<i>Genealogical inferences based on comparison of modern and ancient DNA</i>

Abstract

The study of genetic variation within and between populations can help us understand aspects of human demographic history over the past thousands of years, i.e. well beyond the timescales of historical evidence. Demographic and evolutionary dynamics influence the distribution of the observed genetic diversity, and so one can retrospectively reconstruct episodes in population history on the basis of genetic diversity data. One way to do this is to make extensive use of simulations, considering evolution as a stochastic process in which the genetic data are modeled as random variables. The simulation of genetic data under various scenarios allows one to explore how demographic and evolutionary parameters can affect genetic variation, also making it possible to approximately estimate the historical parameters that produced the observed data. To this aim, many statistical approaches have been developed, but, when models are complex or datasets are large, they often become computationally expensive, or analytically intractable. Approximate Bayesian Computation (ABC) methods overcome these problems allowing, for the first time, to analyze large datasets and to interpret them in the light of realistic (i.e. complex) models, thus enabling the probabilistic comparison among different models of evolution, the simultaneous estimation of demographic and evolutionary parameters, and the quantitative evaluation of the results credibility. In this context, we analyzed datasets of modern and ancient genetic variation in order to understand the demographic histories of these populations, to highlight traces of past genetic variation in modern populations, and to evaluate whether, and to what extent, ancient and modern populations that have lived in the same place in different period of times can be considered genealogically related. We tried to address three anthropological questions, namely the interaction of anatomically modern humans with archaic forms (i.e. Neandertals in Europe), evidence for genealogical continuity in Sardinia since the Bronze-age, and the origins and evolution of the Etruscan population. Within the ABC framework, in each of the three studies, we explicitly compared several models, differing for the demographic processes and the genealogical relationship among population, to identify the model best accounting for the observed variation, and to estimate its demographic and evolutionary parameters. This way, it has been possible to shed light on past population history and to address questions about the nature and the extent of genealogical links between modern and ancient populations, clarifying aspects of human history that have long been controversial in population genetics and evolutionary biology.