Wednesday, October 3, 2018

Cimmerians, Scythians and Sarmatians came from...


Apparently they all came from the eastern Pontic-Caspian steppe. There's a new paper about that at Science Advances (see here). Below is the abstract, emphasis is mine:

For millennia, the Pontic-Caspian steppe was a connector between the Eurasian steppe and Europe. In this scene, multidirectional and sequential movements of different populations may have occurred, including those of the Eurasian steppe nomads. We sequenced 35 genomes (low to medium coverage) of Bronze Age individuals (Srubnaya-Alakulskaya) and Iron Age nomads (Cimmerians, Scythians, and Sarmatians) that represent four distinct cultural entities corresponding to the chronological sequence of cultural complexes in the region. Our results suggest that, despite genetic links among these peoples, no group can be considered a direct ancestor of the subsequent group. The nomadic populations were heterogeneous and carried genetic affinities with populations from several other regions including the Far East and the southern Urals. We found evidence of a stable shared genetic signature, making the eastern Pontic-Caspian steppe a likely source of western nomadic groups.

Krzewinska et al., Ancient genomes suggest the eastern Pontic-Caspian steppe as the source of western Iron Age nomads, Science Advances, 03 Oct 2018: Vol. 4, no. 10, eaat4457, DOI: 10.1126/sciadv.aat4457

Update 04/10/2018: Twenty four of the ancient nomad samples made it into the Global25 datasheets. Look for the following population codes: Cimmerian_MDA, Sarmatian_RUS_Urals, Scythian_MDA, Scythian_UKR and Srubnaya-Alakulskaya_MLBA. Feel free to put them through their paces and share the results with us in the comments below.

Global25 datasheet ancient scaled

Global25 pop averages ancient scaled

Global25 datasheet ancient

Global25 pop averages ancient

See also...

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

52 comments:

  1. This time all of the Scythians belong to R1b, while one of the Sarmatians belongs to R1a.

    Sarmatian chy002 R1a1a
    Sarmatian tem002 R1b1a1a2
    Sarmatian tem003 R1b1a1a2?
    Scythian scy009* R1b1a1a2
    Scythian scy193* R1b1a1a2?
    Scythian scy197* R1b1a1a2
    Scythian scy301 R1b1a1a2
    Scythian scy304 R1b1a1a2
    Scythian scy305* R1b1a1a2
    Cimmerian cim357 R1b1a
    Cimmerian cim358 Q1a1
    Srubno-alakulskaya kzb002 R1a1a1
    Srubno-alakulskaya kzb003 R1a1a1
    Srubno-alakulskaya kzb005 R1a1a1
    Srubno-alakulskaya kzb007 R1a1a1
    Srubno-alakulskaya kzb008 R1a1a1
    Srubno-alakulskaya mur003 R1a1a1?

    ReplyDelete
  2. "This time all of the Scythians belong to R1b"
    !!!!!!

    Exciting. No subclade yet?

    ReplyDelete
  3. In my excitement I neglected to read your entire comment =p

    ReplyDelete
  4. Someone will probably try and check out the subclades from the BAM files soon.

    I guess most of these R1's are Z2103 and Z93, but I'm hoping for a surprise like L51 or even U106.

    ReplyDelete
  5. @Davidski
    Yeah, it would be great to get an R1b that wasn't Z2103 from one of these guys, especially in light of this business with the Hallstatt culture.

    ReplyDelete
  6. Important: the samples that they call Cimmerians are actually belong to Chernogorovka culture. Not everyone agrees with identification of that culture with Cimmerians: http://www.academia.edu/1084093/The_Current_State_of_the_Cimmerian_Problem_Ancient_Civilizations_from_Scythia_to_Siberia_7_2001_307-340

    ReplyDelete
  7. They don’t even mention Iranians or Indo-Iranians with respect to Scythians and Sarmatians here as they don’t seem to be descendent from Srubnaya/Sintashta tribes.

    ReplyDelete
  8. @All

    Lousy trolling attempt from Azarov Dimitry blocked in moderation.

