K = 13 admixture analysis of Ice Age European genomes

Below is a plot for a K = 13 admixture analysis that includes 37 new, mostly Ice Age genomes from Europe. The following table gives some information for these samples.

Sample           Country  Years BP       Culture
Goyet Q116-1     Belgium  35,160–34,430  Aurignacian
Muierii 2        Romania  33,760–32,840  Unassigned
Paglicci 133     Italy    34,580–31,210  Gravettian
Cioclovina 1     Romania  33,090–31,780  Unassigned
Kostenki 12      Russia   32,990–31,840  Unassigned
Krems WA3        Austria  31,250–30,690  Gravettian
Věstonice 13     Czechia  31,070–30,670  Gravettian
Věstonice 15     Czechia  31,070–30,670  Gravettian
Věstonice 14     Czechia  31,070–30,670  Gravettian
Pavlov 1         Czechia  31,110–29,410  Gravettian
Věstonice 43     Czechia  30,710–29,310  Gravettian
Věstonice 16     Czechia  30,710–29,310  Gravettian
Ostuni 2         Italy    29,310–28,640  Gravettian
Goyet Q53-1      Belgium  28,230–27,720  Gravettian
Paglicci 108     Italy    28,430–27,070  Gravettian
Ostuni 1         Italy    27,810–27,430  Gravettian
Goyet Q376-19    Belgium  27,720–27,310  Gravettian
Goyet Q56-16     Belgium  26,600–26,040  Gravettian
El Mirón         Spain    18,830–18,610  Magdalenian
Afontova Gora 3  Russia   16,930–16,490  Unassigned
Rigney 1         France   15,690–15,240  Magdalenian
Hohle Fels 49    Germany  16,000–14,260  Magdalenian
Goyet Q-2        Belgium  15,230–14,780  Magdalenian
Brillenhöhle     Germany  15,120–14,440  Magdalenian
Hohle Fels 79    Germany  15,070–14,270  Magdalenian
Burkhardtshöhle  Germany  15,080–14,150  Magdalenian
Villabruna       Italy    14,180–13,780  Epigravettian
Rochedane        France   13,090–12,830  Epipaleolithic
Iboussières 39   France   12,040–11,410  Epipaleolithic
Continenza       Italy    11,200–10,510  Mesolithic
Ranchot 88       France   10,240–9,930   Mesolithic
Les Closeaux 13  France   10,240–9,560   Mesolithic
Falkenstein      Germany  9,410–8,990    Mesolithic
Bockstein        Germany  8,370–8,160    Mesolithic
Ofnet            Germany  8,430–8,060    Mesolithic
Chaudardes 1     France   8,360–8,050    Mesolithic
Berry-au-Bac     France   7,320–7,170    Mesolithic

The plot shows that these samples can be divided into three groups. The first includes the Aurignacian, Gravettian, and other early samples. These have some of the same dark blue component found in the later European hunter-gatherers, but they also show other components, which demonstrates their more distant relationship to later Europeans. The Magdalenian samples make up the second group. They’re very similar to the later hunter-gatherers, but they still show small amounts of non-European components. The third group includes the most recent hunter-gatherers, who lived during the last few thousand years of the Ice Age and the post-Ice-Age Mesolithic. They all show only the dark blue component.

A striking feature of the plot, which has appeared in my previous plots, and which cries out for an explanation, is the significant amount of the pine green component, which peaks in Georgians, throughout all of the modern European samples. This component does not appear in the hunter-gatherers, or in the early Neolithic farmers, and it is also almost completely absent in the Pit Grave samples from the Pontic-Caspian steppe, and the Corded Ware and Timber Grave samples. It does appear in the Chinchorro mummy sample from around 4000 BC. It also appears in the pre-Bell-Beaker Copper Age samples from El Portalón cave in Spain, and especially in El Portalón 3, which I discovered to belong to Y haplogroup R1b-M269. Some of the pine green component shows up in the Bell Beaker and Unetice samples of Central Europe, and in the Nordic and Irish Bronze Age samples. My best explanation for the significant presence of this component in these samples is that it is somehow linked with the spread of R1b-L51.


Posted in Uncategorized
13 comments on “K = 13 admixture analysis of Ice Age European genomes
  1. norah4you says:

    Reblogged this on Norah4you's Weblog and commented:
    A special thanks to Genetiker who allows me to reblogg this

  2. SF says:

    Do you have an email to be contacted at?

