Modern approaches in association mapping

In genetic association studies, the same marker is often associated with different complex phenotypes. As a result, it is difficult to discern how the genetic locus acts on the phenotype of interest: if the gene has a direct effect on the phenotype of interest or if the gene acts through a secondary phenotype. In paper 1, we devise a method to detect the direct effect of the marker on the phenotype of interest. In paper 2, we devise a method to detect the indirect effect when the marker acts on the phenotype of interest through a secondary phenotype. While the first two papers devise methods to determine the pathway in which the gene acts on the phenotypes of interest, paper 3 approaches population-based genetic association studies from a different angle. In population-based genetic association studies, the standard statistical approach is to model the phenotype of interest as a function of the offspring genotype. We develop a general framework of conditional score-tests that treat the genetic information as the random variable and condition upon the phenotypic information. By treating the phenotype data as deterministic, the power of the approach can be increased by incorporating assumptions about the phenotype distribution into the test statistic, while, at the same time, the validity of the approach is not affected by a mis-specification of the phenotypic assumptions.