Page 1 of 8 123 ... LastLast
Results 1 to 10 of 77

Thread: Saka, Scythians, Sarmatians & Huns: Overview (G25 Data)

  1. #1
    Administrator
    Posts
    3,258
    Sex
    Y-DNA
    R2a*-M124 (L295-)
    mtDNA
    D4j5*

    England

    Lightbulb Saka, Scythians, Sarmatians & Huns: Overview (G25 Data)

    Briefly, this thread will review the differing admixture proportions between the various Eurasian pastoral nomad samples we currently have from the Iron Age and Classical periods based on Eurogenes G25 data and poi's excellent nMonte runner (remember to donate to its' upkeep!). This can be considered an extension of Kulin's assessment here.


    Method
    The ref pop choices are self-explanatory (note that Afanasievo is interchangeable with Yamnaya Samara here). I usually refrain from using more than six ref pops in G25 nMonte simulations, but the breadth of the territory covered does necessitate a somewhat large number. All of these populations are distinct from one another. I've elected to use only one non-ancient as a ref pop (my qpAdm-guided AASI simulation from Pakistan-Afghanistan) for obvious reasons (we don't have any predominantly AASI aDNA just yet).

    The original iteration of this run had a ref pop source for CHG. However, most of the target samples didn't require it (<1%) and only two or three registered anything above 1% (<2%). Ergo, there doesn't appear to have been any significant gene flow from the Caucasus into the Eurasian steppes during the Iron Age to Classical period based on our current samples.

    A second iteration using just the Hovsgol Mongolian BA samples as an East Eurasian source was attempted. While the fits worked very well for everyone else, certain Hun and "Nomad_IA" samples produced substandard fits (>5) and scored 100% Hovsgol BA. Inspecting the check fit demonstrated that those samples seemed to pair fairly well with modern Tungusic-speaking groups. I performed a quick test to discern which of these populations was the least West Eurasian (below in spoilers; cycle=1k, batch=500, pen=def). This suggests an additional East Eurasian source that's present in modern Tungusic speakers played a role in the ethnogenesis of these groups. They aren't a simple expansion of Hovsgol BA westwards. I elected to use the Oroqen (0% W. Eurasian, lowest fit) as a result.

     

    Code:
    	Sample	Details	Fit	Map	Eskimo Naukan	Han	RUS Sintashta MLBA	TKM Parkhai EBA	USA Ancient Beringian
    1	Evenk:Average		21.77 	Open Map	68.4	31.6	0	0	0
    2	Mogush:Average		12.2149 	Open Map	37.8	46.8	14.8	0.6	0
    3	Oroqen:Average		11.1891 	Open Map	29.6	70.4	0	0	0
    4	Tuvinian:Average		13.0665 	Open Map	38.8	47.6	13.4	0.2	0
    5	Ulchi:Average		12.5751 	Open Map	38.6	61.4	0	0	0


    Results
    Please see below.