    Note to Dimitry: if you're going to troll here, at least make up something halfway believable. The claim that Scythians came from Anatolia is not it. LOL

    ReplyDelete
  9. It looks like these R1b Cimmerians, Scythians and Sarmatians originated from some Yamnaya-Caucasus tribes and were very fond of stealing wives from Greek, Celtic or some East Asian tribes. Just like Tatars were stealing Slavic women later on.

    https://i.postimg.cc/g2CkXPtt/Scythians.png

    ReplyDelete
  10. Interesting, no H6a mtDNA haplogroup in these samples

    ReplyDelete
  11. @Lee Albee,

    In 136 mtDNA samples from Sintashta, Andronovo, Srubnaya, unadmixed descendants H6a1 is at a frequency of 2.5%. In 153 mitogenomes from Bell Beaker (xIberia, Italy, Hungary) H6a1 is at a frequency of 2.6%.

    It is not a surprise that H6a1 does not pop up with the small amount of mtDNA samples from this study.

    ReplyDelete
  12. Did the Wietenburg Culture spread into the Ukraine at some point ? If some of those turn out to be L51, could it be that Suvorovo related people survived in the Mures/Tisza area and contributed to these Scythians and earlier to Eastern Bellbeaker ? Just a bit of speculation..

    ReplyDelete
  13. I remember some people dreaming that Pontic Scythians will be R1a and Slavic like:)

    ReplyDelete
  14. @All

    Someone has indeed started to look at the BAM files...

    http://forum.molgen.org/index.php/topic,11272.15.html

    ReplyDelete
  15. Slight remixing of the graphics from the paper, as I thought would be useful to see the ADMIXTURE along with the clustering for the Scythians: https://imgur.com/a/Y3Zozwt

    The North European like Scythians have about two at the intersection of Czech-English in the PCA (scy009 and scy303) and one outside the distribution of modern day Hungarians (scy010). 009 is male while 303 and 010 are female. SE Cluster overlap modern day Greeks while other clusters overlap no one.

    Be interesting to see these samples dropped into the Eurogenes PCAs with higher resolution to modern day intra-European differences. At least scy009 and scy303 should work in the North European PCA, Global25 probably better for 010.

    ReplyDelete
  16. Fifteen comments in and not one mention of Conan.

    ReplyDelete
  17. @All

    Twenty four of the ancient samples are now in the Global25 datasheets. Check out the update with the links.

    ReplyDelete
  18. OK, normal tree with averages I always run: https://imgur.com/a/tY1fIuh (new samples / averages in black)

    As for additional observations, not much. Multiple samples look unusual on PCA separating Han_N from Man_Bac+Nganasan, PC11 (e.g. scy011, cm359, scy300 all very slightly offset towards Han relative to Man_Bac+Nganasan).

    Samples that seem to share in the Balto-Slavic drift trend are: scy009 (quite a bit), cim358, scy010 (both less so). On dimensions that specifically break that out. scy009 tenatively looks to have a long branch off between Polish+Avar_Hungary_Szolad and Slovakian because of unusual WHG+EEF dynamics (e.g. more WHG+EEF, less steppe than typical), but not sure about that.

    Some other trends of samples being outside range in some dimensions but not sure if these are real or artefactual, so not much point talking about them.

    ReplyDelete
  19. @All

    Here are the revised Y-haplogroup assignments by one of the YFull guys.

    https://docs.google.com/spreadsheets/d/1bnVJujBs_bQu_dqSVi_dSXUuz9gNIYFX_XlqRrz92mo/htmlview

    Interestingly, in this list, all of the Sarmatians belong to R1a, one of the Cimmerians belongs to R1a, at least one of the Srubnaya samples belongs to R1a-Z280, and one of the Scythians might belong to R1a as well.

    ReplyDelete
  20. @Matt

    Samples that seem to share in the Balto-Slavic drift trend are: scy009 (quite a bit), cim358, scy010 (both less so).