  3. vsva says:

    G2a-z6030 and z39318 are some minority P303 branches? I saw some Sardinians. what about les chaudardes, Berry au Bac, Burkhardtshohle, Hohefels49 and rest El Mirador samples y snp calls? Thank you.

  4. vsva says:

    Can i ask about some other samples y snp calls? What about one Esperstedt sample i0807(not i0172 I2a1B1 l161), one Els Trocs i0411 listed only as F, i0551 Salzmunde and Barcin samples i1099, i1103, i1583, i0723 and i0727.

  5. Antoine says:

    Great work,…Do you have an Excel spreadsheet with the percentages?

  6. Random_dude06 says:


    Isn’t the “pine green component” simply the signal of the “Caucasian Hunter Gatherers”– ie, Kotias/Satsurblia?

    I also notice the K13 uses a neutral-hued green to group the South Asians…. in most calcs, South-Central Asians like Brahui and Punjabi score very highly (40-50%) in the “pine green” (CHG) component, but only about 20-30% “Ancestral South Indian” (ie, Onge). Is there a reason why ASI and CHG heritage is non-discrete and lumped under one “neutral” green color for some of the South Asians?

    • Genetiker says:

      The pine green component is CHG-related, but it’s not a true CHG component. If it was, then the two CHG samples would be 100% pine green, but instead they’re a mix of medium blue, pine green, and plain green, which is a reflection of the fact that those three components include some CHG DNA. If the dataset included a few more CHG samples, then the analysis would probably produce a true CHG component.

      The plain green component is a Veddoid-Caucasoid hybrid component, with the Caucasoid part being part CHG. If the dataset included Indian tribal groups like the Paniya and Pulayar, then the analysis would produce a more pure Veddoid component.

      The pine green component is informative, because it shows that all modern Europeans have a significant amount of Caucasus-related admixture beyond what was present in the Pit Grave and Corded Ware people. As I said above, this extra Caucasus-related admixture seems to be due at least in part to the spread of R1b-L51 people.

  7. Genetiker says:

    Here is a plot for another K = 13 analysis. The component colored pine green in this plot peaks in Bedouins, not Georgians. This analysis didn’t produce a component that peaks in the Caucasus region. In the above plot, Europeans have large amounts of the three blue components and the pine green component, but in this second plot most Europeans only have large amounts of the three blue components.

  8. Genetiker says:

    Here is a plot for a K = 14 analysis in which I included two copies of each of the two Caucasus hunter-gatherers so that the analysis would produce a true CHG component. The European samples show a lot less of the pine green component in this new plot than they do in the plot above, although it is still consistently present in significant amounts.

  9. Gene says:

    Any idea how those australoid, south asian, papuan, east asian, siberian and amerindian components present in paleolithic europeans could disappear in mesolithic europeans. Could it be through genetic drift or negetive selection and how did they got in there in the first place? Shared ancestry or leftover alleles from before the west and east eurasian split?
    Siberian and amerindian components are strange because Paleoeuropeans aren’t suppesed to have any ANE ancestry yet, unlike mesolithic and neolithic ones.
    What is the light blue component that peaks in Mozabites and Bedouins and is also present in Cioclovina and Paglicci in large quantities. Does it suggest basal eurasian admixture in paleolithic europeans?
    Also what is that blue component that peaks in Pit Grave CA and how it is different from the dark blue WHG component? It might be WHG+ANE or EHG+ANE but MA1 has both here.

  10. Gene says:

    By the way could you run more tests on especially paleolithic genomes and rarely tested mesolithics like Finkelsta, . Could be with Gedmatch calculators or some kind of software like ADMIXTURE, STRUCTURE and fineStructure. I think you missed a few important samples like RaRochette, Linatzeta, La Pasiega and La Chora, also more Vestonice samples would be good. Also Oase1/2, Kostenki12/14, Tianyuan and Ust’-Ishim.

    I know you once tested Oase1, K14 and Tianyuan but more thorough tests would be useful.
    Nice list https://docs.google.com/spreadsheets/d/1lMeojSM7Lrep4GYi8qV31QbDPIw5w-bEygm8esoSHF4/edit#gid=991990924

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: