Contributions to statistical inference from genomic data. (Contributions à l'inférence statistique pour des données génomiques)

Pierre Neuvial

Abstract: This thesis summarizes my research contributions to the development, mathematical study, and application of statistical models and methods that take advantage of the complexity and structure of genomic data to address biomedical questions. These contributions deal with (1) multiple testing and post hoc inference, (2) inference from heterogeneous data (combining several data types that correspond to the same set of observations or variables) and (3) inference from ordered data (segmentation, classification under contiguity constraints).