    I noticed that, and checked out scy009 in some detail. That's a very Polish Scythian indeed.

    ReplyDelete
  21. SCY009 seems to be yDNA R1b1a1a2-P312 and mtDNA J2b1a6. On Ian Logan mtDNA database, J2b1a6 is found in Poland, Russia, Moldova, Hungary and Caucasus(http://www.ianlogan.co.uk/sequences_by_group/j2b_genbank_sequences.htm)

    ReplyDelete
  22. Lol Dave everyone is Polish! .
    BTW Thanks Dave.

    ReplyDelete
  23. @All

    There's a new paper on the Neolithic transition in North-Central Poland.

    https://www.nature.com/articles/s41598-018-33067-w#Sec21

    R1a shows up during the Early Bronze Age, but we knew this already, and the sample in question is this one...

    http://eurogenes.blogspot.com/2017/01/r1a-z280-from-early-bronze-age-northern.html

    ReplyDelete
  24. This time Corded Ware is I2a2a. The two Corded Ware samples from Pikutkowo Poland (c. 2500-2300 BC) that had their yDNA resolved are I2a2a.

    ReplyDelete
  25. One of the Czech Corded Ware samples is also I2a2a.

    Awesome substructure within Czech Corded Ware

    Paternal farmer admixture.

    ReplyDelete
  26. I wonder if BKG N22 will fit as the source of WHG in Baltic_BA.
    Also two CWC samples (N49 and N47) plot in an interesting position.

    ReplyDelete
  27. Both CWC individuals with I2a2a (N47 and N49) were related to each other in the third order (Supplementary Table S6. Relatedness coefficients for the Corded Ware individuals.)

    ReplyDelete
  28. I'm curious as to how the original authors and the YFull people can get Y-haplogroup calls that are so different from each other (even disagreeing on R1b/R1a assignments).
    Did someone make a mistake somewhere, or do you have any other suggestions to explain this difference?

    ReplyDelete
  29. Looks like scy009 is R1b->U152->L2. Same paternal line as myself. Maybe Davidski is on to something with his theory of slight Celtic-Scythian genetic ties in ancient times.

    ReplyDelete
  30. @Samuel Andrews

    So Bell beaker--Does that include Iberia or exclude Iberia.

    It matters.

    ReplyDelete
  31. I wonder when and where I1 will turn up. Considering its current highest diversity it will be NW Germany IIRC.

    ReplyDelete
  32. Mlukas has some real issues with slavic people. Not only here, but on Polish sites as well. Anyway, any chance od Z282 popping up among Sarmatians?

    ReplyDelete
  33. Davidski: I noticed that, and checked out scy009 in some detail. That's a very Polish Scythian indeed.

    yep, scy009 is very close to the Polish average in the dimensions tracking Baltic_BA relatedness specifically, seems a little out of scope in WHG:EEF:Steppe, hence placing closest to Polish+Avar but the long branch in clustering.

    graphically, using a reprocessing PCA without Africans: https://imgur.com/a/fUiqYmc

    ReplyDelete
  34. @Lee Albee,

    Excluding Iberia, Italy, Hungary Bell Beaker. I also include Early Bronze age samples from Britain & Germany because they're probably overwhelmingly of Beaker origin.

    ReplyDelete
  35. @epoch,

    I1 will pop up in Scandinavia. It already has. One I1 dating to 2200 BC, one to 1400 BC.

    ReplyDelete
  36. @Mikkel

    I think the problem is that there's currently no standard method to assign Y-haplogroups available to academic researchers that has been proven to work accurately with ancient DNA, and especially with low coverage genomes. They just seem to do their own thing each time.

    ReplyDelete
  37. @All

    Interestingly, the YFull Y-haplogroup assignment for scy197 (E-V13) correlates very well with his placement in Matt's tree. He's located very close to Albanians and other Balkan groups.

    https://imgur.com/a/tY1fIuh

    ReplyDelete
  38. Because Scythians were so diverse genetically the question should be asked if all Scythians were indeed Iranians as some, for example Carlos, assume.

