The first horse herders and the impact of early Bronze Age steppe expansions into Asia

Abstract

The Yamnaya expansions from the western steppe into Europe and Asia during the Early Bronze Age (~3000 BCE) are believed to have brought with them Indo-European languages and possibly horse husbandry. We analyzed 74 ancient whole-genome sequences from across Inner Asia and Anatolia and show that the Botai people associated with the earliest horse husbandry derived from a hunter-gatherer population deeply diverged from the Yamnaya. Our results also suggest distinct migrations bringing West Eurasian ancestry into South Asia before and after, but not at the time of, Yamnaya culture. We find no evidence of steppe ancestry in Bronze Age Anatolia from when Indo-European languages are attested there. Thus, in contrast to Europe, Early Bronze Age Yamnaya-related migrations had limited direct genetic impact in Asia.

Figure Caption

Model-based admixture proportions for selected ancient and present-day individuals, assuming k=6, shown with their corresponding geographical locations. Ancient groups are represented by larger admixture plots with those sequenced in the present work surrounded by black borders, and others used for providing context with blue borders. Present-day South Asian groups are represented by smaller admixture plots with dark grey borders.

Fig. 1

Geographic location and dates of ancient samples. A) Location of the 74 samples from the steppe, Lake Baikal region, Turkmenistan, and Anatolia analyzed in the present study. MA1, KK1, and Xiongnu_IA were previously published. Geographical background colors indicate the western steppe (pink), central steppe (orange) and eastern steppe (gray). B) Timeline in years before present (BP) for each sample. ML – Mesolithic, EHG – Eastern hunter-gatherer, EN – Early Neolithic, LN – Late Neolithic, CA – Copper Age, EBA – Early Bronze Age, EMBA – Early/Middle Bronze Age, MLBA – Middle/Late Bronze Age, IA – Iron Age.

Fig. 2

Principal component analyses using ancient and present-day genetic data. A) PCA of ancient and modern Eurasian populations. The ancient steppe ancestry cline from EHG to Baikal_EN is visible at the top outside present-day variation, while the YamnayaKaragash_EBA sample has additional CHG ancestry and locates to the left with other Yamnaya and Afanasievo samples. Additionally, a shift in ancestry is observed between the Baikal_EN and Baikal_LNBA, consistent with an increase in ANE-related ancestry in Baikal_LNBA. B) PCA estimated with a subset of Eurasian ancient individuals from the steppe, Iran, and Anatolia as well as present-day South Asian populations. PC1 and PC2 broadly reflect West-East and North-South geography, respectively. Multiple clines of different ancestry are seen in the South Asians, with a prominent cline even within Dravidians in the direction of the Namazga_CA group, which is positioned above Iranian Neolithic in the direction of EHG. In the later Turkmenistan_IA sample, this shift is more pronounced and towards Steppe EBA and MLBA. The Anatolia_CA, EBA and MLBA samples are all between Anatolia Neolithic and CHG, not in the direction of steppe samples.

Fig. 3

Model-based clustering analysis of present-day and ancient individuals assuming K = 6 ancestral components. The main ancestry components at K = 6 correlate well with CHG (turquoise), a major component of Iran_N, Namazga_CA and South Asian clines; EHG (pale blue), a component of the steppe cline and present in South Asia; East Asia (yellow ochre), the other component of the steppe cline also in Tibeto-Burman South Asian populations; South Indian (pink), a core component of South Asian populations; Anatolian_N (purple), an important component of Anatolian Bronze Age and Steppe_MLBA; Onge (dark pink) forms its own component.

Fig. 4

Demographic model of 10 populations inferred by maximizing the likelihood of the site frequency spectrum (implemented in momi).We used 300 parametric bootstrap simulations (shown in gray transparency) to estimate uncertainty. Bootstrap estimates for the bias and standard deviation of admixture proportions are listed beneath their point estimates. Note that the uncertainty may be underestimated here, due to simplifications or additional uncertainty in the model specification.

Fig. 5

Y-chromosome and mitochondrial lineages identified in ancient and present-day individuals. A) Maximum likelihood Y-chromosome phylogenetic tree estimated with data from 109 high coverage samples. Dashed lines represent the upper bound for the inclusion of 42 low coverage ancient samples in specific Y-chromosome clades on the basis of the lineages identified. B) Maximum likelihood mitochondrial phylogenetic tree estimated with 182 present-day and ancient individuals. The phylogenies displayed were restricted to a subset of clades relevant to the present work. Columns represent archaeological groups analyzed in the present study, ordered by time, and colored areas indicate membership of the major Y-chromosome and mtDNA haplogroups.

Fig. 6

A summary of the four qpAdm models fitted for South Asian populations. For each modern South Asian population, we fit different models with qpAdm to explain their ancestry composition using ancient groups and present the first model that we could not reject in the following priority order: 1. Namazga_CA + Onge, 2. Namazga_CA + Onge + Late Bronze Age Steppe, 3. Namazga_CA + Onge + Xiongnu_IA (East Asian proxy), and 4. Turkmenistan_IA + Xiongnu_IA. Xiongnu_IA were used here to represent East Asian ancestry. We observe that while South Asian Dravidian speakers can be modelled as a mixture of Onge and Namazga_CA, an additional source related to Late Bronze Age steppe groups is required for IE speakers. In Tibeto-Burman and Austro-Asiatic speakers, an East Asian rather than a Steppe_MLBA source is required.