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

The recent increase in openly available ancient human DNA samples allows for large-scale meta-analysis applications. Trans-generational past human mobility is one of the key aspects that ancient genomics can contribute to since changes in genetic ancestry—unlike cultural changes seen in the archaeological record—necessarily reflect movements of people. Here, we present an algorithm for spatiotemporal mapping of genetic profiles, which allow for direct estimates of past human mobility from large ancient genomic datasets. The key idea of the method is to derive a spatial probability surface of genetic similarity for each individual in its respective past. This is achieved by first creating an interpolated ancestry field through space and time based on multivariate statistics and Gaussian process regression and then using this field to map the ancient individuals into space according to their genetic profile. We apply this algorithm to a dataset of 3138 aDNA samples with genome-wide data from Western Eurasia in the last 10,000 y. Finally, we condense this sample-wise record with a simple summary statistic into a diachronic measure of mobility for subregions in Western, Central, and Southern Europe. For regions and periods with sufficient data coverage, our similarity surfaces and mobility estimates show general concordance with previous results and provide a meta-perspective of genetic changes and human mobility.

All human behavior is spatial behavior, and spatial perception and interaction are deeply rooted in the human mind. Understanding movements in space—mobility—on different orders of magnitude is therefore a major component for understanding human behavior throughout history (1), from the Iceman’s quest through the Ötztal Alps, to the Viking expansion even beyond Medieval Europe, and maybe eventually humankind’s journey to the stars.Anthropological theory provides different concepts and categories to classify mobility. Mobility can be permanent or cyclical, a group property or individual behavior, and finally motivated by economic, social, or cultural incentives. It has complex implications for the formation, perception, and interaction of identity (2–4). Migration is an especially challenging and controversial topic (5, 6) as it is notoriously difficult to prove and to uncover its causes among the interdependencies of microprocesses and macroprocesses (7). Narratives of migration are particularly vulnerable to political instrumentalization (8).The field of archaeogenetics now provides a perspective on mobility, which is at its very core influenced by population genetics theory. The emergence, change, and distribution of human ancestry components—mediated by the mobility of their hosts—are in fact some of its most important research questions (e.g., refs. 9–11), causing fruitful and corrective friction with the humanities (12–14). While so far much archaeogenetic research focuses on particular cultural–historical contexts, the recent growth of published ancient DNA samples from all around the world enables a unique category of quantitative meta-analysis.Large, explicitly spatiotemporal datasets have been part of population genetics research for a long time already (15), sometimes even with a focus on mobility quantification (16–19). But to our knowledge, only few attempts have been made to systematically derive a continuous, large-scale and diachronic measure of human mobility with ancient genetic data. These are most notably a pioneering publication by Loog et al. (20) and another approach by Racimo et al. (21). Loog et al. measure mobility in prehistoric Europe by comparing the distance matrix correlation among spatial, temporal, and genetic distance for aDNA samples in moving 4,000-y windows. As a result, they generate an unscaled mobility proxy curve that indicates elevated levels of mobility correlating with the Neolithic expansion, the Steppe migration, and, finally, the European Iron Age. Racimo et al., on the other hand, employ admixture analysis to model the dynamics of specific ancestry components through time: Mesolithic hunter-gatherers, Neolithic farmers with ancestry originating in the Near East, and Yamnaya steppe herders, arriving in Europe during the third millennium BC. They derive mobility as a wave front speed of surpassed ancestry component thresholds. To overcome sample sparsity and to correlate the arrival of certain ancestry components with biogeographic metrics, they use Gaussian process regression for the interpolation of relative ancestry component occurrence—an idea we also took as a starting point for our proposed mobility estimation method.In this paper, we present an algorithm to estimate past human mobility on the individual level. For each individual, we determine a probability distribution in space, which yields locations of likely genetic similarity to the sample in question. We call this the similarity probability surface, which, as we show, is generally informative on where an individual’s ancestors might have lived. The distance between the location where an individual was buried and a point of maximum likelihood in the similarity surface serves us as a simplified proxy for personal mobility in an individual’s (or their ancestors’) lifetime. We apply this algorithm to several thousand previously published ancient genomes from Western Eurasia dating from between 8000 BC and 2000 AD (excluding modern genomes) taken from the Allen Ancient DNA Resource (AADR) (22). And, we show that, while the average results largely match expectations including known and large-scale movements at the beginning and end of the Neolithic, these large-scale patterns are accompanied by considerable individual-level heterogeneity.