    A very interesting article by Constantine Borissoff

    “It is difficult to imagine that the Scythians, which was for the Greek a generic name applied to any ‘barbaric’ (in their opinion) people living north of their confines, had remained a single ethnos speaking a uniform ‘Scythian’ language over the vast territory for a thousand of years.

    In his analysis Abaev completely ignored any possibility that on the vast territory of what the Greeks referred to as ‘Scythia’ could also live numerous IE tribes who never migrated to Iran and, therefore, never returned from there bringing with them the already changed ‘Iranian’ dialects.”

    https://borissoff.wordpress.com/2013/02/12/was-scythian-an-iranian-language/

    ReplyDelete
  39. With those R1bs in mind I wonder how the Scythian Language affinities will look....

    ReplyDelete
  40. These Scythians are from the western edge of the steppe, so probably mostly acculturated Scythians.

    And, in fact, it looks like only three of them belong to R1b, with one being R1b-P312, which might have something to do with the contacts between Hallstatt and steppe peoples.

    ReplyDelete
  41. Article in ‘Science Direct’:
    “The genetic makings of South Asia”by Mait Metspalu, Mayukh Mondal and Gyaneshwer Chaubey.


    “ANI in turn forms when IP admixes with the incoming Middle and Late Bronze Age Steppe (Steppe_MLBA) component, (rather than the Steppe_EMBA groups suggested earlier)”

    https://www.sciencedirect.com/science/article/pii/S0959437X17301752?via%3Dihub

    So they confirm that there was a migration to India from Steppe_MLBA (i.e. Andronovo) and not from Steppe_EMBA (i.e. Yamanaya) as many claim even now. On the maps they show that there was a migration from CWC to Central Asia which participated in the formation of Steppe_MLBA.

    https://i.postimg.cc/yx6Hpy0h/screenshot_443.png

    ReplyDelete
  42. Mikkel Nørtoft said...
    I'm curious as to how the original authors and the YFull people can get Y-haplogroup calls that are so different from each other (even disagreeing on R1b/R1a assignments).
    Did someone make a mistake somewhere, or do you have any other suggestions to explain this difference?

    Is it possible to parse the timeline map in more detail, instead of clumping R1's all together[more colors and or symbols]?
    For example Steppe derived R1a and R1b differ in clades, found in Europe, India, Iran etc....Example
    1- R1b=z2103>z2106...Yamnaya,Afanasievo,Vucedol,Sarmatian,Scythian,Bell Beaker ..etc and L51+


    http://homeland.ku.dk/

    ReplyDelete
  43. Reprocessing the G25 data with these samples and various population averages, hit on one which in which the Baltic BA vs Iberia Chl dimension pops out by PC3: https://imgur.com/a/WaGKk8V

    Still not hugely clear (lots of these samples are so admixed it inevitably ends up looking a bit like a cloud) but hopefully useful.

    Tangentially, it strikes me that these patterns are pretty good supportive evidence for transfer of culture without genes (and maybe language?), though cultural inclusion, and so that makes me question whether we should hold other similar scenarios to the same standard (or accept other geneflow free transfer scenarios *on* the same standard).

    ReplyDelete
  44. @Matt

    I think the most important thing to keep in mind is that these were Iron Age interactions between Iron Age peoples.

    It's easy to fall back into "Bronze Age" thinking where we imagine nomads with metal and horses encountering Neolithic farmers for the first time, but of course the scenario here is quite different. I'm not surprised that cultural diffusion became more important in the Iron Age, since these Iron Age Europeans were much better equipped to resist genetic replacement.

    ReplyDelete
  45. I actually wonder about that, whether material cultural differences were starker between say, the Bronze Age peoples or Iron Age groups.

    On the one hand, there's the scenario as you describe it, but then if I think about it, for instance, though Middle Neolithic cultures of Northern Europe were not metal users, neither were the Corded Ware. And the cultural differentiation between the pastoralist CW and agropastoralist GAC is not clearly greater or lesser than between the Scythians who in classical frame are much more mobile than EBA pastoralists, and various agro-pastoralist societies that would've followed the expansion of Indo-Europeans.

