koby, despite being conquered by the Xiongnu, likely didn't adopt a Turkic tongue. '
Although the authors of the paper (and others elswhere) make the claim, it doesn't even make chronological sense. The Bulan Koba monuments date 300 - 500 AD, which is long after the Xiongnu 'empire' ended.
(e.g. Ref 1 and 2.
Even during the actual Xiongnu period, it is not clear they really 'conquered' the northwest periphery regions such as the Altai, although it would have been integrated into some sort of compelx sphere (See Permutations of Peripheries in The Xiongnu Empire, Bryan Miller).
The collapse of Scythian groups such as Pazyryk is often blamed to Xiongnu, but this was a systemic issue offending even the 'western Scythians' in the Black Sea c. 300 BC, far away from any Xiongnu implact. So the explanation must be more complex
'' A similar story happens with the Tagar culture and the subsequent Tashtyk culture''
Indeed, continuity
As for your second comment ''likely didn't adopt a Turkic tongue. ''
I gather a recent abstract by another team suggests similar, but BK were apparently assimilated into common Turkic propper, whilst Huns/Oghurs/Bulgars lack such ancestry.
Note the Mongolian input in some of (post-Huno-Bulgar) Turkic groups
However, again I would highlight the importance of chronology: Huns appear in Europe 300 years after the Xiongnu ended. This is more in line with the Xianbei period. Things are always more complex than the irresistable urge for simple equations (in this case XIongnu= Huns)
qpAdm Model
Target: Chamar_Haryana
Russia_MLBA_Sintashta.AG — 27.3%
SE: 0.0329 | Z: 8.30
IruIa.DG — 72.7%
SE: 0.0329 | Z: 22.1
Fit: p = 0.423
qpAdm Model
Target: Brahmin_Haryana
Iran_ShahrISokhta_BA2.AG — 65.4%
SE: 0.0595 | Z: 11.0
Russia_MLBA_Sintashta.AG — 25.6%
SE: 0.0455 | Z: 5.63
Onge.DG — 8.97%
SE: 0.0262 | Z: 3.43
Fit: p = 0.899
Both Upper caste Brahmins and Lower caste Chamar of Haryana have similar levels of Steppe. Only difference being Brahmins in this region have 26-30% AASI vs 45-50% AASI for Chamar.
Hard to tell if Chamar formed from mixing of high AASI population and steppe independent of Brahmins and Jats of the region or an IVC + high steppe population later mixed with high aasi population and became low caste. Or Brahmins formed from a high aasi population that mixed with steppe people with BMAC affinity.
Are there tools to test this?
@rob @davidski
Another model
Target
Brahmin_Haryana
PASS ✓
Feasible ✓
2-way admixture
p-value
0.495
Fit Quality
Excellent
chisq
9.39
Model
Weight
SE
z
Russia_MLBA_Sintashta.AG 19.3% 0.0452 4.27
Iran_ShahrISokhta_BA2.AG 80.7% 0.0452 17.9
My question is, in this model my chisq/dof ratio is 0.939 which is closer to 1 while previous model had it at 0.46.
Is chisq/dof ratio of 0.46 over fitting? And how to interpret it? Overfitted model has 25 steppe vs 20 on the second one. Chamar model has 1.02 ratio making it a better model?
I avow I prefer the reasoning of Dospaises to the Gioiello and Gaska 's ones concerning the still uncertain geographic origin of Y-R1b P297 and M269!
qpAdm Ancestry Model: Brahmin_UP
A high-confidence 2-way model for the Uttar Pradesh Brahmin profile.
Ancestry Weights:
• 64.8% I8728 (IVC-related)
• 35.2% Sintashta (Steppe MLBA)
Statistical Fit:
• P-value: 0.536 (Strong fit)
• Chi-sq (χ²): 9.93
• DOF: 11
• Z-score: 14.8 / 8.08
• Std. Error: 0.0436
Right:
Mbuti.DG
Iran_GanjDareh_N.AG
Georgia_Satsurblia_LateUP.SG
Turkey_Marmara_Barcin_N.AG
Israel_Natufian.AG
Morocco_Iberomaurusian.AG
Russia_Steppe_Eneolithic.AG
Russia_Tyumen_N.AG
Russia_Shamanka_EN.SG
Luxembourg_Mesolithic.AG
Mongolia_North_N.AG
China_Xinjiang_Xiaohe_BA.AG
Onge.DG
One interesting thing to note is:
Brahmin_UP (35% steppe) and Jatt (35% steppe) even though share about the same steppe have different paternal origins. Even though they share R1a line, ( would be interesting to see which group is the sub group or if they broke away early and took different paths ), Jats have >50% males under L1a2 and Q paternal lines.