    Code:
    Sample				Fit	Botai	Hovsgol BA	Oroqen	Globular Amphora PL	Afanasievo	Sintashta	Sim AASI NW By DMXX	Parkhai EN
    Hun_Tian_Shan:Average		2.0889	5.2	27.8		4.6	2.2			16		36.8		2.6			4.8
    Hun_Tian_Shan_o:Average		2.9309	0	67.4		25.6	0.2			0.6		5.6		0.2			0.4
    KAZ_Hun-Sarmatian:Average	2.5186 	0	52.4		46.8	0			0.2		0.6		0			0
    KAZ_Nomad_HP:Average		3.6153 	0	57.2		41.2	0.4			0.2		1		0			0
    KAZ_Nomad_IA:Average		2.8719 	12	16.8		2.6	2.4			25		33.8		2.2			5.2
    KAZ_Nomad_Med:Average		2.8806 	0	60.6		10.6	1.4			8.6		16.4		1			1.4
    KGZ_Nomad_Med:Average		3.725 	0	60.6		34	0.4			1		3.4		0			0.6
    RUS_Nomad_Med:Average		2.4454 	0	37.8		6.8	3			0.8		47.8		0.6			3.2
    Saka_Kazakh_steppe:Average	1.6451 	1.2	49.8		3.6	2.4			15.4		24.6		0.4			2.6
    Saka_Kazakh_steppe_o1:Average	3.2748 	0.8	8.2		1.6	0			18.6		57.2		2.6			11
    Saka_Kazakh_steppe_o2:Average	3.3434 	0.2	9		1.8	0			17.4		65.6		0.8			5.2
    Saka_Tian_Shan:Average		3.0204 	3	16		4	1			29.8		37.6		2.2			6.4
    Saka_Tian_Shan_o:Average	2.6643 	10.8	26.6		5	3			3.6		44.8		1.8			4.4
    Sarmatian_KAZ:Average		2.0229 	0	8.4		1.4	3			38		43		0.2			6
    Sarmatian_RUS_Casp_ste:Average	1.5573 	2	9		2	2			29		48.6		0.4			7
    Sarmatian_RUS_Pokrovka:Average	1.5593 	0.6	7.8		1.8	0			31.2		52.6		0.6			5.4
    Sarmatian_RUS_Urals:Average	1.6352 	1.6	7.4		2.8	2.2			26		53.8		0.6			5.6
    Scythian_Aldy_Bel_IA:Average	1.9364 	4	51.6		3	1.8			14.4		23		0.4			1.8
    Scythian_HUN:Average		2.8199 	0	0		0	42.4			0		57.4		0			0.2
    Scythian_MDA:Average		4.9562 	0	1		0.2	45.2			0		45.8		0.4			7.4
    Scythian_MDA_o2:Average		2.8724 	0	0		0	18.8			0		81.2		0			0
    Scythian_UKR:Average		3.0679 	0	0.4		0.6	16.6			0		80.8		0			1.6
    Scythian_Zevak_Ch_IA:Average	2.954 	0	44.6		4.6	3.4			12.4		30.6		0.8			3.6
    Analysis
    • Botai persists sporadically around the Tian Shan, but the highest value is seen in a Kazakh nomad.
    • Hovsgol BA is present in most of the samples. The frequency varies wildly. There isn't an obvious relationship between the Hun vs. Saka assignment and the observed frequency. Those Scythians from the European steppe (Hungary, Ukraine, Moldova) lack this form of ancestry.
    • Oroqen-like ancestry largely mirrors the pattern seen from the Hovsgol BA component.
    • Trace surplus EEF (Global Amphora) is located east of the Urals (0-3%). It forms the majority of the non-steppe ancestry in the European Scythians. The "Moldovan_outlier2" sample is so because it contains less GA and more MLBA steppe relative to the others. However, it's identical to the Ukrainian Scythian.
    • Afanasievo ancestry is truly "Afanasievo" in those samples from East-Central Asia. Here, we see it reach a maximum of 29% in the Saka Tian Shan average. A moderate presence exists in the Kazakh steppe (Saka, Hun). The non-Hun Sarmatians register plenty, but this is probably persistent ancestry from the EMBA period. The European Scythians score none.
    • Sintashta is, unsurprisingly, the primary component across most of the Scythian and Saka individuals. It reaches a maximum in the European Scythians (though it's not possible to determine from this run whether that's strictly true).
    • My AASI simulation is picked up in a handful of the East-Central Asian samples (approaching 3% in some). It seems to correlate with Iranian-Turanian agriculturalist admixture, and we know AASI was present in South-Central Asia around the Bronze Age onwards. As expected, absent everywhere west of the Kazakh steppe.
    • Parkhai_EN is present in a handful of East-Central Asian samples. Curiously, it's picked up in two of the European Scythians (the Moldovan reaches up to 8%).