    But even if I'm not sold on thinking that the degree of cultural differentiation in Iron Age of course allows purely cultural dispersals while the degree of cultural differentiation in early Bronze Age of course rejects purely cultural dispersals, I think you're dead on to raise cultural and technological differentiation as part of the question.

    ReplyDelete
  46. a said:
    "Is it possible to parse the timeline map in more detail, instead of clumping R1's all together[more colors and or symbols]?
    For example Steppe derived R1a and R1b differ in clades, found in Europe, India, Iran etc....Example
    1- R1b=z2103>z2106...Yamnaya,Afanasievo,Vucedol,Sarmatian,Scythian,Bell Beaker ..etc and L51+"


    We have thought a lot about that. Since we have sort of run out of contrasting colours, and to help people that are not already into all the subclades understand the map, we have for now only distinguished the steppe-derived R1a from the R1b individuals on the map (red squares and triangles).
    But ”for the nerds” we have also added the layer that highlights the specific subclade SNP numbers on the map, which can then be referenced in the R1 tree which has a link the legend just below the map.
    I think different colours or even more symbols for every subclade would be visually confusing for many people. I already hear from some that it is too detailed as it is now. :)

    Here's a direct link to the R1 tree (which is also visible in the legend). The tree also features the R1 individuals that are outside the timerange of the map.
    But obviously, I now have a problem with displaying the haplogroups from this new Scythian study, since there is apparently not a concensus as to which haplogroups they belong to.

    http://homeland.ku.dk/CSS/images/R1_tree_sign.png

    ReplyDelete
  47. Mikkel Nørtoft said...
    We have thought a lot about that. Since we have sort of run out of contrasting colours, and to help people that are not already into all the subclades understand the map, we have for now only distinguished the steppe-derived R1a from the R1b individuals on the map (red squares and triangles).
    But ”for the nerds” we have also added the layer that highlights the specific subclade SNP numbers on the map, which can then be referenced in the R1 tree which has a link the legend just below the map.
    I think different colours or even more symbols for every subclade would be visually confusing for many people. I already hear from some that it is too detailed as it is now. :)

    Here's a direct link to the R1 tree (which is also visible in the legend). The tree also features the R1 individuals that are outside the timerange of the map.
    But obviously, I now have a problem with displaying the haplogroups from this new Scythian study, since there is apparently not a concensus as to which haplogroups they belong to.

    http://homeland.ku.dk/CSS/images/R1_tree_sign.png


    Maybe in future you can add R1b-Z2103>Z2109+ samples from Sarmatian,Bell Beaker, Scythians to your tree.

    ReplyDelete
  48. a said:
    "Maybe in future you can add R1b-Z2103>Z2109+ samples from Sarmatian,Bell Beaker, Scythians to your tree."

    Yep, I'll do it soon, but probably only the individuals that this article and YFull agree on. :)

    ReplyDelete
  49. Mikkel Nørtoft said...

    Yep, I'll do it soon, but probably only the individuals that this article and YFull agree on. :)

    Thanks. Maybe you can also add this sample to your Y-Rtree[Z2109/8] from Yfull-R-Z2108Z2109/CTS1843> R-KMS67>R-Y20993*while waiting for the Scythian samples to get sorted out.
    R-Y20993*=Sarmatian-id:YF03134 I0575, Pokrovka, EarlySarmatian_IA 5th–2nd c. BC

    ReplyDelete
  50. " Scythians, which was for the Greek a generic name applied to any ‘barbaric’ (in their opinion) "

    Right. I recall that the Greeks believed there are only 2 types of Barbarians in the north: The Celts in the west and the Scythians in the East. No knowledge of Slavs or Germans.

    ReplyDelete

Read the rules before posting.

Comments by people with the nick "Unknown" are no longer allowed.

See also...


New rules for comments

Banned commentators list