CONTRIBUTION A LA CARTOGRAPHIE, PAR ETUDES DE LIAISON ET D’ASSOCIATION, DE LOCI D’INTERET CHEZ LES ANIMAUX DOMESTIQUES. | CONTRIBUTION TO LINKAGE AND ASSOCIATION MAPPING OF TRAIT LOCI IN LIVESTOCK.

Until recently, breeding values were estimated based on phenotypes measured on the individual and its relatives, and the notion that the covariance between breeding values is proportionate to the kinship coefficient. Advances in genomics now allow for direct analysis of the genome and identification of the loci that determine the breeding values of individuals. As a consequence, marker assisted selection and genomic selection have become more effective and are replacing conventional selection. The identification of loci influencing the traits of interest requires the use of advanced statistical methods that are constantly evolving. In the context of this thesis, we have (i) contributed to the development of gene mapping methods, (ii) applied these methods to map loci influencing both metric and meristic traits, and (iii) contributed to the development of methods for the integration of genomic information in livestock breeding and management. The mapping methods that we have helped developing distinguish themselves mainly by the fact that (i) they exploit haplotype information (by means of a hidden markov model) which should increase the linkage disequilibrium with causative variants and hence detection power, (ii) they can simultaneously extract linkage information within families, and linkage disequilibrium information across the population, and (iii) they correct for population stratification by means of a random polygenic effect, and (iv) they can be applied to binary as well as quantitative traits. We have applied these and other methods to map loci influencing (i) quantitative hematological parameters in a porcine line-cross, and (ii) binary traits including diseases in bovine and non-syntenic Copy Number Variants in cattle, horse and human. In fine, we have contributed to the development of methods for the utilization of marker information in animal selection and production. We have extended the haplotype-based mapping method to allow imputation and have evaluated the utility of this approach in scenarios mimicking reality. We have also contributed to the development of a method to quantify somatic cell counts in the milk of individual cows by genotyping a sample of milk from the farm’s tank (hence a mixture of milk from all cows on the farm) Our work has resulted in the development of a software package (“GLASCOW”) that is increasingly used by the community to map genes influencing complex traits, primarily binary. By using this tool, we have contributed to the localization of several trait loci in pig, cattle, horse and human. We have contributed to the development of approaches that reduce the costs of genomic analyses in livestock by, on the one hand, complementing real SNP genotypes with genotypes obtained in silico by means imputation, and, on the other hand, by developing a method to deconvolute genotypes obtained on DNA pools.