    All thoughts welcome.
    Last edited by DMXX; 05-04-2019 at 11:54 AM. Reason: correction to sentence in method RE: CHG

  2. The Following 28 Users Say Thank You to DMXX For This Useful Post:

     Просигој (05-04-2019),  agent_lime (05-06-2019),  Alain (05-14-2019),  anglesqueville (05-04-2019),  Arame (05-07-2019),  Awale (05-08-2019),  Drewcastle (05-04-2019),  Garimund (05-08-2019),  Hando (05-06-2019),  Kopfjäger (05-04-2019),  Kulin (05-06-2019),  Megalophias (05-07-2019),  NiloSaharan (05-03-2019),  oz (05-12-2019),  palamede (05-04-2019),  parasar (05-04-2019),  pmokeefe (05-08-2019),  poi (05-06-2019),  Power77 (05-06-2019),  prashantvaidwan (05-06-2019),  Pribislav (05-06-2019),  Psynome (05-03-2019),  rms2 (05-04-2019),  Ruderico (05-06-2019),  Ryukendo (05-06-2019),  szahel (05-04-2019),  Táltos (05-10-2019),  Targum (05-03-2019)

  3. #2
    Administrator
    Posts
    3,258
    Sex
    Y-DNA
    R2a*-M124 (L295-)
    mtDNA
    D4j5*

    England
    Repeated the same procedure for all the individual Medieval Turkish samples from the steppe (Kimak, Karluk, Karakhanid). At work right now, but will post the outputs when I return home.

    They were largely a mixture of Hovsgol-Oroqen and Sintashta, the ratio of which was around 30-50% vs. 30-40% from memory. Botai was registered but not significant. Somewhat more Iranian-Turanian agriculturalist ancestry is present relative to the Tian Shan samples. They also scored a pinch of extra EEF.

    Afanasievo ancestry appeared consistently, though at lower levels than what was observed in the Tian Shan samples. The significant showing made by Afanasievo in these samples neatly accompanies the linguistic evidence of multiple Tocharian loanwords in Turkic. The general decrease in Afanasievo-like ancestry from the EMBA steppe period onwards in E-C Asia suggests that they simply melted into the surrounding populations in an irregular fashion over time.

    Those of us who've insisted for years (such as yours truly) that the the first Turks had to have been a West-East Eurasian intermediary population based on the physical anthropological and historical data appear to have been correct on that call.

    It looks to me like the Medieval Turks were an (again) irregular mixture of a group similar to the "Scythian" Zev_Chik site and a more Sintashta and Iranian agriculturalist-rich population. The Kangju look like ideal candidates for that. I might have a go at modelling them later on as well.

    The above has implications with respect to modelling Turkish, Azerbaijani, Turkmen and other related groups. Clearly, using East Asians as a catch-all for the sum East Eurasian admixture and reverse-calculating the values in "elemental" models is no longer sensible, because...
    1) "Han" or "Nganassan" do not serve as reliable proxies for Botai admixture
    2) The amount of East Eurasian admixture varies substantially between the five Medieval Turks examined

    I also assessed the Medieval Ottoman Turkish samples - One of them (ID ...95, with whom I share mtDNA with) was up to 90% Kimak-Karakhanid-Karluk-like, with the remainder being West Asian. The other was predominantly Anatolian overall.

    G25 is great for modelling purposes and for revealing overall trends, so I'd expect these qualitative inferences to mostly match any formal stats-based assessment of these groups, though the proportions are likely to differ. Just pointing that out in advance.

    Stay tuned!

  4. The Following 7 Users Say Thank You to DMXX For This Useful Post:

     Amber29 (05-06-2019),  Garimund (05-08-2019),  Hando (05-06-2019),  pmokeefe (05-08-2019),  Power77 (05-06-2019),  prashantvaidwan (05-06-2019),  Táltos (05-10-2019)

  5. #3
    So is Sintashta actually Proto-Tocharian rather than Proto-Indo-Iranian?

  6. The Following User Says Thank You to talljimmy0 For This Useful Post:

     Hando (05-06-2019)

  7. #4
    Administrator
    Posts
    3,258
    Sex
    Y-DNA
    R2a*-M124 (L295-)
    mtDNA
    D4j5*

    England
    Quote Originally Posted by talljimmy0 View Post
    So is Sintashta actually Proto-Tocharian rather than Proto-Indo-Iranian?
    No. Per the Anthony-Ringe model, Sintashta is most likely an Indo-Iranian culture, with Tocharian most likely being the language of the Afanasievo culture. The material items observed in Sintashta match Indo-Iranian's reconstructed words better than Afanasievo does (that and the multitude of other features of Tocharian which suggest prolonged separation from the IE urheimat, which Afanasievo represents geographically).