Brahmins of UP though have a dominant R2 male line ~32% and with ~23% R1a makes more than half of the Brahmin_UP paternal lines.
qpAdm results for Norwegian.DG (1.15M SNPs)
Ancestry model:
• Yamnaya (Russia Samara EBA): 42.2% ±5.3 (z=7.90)
• Globular Amphora (Poland): 42.8% ±5.3 (z=8.05)
• Latvia EN: 15.0% ±4.7 (z=3.21)
Model fit:
• dof = 11
• chisq = 11.0
• p = 0.439 → good fit
Never modeled Norwegian on qpadm before. Is it accurate?
The Khvalynians have been returned to their places! As with the old one, the new Khvalynians fall into the same age range of 15,000 BC, which doesn't quite fit with the Chokh Mesolithic.
https://discover.familytreedna.com/y-dna/J-Z1841/tree
Ran QPADM and F3 for Pashtun_Yosufzai
https://ibb.co/fd0NrXn1
https://ibb.co/gb79m6qJ
They look like an Indo aryan population that switched to Iranian language. 68% IVC and 32% steppe.
@Davidski
Dear Davidski, could you please extract from Progress, Berezhnovka, and other Eneolithic steppe samples (from their raw files) the part that is autosomally Caucasian? Are there any tools that allow this to be done? And then convert it into G25 coordinates. I have already tried doing this in various ways using Vahaduo coordinates, but that’s not quite right — those end up being synthetic coordinates based on an optimal vector.
@Rob
BTW did you know I was the first (or second) to properly sequence Y-hg R1b-PH155, from a modern Italian man ?
Oh fascinating ! When was this done
@Davidski
David can you please get rid of this troll Ash/Rudraman spamming the comment section with Bullshit models, giving serious Sridhar/Vashishta vibes. Merci
@ Kouros
It was approx 2015 or 2016, we noted there was a chap with STR c/w M335, first known to me via the study Excavating Y-chromosome haplotype strata in Anatolia; Cinnioğlu et al 2004. IIRC he might have been Italian American, but to protect privacy we wont go into more details. I submitted the data to ftDNA & YFull for expert Y DNA analysis, freely of course for the greater good.
@ rozbójnik
''Thanks I read your link. Good read. And which model for celtic do you prefer currently?''
If you asked me 10 years ago, when the buzz around steppe, BB and modern continuity led some to propose 'Atlantic Bronze Age Celtic', I would have stood in the 'traditionalist' Continental camp. But now, as recent genetics papers are swinging back to Urnfield models, Im not so sure.
For one, at its core in the Carpathian basin the Urnfield phenomenon seems to arise amongst the 'Tell communities' which have a significant 'pre-IE' genetic substrate. Obviously 'pre-IE' is a somewhat anachronistic term when talking about 2000-1200 BC, but as Alex & Radio have pointed out there must have been pre-IE speech communities through Europe, not just the Aegean & Iberia. Although some of these Urnfielders migrated about (e.g. see https://link.springer.com/article/10.1007/s10963-022-09164-0), in many other regions cremation was gradually adopted by local groups, incl IE Nordic, Slavic, Celtic, etc. Secondly, I vaguely recall reading the Patterson suppl. that the Urnfield -like migrants seen in southeast Britain weren't exactly elite / warrior groups, in fact they didn't seem to be of an elevated status and could have been pragmatic migrants from central-east Europe, somewhat like recent times :) Thirdly, even if we can acccept that continental groups created a language shift in LBA southeast Britain (from whatever pre-Celtic idiom); we would need to explain how Scotland and Ireland became 'Celtic'. Because they do not appear to have been impacted by these genetic shifts, nor can we easily invoke 'language shift' as there was not unified British 'empire' or confedracy at the time. There might be an element of both. .
Post a Comment