Classical scrapie is a transmissible spongiform encephalopathy that affects small ruminants. Due to potential health risks, it is submitted to eradication measures. It is characterised in sheep by a genetic factor based on the PrP major susceptibility gene. Therefore, control strategies partly rely on selective breeding. Scrapie transmission mechanisms are still incompletely understood. The long infectious and undetectable incubation period makes direct data analyses difficult, hence the interest of a modelling approach. Our aim was to study the impact of control strategies, based on selective breeding, on the disease spread within a flock. To achieve this goal, an epidemiological dynamic model was coupled to the optimal outputs of a genetic model, through the breeding function. \\ The realistic SI-based epidemiological model takes into account demographic and epidemiological processes, including seasonality in breeding and transmission, genetic and age-dependent susceptibilities, long and variable incubation periods, etc. The genetic deterministic dynamic model represents breeding programs including overlapping generations, different selection steps and strategies for males and females, and non-random mating. It maximises scrapie resistance, while simultaneously minimising the loss of genetic progress in milk production. The resulting model was applied to French dairy flocks exhibiting contrasting PrP genotype frequencies. The epidemiological model was calibrated using French outbreak data. The robustness of the results was tested by a sensitivity analysis. \\ Several breeding strategies, corresponding to various levels of genetic loss, were compared. They had a strong impact on the disease outbreak and are therefore an efficient tool to control scrapie. However, their efficacy in terms of disease eradication strongly depended on the scrapie introduction scenarios (e.g. one initial infected ewe vs. one per year). These results hence emphasise the contribution of our original approach, which combines genetic and epidemiological models in a realistic framework.},