    Sintashta-derived ancestry generally predominates over Afanasievo-derived ancestry across East-Central Asia, which is also in keeping with the archaeological data, where all the Iron Age cultures observes there (Tagar, Tashyk, Karasuk etc.) look Andronovo-derived (which in turn appears Sintashta-derived). The only Afanasievo-derived culture I'm aware of (Okunevo) was materially replaced by the above.

    So, there's a real-world, observable, prehistoric phenomenon which seems to match what we're observing here.

    (If I have to type "derive" one more time today, I'll "..." no pleasure in doing so, that's for sure!)

  8. The Following 8 Users Say Thank You to DMXX For This Useful Post:

     Awale (05-08-2019),  Garimund (05-08-2019),  Hando (05-06-2019),  Nibelung (05-06-2019),  palamede (05-06-2019),  poi (05-06-2019),  Power77 (05-06-2019),  Ruderico (05-06-2019)

  9. #5
    Administrator
    Posts
    3,258
    Sex
    Y-DNA
    R2a*-M124 (L295-)
    mtDNA
    D4j5*

    England
    Quick comment, relating to the Saka Tian Shan samples.

    There's been a tendency in our community to use these to model SC Asian populations (particularly Pashtuns) based on the hypothesis linking them to Hepthalite and/or Yuezhi admixture.

    While it's reasonable prima facie, notice that the STS average and the outlier have between 10-12% non-MLBA steppe admixture with a form that just so happens to represent the types of ancestry that are typically observed in SC Asians (AASI, extra Iranian-Anatolian-Turanian agriculturalist).

    So, the statistical favouring of STS over Sintashta in certain SC Asians looks like a red herring (due to the former having minor but significant SC Asian-like ancestry that Sintashta, Srubnaya etc. lack). That's the simplest reason regarding why some Pashtuns or Tajiks seem to love STS (Occam's razor applies here). One needs a more robust indicator of direct ancestry to prove a special relationship, given this.

  10. The Following 5 Users Say Thank You to DMXX For This Useful Post:

     Garimund (05-08-2019),  Hando (05-06-2019),  Kulin (05-06-2019),  Ruderico (05-06-2019),  Sapporo (05-07-2019)

  11. #6
    Moderator
    Posts
    651
    Sex
    Location
    Sarai-Batu
    Ethnicity
    Bengali
    Nationality
    Canadian
    Y-DNA
    J2b2a
    mtDNA
    A11-a*

    Canada Bangladesh Jammu and Kashmir India Pakistan Sri Lanka
    Thank you for the really good runs, both historical and modern Central Asia is very fascinating, and the Saka/Scythians are a very interesting population in themselves.

  12. The Following 6 Users Say Thank You to Kulin For This Useful Post:

     Amber29 (05-06-2019),  DMXX (05-06-2019),  Garimund (05-08-2019),  Hando (05-06-2019),  poi (05-06-2019),  Ruderico (05-06-2019)

  13. #7
    Registered Users
    Posts
    143

    Quote Originally Posted by DMXX View Post
    .
    Those of us who've insisted for years (such as yours truly) that the the first Turks had to have been a West-East Eurasian intermediary population based on the physical anthropological and historical data appear to have been correct on that call.
    Just to be clear, you are referring to to the first groups to be historically recorded as Turks--Gokturks and onward, is that right?

    I suspect that proto-Turkic speakers probably were a mainly Baikal HG-like east Eurasian population, but mixed with Afanasievo and Sintashta derived peoples beginning in the Bronze age and with other groups like Iran related farmers through the Iron age and during historical times. Would you agree with that scenario?

  14. The Following 4 Users Say Thank You to Psynome For This Useful Post:

     Awale (05-08-2019),  DMXX (05-06-2019),  palamede (05-06-2019),  parasar (05-06-2019)

  15. #8
    Administrator
    Posts
    3,258
    Sex
    Y-DNA
    R2a*-M124 (L295-)
    mtDNA
    D4j5*

    England
    Quote Originally Posted by Psynome View Post
    Just to be clear, you are referring to to the first groups to be historically recorded as Turks--Gokturks and onward, is that right?

    I suspect that proto-Turkic speakers probably were a mainly Baikal HG-like east Eurasian population, but mixed with Afanasievo and Sintashta derived peoples beginning in the Bronze age and with other groups like Iran related farmers through the Iron age and during historical times. Would you agree with that scenario?
    Correct - By "first Turks", I refer to those who were historically attested.

    I also agree with your proposition. Early influence from IE speakers looks likely to me.

  16. The Following 4 Users Say Thank You to DMXX For This Useful Post:

     Awale (05-08-2019),  Hando (05-06-2019),  Psynome (05-06-2019),  Ruderico (05-06-2019)

  17. #9
    Registered Users
    Posts
    6,386

    Quote Originally Posted by Psynome View Post
    Just to be clear, you are referring to to the first groups to be historically recorded as Turks--Gokturks and onward, is that right?

    I suspect that proto-Turkic speakers probably were a mainly Baikal HG-like east Eurasian population, but mixed with Afanasievo and Sintashta derived peoples beginning in the Bronze age and with other groups like Iran related farmers through the Iron age and during historical times. Would you agree with that scenario?
    I would agree based on where they were initially located 1085 miles east of Loulan to the west of the Wei lands.

    "In the suburb Wen-I, to the north-east of the city of Lo-Yang, was the dwelling of Sung-Yun of Tun-hwang who, in company with the Bhikshu Hwei Säng, was sent on an embassy to the western countries by the Empress Dowager (Tai-Hau) of the Great Wei dynasty to obtain Buddhist books. This occurred in the eleventh month of the first year of the period Shén kwei (517-518 A.D.) They procured altogether 170 volumes, all standard works, belonging to the Great Vehicle. First of all, having repaired to the capital, they proceeded in a westerly direction forty days, and arrived at the Chih Ling (Barren Ridge), which is the western frontier of the country. On this ridge is the fortified outpost of the Wei territory. The Chih-Ling produces no trees or shrubs, and hence its name (Barren). Here is the common resort (cave) of the rat-bird. These two animals being of different species (chung), but the same genus (lui), live and breed together. The bird is the male, the rat the female. From their cohabiting in this manner, the name rat-bird cave is derived. Ascending the Chih-Ling and proceeding westward twenty-three days, having crossed the Drifting Sands, they arrived at the country of the Tuh-kiueh-'hun.
    Along the road the cold was very severe, whilst the high winds, and the driving snow, and the pelting sand and gravel were so bad, that it was impossible to raise one's eyes without getting them filled. The chief city of the Tuh-kiueh-'hun and the neighbourhood is agreeably warm. The written character of this country is nearly the same as that of the Wei. The customs and regulations observed by these people are mostly barbarous in character (after the rules of the outside barbarians or foreigners). From this country going west 3500 li, we arrive at the city of Shen-Shen [Loulan]. This city, from the time it set up a king, was seized by the Tuh-kiueh-hun, and at present there resides in it a military officer (the second general) for subjugating (pacifying) the west. The entire cantonment” amounts to 3000 men, who are employed in withstanding the western Hu."


  18. The Following 5 Users Say Thank You to parasar For This Useful Post:

     Alain (05-15-2019),  Garimund (05-08-2019),  Hando (05-10-2019),  palamede (05-07-2019),  spruithean (05-06-2019)

  19. #10
    Administrator
    Posts
    3,258
    Sex
    Y-DNA
    R2a*-M124 (L295-)
    mtDNA
    D4j5*

    England
    I've performed the same setup for the individual Alan and Moldovan Cimmerian samples (replaced AASI with Maykop):

