Allentoft and Iron Age Sweden

[Anglesqueville]

I started using data from the big Allentoft study. For me, they are of major interest in having been calculated (by Glimpse) by professionals. I have in fact always had reservations about my own imputations (by Beagle) for data with very low coverage. This is particularly the case for the Finnish Iron Age micro-group (Levänluhta). In addition, Allentoft has given us, among other things, a gift of valuable new data, in particular from the Swedish Falköping group (between 2000 BC and 1700 BC). This group is of major importance for at least three reasons:
1) It contains the first representatives of I-M253 which have a reasonable chance of corresponding to the roots of this haplogroup as we know it today (TMRCA and location).
2) Falköping is located in the heart of the Swedish corridor to which I have often alluded on the late proto-Germanic thread. This corridor, which links the Lake Mälaren region and the southwest coast of Sweden, is the area where most of the early iron industries were distributed, as well as the finds of Mälar axes.
3) Isotopic analyzes attest that most of the individuals from Falköping are not native to this area, but to regions located in the northeast of Sweden. (see: Open Archeology 2018; 4:1–35 Malou Blank*, Anna Tornberg, and Corina Knipper "New Perspectives on the Late Neolithic of South-Western Sweden. An Interdisciplinary Investigation of the Gallery Grave Falköping Stad 5").
I first carried out a PCA (by Plink2) by projecting onto modern populations the groups Falköping (Sweden_N), Sweden_IronAge (which includes the outlier individual of Levänluhta), Sweden-BattleAxes, and finally the micro-group of Levänluhta (which merges with that of the modern Saamis). This PCA reveals an indisputable cline, with an obvious shift of the Sweden_N group relative to the Falköping group in the direction of Levänluhta.



I obviously wondered whether this shift was verifiable by qpAdm. I didn't have much hope, remembering the sad experiences I had previously had with Levänluhta. My pessimism was unwarranted, because what qpAdm gives me is nothing less than the best model I have ever gotten for Iron Age Swedes. This model absolutely converges with Allentoft's conclusions regarding the leading role of the Falköping group. As for the Iron Age, it undoubtedly points to a genetic influx from Finland.

left pops:
Sweden_IA.imputed_allentoft
Sweden_N_Falköping.imputed_Allentoft
Finland_IA.imputed_allentoft

right pops:
Russia_Ust_Ishim.DG
Cameroon_SMA.DG
Italy_North_Villabruna_HG
Czech_Vestonice16
Belgium_UP_GoyetQ116_1
Russia_MA1_HG.SG
Iran_GanjDareh_N
Kazakhstan_Botai_Eneolithic.SG
Russia_Kostenki14.SG
Ukraine_Mesolithic
Indian_GreatAndaman_100BP.SG
Russia_Arkhangelsk_Veretye_Mesolithic.SG
Israel_PPNB
Georgia_Kotias.SG
Turkey_N_I0707



best coefficients:    0.919    0.081
totmean:      0.919    0.081
boot mean:    0.919    0.081
      std. errors:    0.033    0.033

    fixed pat  wt  dof    chisq      tail prob
          00  0    13    6.435        0.928885    0.919    0.081
          01  1    14    12.129        0.595957    1.000    0.000
          10  1    14  318.653              0    0.000    1.000
best pat:          00        0.928885              -  -
best pat:          01        0.595957  chi(nested):    5.694 p-value for nested model:      0.0170234



Samples:

NEO221_A U Sweden_N_Falköping.imputed_Allentoft
NEO223_A U Sweden_N_Falköping.imputed_Allentoft
NEO224_A U Sweden_N_Falköping.imputed_Allentoft
NEO225_A U Sweden_N_Falköping.imputed_Allentoft
NEO226_A U Sweden_N_Falköping.imputed_Allentoft
NEO227_A U Sweden_N_Falköping.imputed_Allentoft
NEO228_A U Sweden_N_Falköping.imputed_Allentoft
NEO259_A U Sweden_N_Falköping.imputed_Allentoft


RISE174_A U    Sweden_IA.imputed_allentoft
VK522_A U    Sweden_IA.imputed_allentoft
VK523_A U    Sweden_IA.imputed_allentoft

DA234_A U    Finland_IA.imputed_allentoft
DA238_A U    Finland_IA.imputed_allentoft

Here for once is a model of easy interpretation: the p-value of the alternative model is too low for this alternative model to be admissible. The intervention of the Finnish group is clearly essential. The monumental tail-prob of the complete model left little doubt as to this in any case.