Human genetic differentiation across the Strait of Gibraltar

Mathias Currat,Estella S Poloni,Alicia Sanchez-Mazas

doi:10.1186/1471-2148-10-237

Abstract

BackgroundThe Strait of Gibraltar is a crucial area in the settlement history of modern humans because it represents a possible connection between Africa and Europe. So far, genetic data were inconclusive about the fact that this strait constitutes a barrier to gene flow, as previous results were highly variable depending on the genetic locus studied. The present study evaluates the impact of the Gibraltar region in reducing gene flow between populations from North-Western Africa and South-Western Europe, by comparing formally various genetic loci. First, we compute several statistics of population differentiation. Then, we use an original simulation approach in order to infer the most probable evolutionary scenario for the settlement of the area, taking into account the effects of both demography and natural selection at some loci.ResultsWe show that the genetic patterns observed today in the region of the Strait of Gibraltar may reflect an ancient population genetic structure which has not been completely erased by more recent events such as Neolithic migrations. Moreover, the differences observed among the loci (i.e. a strong genetic boundary revealed by the Y-chromosome polymorphism and, at the other extreme, no genetic differentiation revealed by HLA-DRB1 variation) across the strait suggest specific evolutionary histories like sex-mediated migration and natural selection. By considering a model of balancing selection for HLA-DRB1, we here estimate a coefficient of selection of 2.2% for this locus (although weaker in Europe than in Africa), which is in line with what was estimated from synonymous versus non-synonymous substitution rates. Selection at this marker thus appears strong enough to leave a signature not only at the DNA level, but also at the population level where drift and migration processes were certainly relevant.ConclusionsOur multi-loci approach using both descriptive analyses and Bayesian inferences lead to better characterize the role of the Strait of Gibraltar in the evolution of modern humans. We show that gene flow across the Strait of Gibraltar occurred at relatively high rates since pre-Neolithic times and that natural selection and sex-bias migrations distorted the demographic signal at some specific loci of our genome.

Highlights

The Strait of Gibraltar is a crucial area in the settlement history of modern humans because it represents a possible connection between Africa and Europe
Such effects may be confounding in a number of cases: directional selection generally leads to a loss of genetic diversity within populations, which can hardly be distinguished from an effect of rapid genetic drift; balancing selection maintains genetic variation, which is expected in case of intensive gene flow
Recent studies on the genetic history of human populations focus on the analysis of non-coding (e.g. Short Tandem Repeats (STRs) [5]) or genome-wide (e.g. SNPs [6,7]) markers, allowing to get rid of selective effects acting on individual loci, data on classical and molecular polymorphisms related to either coding or specific DNA regions have been widely used to reconstruct human history and need a specific attention

Summary

Introduction

Geneticists are often faced with the acute problem of disentangling the effects of natural selection and demographic history on the evolution of different polymorphisms [1,2] Such effects may be confounding in a number of cases: directional selection generally leads to a loss of genetic diversity within populations, which can hardly be distinguished from an effect of rapid genetic drift; balancing selection maintains genetic variation, which is expected in case of intensive gene flow. Other hypotheses have been proposed like frequency-dependent selection conferring selective advantage to rare alleles to which pathogens would not have had time to adapt, as well as fluctuating selection depending on environmental changes over time and space (see [15], for a review) These different forms of balancing selection explain why an excess of nonsynonymous compared to synonymous substitutions are found at the peptide-binding regions of the HLA molecules

Methods

Results

Conclusion