    Code:
    	Sample	Details	Fit	Map	KAZ Botai	MNG Hovsgol BA	Oroqen	POL Globular Amphora	RUS Afanasievo	RUS Maykop Late	RUS Sintashta MLBA	TKM Parkhai Eneolithic
    1	MDA_Cimmerian:cim357	Ancient; BCE:875 	2.9568 	Open Map	0.2	7.4	2.4	3.4	17.6	4.6	63	1.4
    2	MDA_Cimmerian:cim358	Ancient; BCE:875 	3.8211 	Open Map	1.2	17.4	2.8	0.4	12.2	13.4	46.8	5.8
    3	MDA_Cimmerian:cim359	Ancient; BCE:923 	2.7322 	Open Map	6.2	41.2	3.8	2.8	13.8	3.6	27	1.6
    4	RUS_Alan_Med:DA146	Ancient; CE:700 	4.3008 	Open Map	0	0	0	0.8	7.8	55.2	17	19.2
    5	RUS_Alan_Med:DA160	Ancient; CE:700 	3.3658 	Open Map	0	2.6	0.8	8	1.4	77.2	10	0
    6	RUS_Alan_Med:DA162	Ancient; CE:250 	1.8012 	Open Map	0	1.4	0.2	7.2	1.4	77.6	11.8	0.4
    7	RUS_Alan_Med:DA164	Ancient; CE:1150 	2.7429 	Open Map	0	1.8	0.4	4.6	1.4	67.8	24	0
    8	RUS_Alan_Med:DA243	Ancient; CE:350 	1.9302 	Open Map	0	1.8	0.8	2.8	8	66.6	20	0
    I've italicised the ones with fits >3 for easier interpretation.

    I'm surprised by the East Eurasian admixture in the Cimmerians (Cim359 is almost half). This is somewhat baffling.

    The Alan samples are predominantly Maykop-derived, with their MLBA steppe ancestry being between 10-24%. They are effectively carbon copies of modern Northern Caucasians. I inspected the fit list for four of them and was somewhat surprised to see that they were closer overall to modern Iranians than they were to Pamiri Tajiks. Another interesting find was that Dzharkutan2_BA (one of the Central Asian samples) was quite close to them overall, as well. Afanasievo-related ancestry once again appears, albeit with a lower frequency in comparison to the Sarmatian samples.

    The Alans look like a very simple, two-ingredient mixture of Sarmatians and post-Maykop indigenous Northern Caucasians:

    Code:
    Sample	Details	Fit	Map	RUS Maykop Late	Sarmatian RUS Caspian Steppe
    1	RUS_Alan_Med:Average		2.1284 	Open Map	72.2	27.8
    This is consistent with their overall similarity with the modern populations in the region (as well as explaining why they are closer to modern Iranians than the more steppe-rich Pamiris).

    I also note that the Alans with decent fits don't register any surplus Parkhai_EN, despite the Sarmatians doing so (OP). I'm assuming the Maykop Late base is absorbing some of this. I'm reminded of Herodotus' account in Histories regarding the Sarmatians being a back-migration from Central Asia into the P-C steppes. Looks like we have some (weak, informal) evidence of this here.
    Last edited by DMXX; 05-08-2019 at 02:21 AM. Reason: typos

  20. The Following 10 Users Say Thank You to DMXX For This Useful Post:

     Просигој (05-08-2019),  Alain (05-15-2019),  Awale (05-08-2019),  FrostAssassin0701 (05-08-2019),  Garimund (05-08-2019),  Hando (05-10-2019),  Kulin (05-08-2019),  palamede (05-08-2019),  Psynome (05-09-2019),  Ruderico (05-08-2019)

Page 1 of 8 123 ... LastLast

Similar Threads

  1. Are the Scythians/Sarmatians Iranic?
    By Brent.B in forum Ancient (aDNA)
    Replies: 206
    Last Post: 04-24-2019, 03:43 PM
  2. Replies: 123
    Last Post: 01-24-2019, 09:15 PM
  3. Q-L938 Huns?
    By kharrington in forum Q
    Replies: 5
    Last Post: 12-12-2018, 07:40 PM
  4. Replies: 5
    Last Post: 10-06-2018, 04:59 AM
  5. [Split] Origins of the Saka
    By pegasus in forum History (Ancient)
    Replies: 20
    Last Post: 07-11-2015, 01:52 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •