"Statistical methods play a central role in the field of modern genetics. New technologies are driving an explosion of high-dimensional datasets that will sustain a continuing need for new methods, theory and computationally efficient software. My proposal has two parts that will address the statistical challenges and creation of resources at the frontier of the science in this area. The methods development will be driven by, and applied to, several cutting-edge datasets in the fields of human disease genetics, population genetics and plant and animal breeding, guaranteeing impact on exciting scientific questions.The first part concerns a wide circle of ideas around haplotype estimation, genotype imputation and analysis of sequencing data. The overarching aim is to provide a suite of methods that can estimate haplotypes and impute genotypes in a unified and computationally efficient manner. In addition, we will work to create a reference set of haplotypes from tens of thousands of European and worldwide samples that will form a central resource for human disease and population genetic studies.The second part concerns the development of models for high-dimensional phenotypic data in genome-wide association studies. This is poorly developed area of human disease genetics with great potential for methods development and wide ranging applications."