نبذة مختصرة : Understanding the reciprocal influence between genotype and phenotype has been a long-standing goal of modern biology. Many aspects of evolution at the molecular level are well known to respond to demographic or life history characteristics of species. In particular, the nearly-neutral theory postulates that small populations accumulate a heavier load of slightly deleterious substitutions in their genome as a result of increased genetic drift. Base composition has also been shown to reflect the influence of macroscopic parameters through the mechanism of GC-biased gene conversion. However, the development and empirical validation of these theories are mostly based on a restricted diversity of organisms, in which mammals stand as a major contributor. In this thesis, using a comparative approach and tens of transcriptomes, we aimed at extending to Amniota our understanding of the determinants of molecular evolutionary patterns. With the incorporation of all clades of reptiles, we confirmed the major role of the effective population size on species ability to purge deleterious amino-acid changes, while revealing a paradoxical response of the dN/dS ratio in birds, raising a stimulating enigma. The biased gene conversion also emerged as the main driver of coding sequence GC content in vertebrates, including reptiles and fishes, whose genomic homogeneity had kept its signal hidden for long. In parallel, the relations between life-history traits and molecular parameters have enabled us to investigate and make progress in the field of ancestral body mass reconstruction. We focused on the Cetartiodactyla order, a group which is mainly characterized by large extant species (such as camel, giraffe or whales). The combined analysis of the yet untested mitochondrial marker and nuclear genes, including 21 newly sequenced transcriptomes, testified in favor of the singular result of a small cetartiodactyl ancestor, in agreement with the palaeontological record, demonstrating the strong potential of DNA sequences to reveal the ...
Processing Request
No Comments.