Estimating human mobility in Holocene Western Eurasia with large-scale ancient genomic data

Abstract

Ancient human DNA (aDNA) extracted from archaeological contexts allows reconstructing past population movements. Previous methods work by calculating proportions of shared ancestry among individuals or groups in order to answer specific, regional research questions. Here, we propose a large-scale algorithm to quantify human mobility through time and space using bulk aDNA data. The algorithm has two core components: i) interpolation of the spatiotemporal distribution of genetic ancestry to obtain a continuous ancestry information field and ii) probabilistic estimation of a spatial genetic similarity surface for each input sample by projecting its ancestry profile into this field. We apply this to thousands of published genomic samples in the last 10,000 y to trace diachronic mobility patterns in Western Eurasia. Keywords: aDNA, prehistory, mobility estimation, Gaussian process regression