Abstract
Genomes are affected by conflicting selective regimes. This is particularly well illustrated by the concept of semi-permeable barriers to gene flow, as found in the hybrid zones literature. Some genes contribute to the prevention of mixing between differentiated genetic lineages, either because they are involved in adaptation to local environmental conditions, or because they are incompatible with alleles from other genetic lineages. Other parts of the genome are either neutral, or subjected to selection which tends to homogenize the genetic lineages. In the first part of this thesis, models of the evolution of reproductive isolation are presented to explain the isolation patterns observed in experimental hybridizing crosses between incipient species. Using standard models of Dobzhansky-Muller genetic incompatibilities, it is shown that the asymmetry and complexity of incompatibilities are not well explained by there being an “evolutionary sieve”, i.e. a different rate of accumulation between incompatibilities. A complementary approach to quantitative modeling (an extension of Fisher's Geometric Model) then clarifies which conditions of divergence between allopatric lines led to highly deleterious effects in hybrid genotypes. The relative importance of mean levels of fitness epistasis, the distribution of mutation sizes, and the way lineages adapt to new environmental conditions is discussed. The second part of this thesis takes advantage of technical advances in genomics to study the history of speciation and adaptation in a non-model species complex, Mytilus mussels. A statistical method of inferring speciation scenarios is presented. Results show that European mussels experienced a complex history of strict divergence followed by a period of periodic connectivity. In agreement with the concept of semi-permeable barriers to gene flow, it is shown that introgression rates are heterogeneous along the genome. Next, genome scans of differentiation were conducted between pairs of populations of the species complex. The analysis of genetic variation and allele genealogies on a small chromosomal scale allowed to reconstruct the evolutionary history of more than 1000 genomic regions. This analysis reveals that a major cause of intraspecific differentiation is the differential introgression of foreign alleles. Overall, this thesis shows not only that biogeography of speciation, i.e. the temporal and spatial patterns of gene flow, play a major role in our understanding of existing biodiversity, but also its amazing complexity and extent of its impact on genome evolution.