Massive migration from the steppe is a source for Indo-European languages in Europe

Abstract

In this section we devised methods for detecting admixture and estimating its proportions without detailed phylogenetic modeling. Our assumptions were that a Test population is formed as a mixture of an arbitrary number of other reference populations that are differentially related to a set of outgroup populations. This approach allowed us to study admixture in ancient Europe without using a tree or admixture graph model. We were able to show that Middle Neolithic Europeans had experienced <30% resurgence of western European hunter-gatherer ancestry while the Corded Ware (the earliest of the Late Neolithic populations) had experienced >1/3 gene flow from a population related to eastern European hunter-gatherers. Furthermore, we showed that the Yamnaya population from the early Bronze Age in the Russian steppe may have been related to the population that admixed into Late Neolithic central Europeans, as it was not of purely eastern European huntergatherer descent but had additional ancestry of a farmer-related population that split off before the differentiation of European farmers, perhaps from the Caucasus or Near East, although more work is needed to identify the source of this population. Taking the Yamnaya as a source of the eastern admixing population, the level of population replacement between the Middle Neolithic and the Corded Ware becomes ~3/4, suggesting a massive influence of eastern populations in the formation of the Late Neolithic in central Europe. This influence is evident in other Late Neolithic populations (at a lower level), as well as in present-day Europeans, who differ from the Late Neolithic central Europeans by having additional farmer or west European hunter-gatherer ancestry. The robustness of our inference is supported by the fact that by adding a different set of outgroups from ancient Europe, the ~3/4 estimate was little affected. Further research of populations from central Europe, Eastern Europe, and the Caucasus may reveal the formation of steppe groups that succeeded the eastern European hunter-gatherers and how they affected (and were affected by) the populations surrounding them. The fact that sizeable admixture accompanied major transitions in European prehistory in at least three locations (Spain, Germany, and European Russia) suggests that the formation of presentday European populations was effected by sizeable local population replacement. Archaeological models of the past must accommodate the possibility of migration and admixture at a massive scale long after the Neolithic transition in Europe. 123 Figure S9.27: Adding BedouinB as a 5th reference population improves residuals for Maltese, Sicilians, Ashkenazi Jews, and Spanish. 124 Table S9.3: Populations that improve resnorm for European outlier populations when added to a model of EN/WHG/EHG admixture as a 4th ancestral population. We show the top 20 populations that reduce resnorm the most when added to the mixture model. For Maltese, Sicilians, and Ashkenazi Jews these populations tend to be from the Middle East and North Africa. For Finns, Russians, and Mordovians they tend to be populations from Central Asia, Siberia, and Turkic/Uralic populations from Western Eurasia. Maltese Sicilian Ashkenazi_Jew Finnish Russian Mordovian 4th anc. pop. resnorm 4th anc. pop. resnorm 4th anc. pop. resnorm 4th anc. pop. resnorm 4th anc. pop. resnorm 4th anc. pop. resnorm Moroccan_Jew 0.000006 Turkish_Jew 0.000006 Cypriot 0.000006 Chuvash 0.000013 Chuvash 0.000016 Turkmen 0.000011 Lebanese 0.000010 Cypriot 0.000006 Iraqi_Jew 0.000018 Mansi 0.000020 Turkmen 0.000020 Uzbek 0.000012 Syrian 0.000010 Moroccan_Jew 0.000009 Turkish_Jew 0.000018 Uzbek 0.000020 Uzbek 0.000021 Nogai 0.000012 Tunisian_Jew 0.000011 Druze 0.000012 Moroccan_Jew 0.000020 Even 0.000021 Nogai 0.000022 Chuvash 0.000014 Saudi 0.000014 Iraqi_Jew 0.000012 Druze 0.000025 Selkup 0.000021 Altaian 0.000025 Altaian 0.000017 Turkish_Jew 0.000016 Syrian 0.000014 Lebanese 0.000027 Dolgan 0.000021 Mansi 0.000025 Tubalar 0.000017 Libyan_Jew 0.000017 Lebanese 0.000016 Syrian 0.000028 Turkmen 0.000022 Tubalar 0.000025 Hazara 0.000017 Jordanian 0.000017 Tunisian_Jew 0.000019 Tunisian_Jew 0.000037 Nogai 0.000022 Selkup 0.000026 Kyrgyz 0.000019 Palestinian 0.000019 Saudi 0.000021 Saudi 0.000039 Nganasan 0.000022 Even 0.000026 Tuvinian 0.000019 Druze 0.000022 Jordanian 0.000022 Iranian_Jew 0.000039 Yakut 0.000022 Yakut 0.000026 Uygur 0.000019 Yemenite_Jew 0.000022 Libyan_Jew 0.000025 Jordanian 0.000039 Altaian 0.000023 Dolgan 0.000026 Yakut 0.000019 BedouinB 0.000022 Palestinian 0.000026 Palestinian 0.000044 Tubalar 0.000023 Tuvinian 0.000027 Oroqen 0.000020 BedouinA 0.000023 BedouinB 0.000027 Libyan_Jew 0.000044 Tuvinian 0.000024 Kyrgyz 0.000029 Kalmyk 0.000020 Tunisian 0.000024 Yemenite_Jew 0.000028 Yemenite_Jew 0.000045 Yukagir 0.000025 Hazara 0.000029 Even 0.000020 Mozabite 0.000024 Tunisian 0.000028 BedouinB 0.000045 Oroqen 0.000026 Nganasan 0.000029 Dolgan 0.000021 Algerian 0.000025 Mozabite 0.000028 BedouinA 0.000046 Kyrgyz 0.000026 Oroqen 0.000029 Selkup 0.000021 Egyptian 0.000025 BedouinA 0.000028 Tunisian 0.000046 Kalmyk 0.000027 Yukagir 0.000030 Hezhen 0.000021 Saharawi 0.000026 Egyptian 0.000028 Egyptian 0.000046 Hazara 0.000029 Kalmyk 0.000030 Daur 0.000021 Yemen 0.000028 Algerian 0.000029 Mozabite 0.000047 Hezhen 0.000030 Uygur 0.000032 Mansi 0.000022 Esan 0.000030 Saharawi 0.000029 Algerian 0.000047 Aleut 0.000031 Hezhen 0.000032 Xibo 0.000023 125