Forward genetic studies using animal models represent an important tool for the discovery of physiological and biochemical mechanisms that cause human disease. Genetic studies in model organisms complement direct human studies, with advantages that include the ability to apply experimental perturbations and to control both environmental conditions and genetic makeup of study populations. Access to disease-relevant tissues provides opportunities for large-scale molecular profiling and deep physiological phenotyping. Traditional forward genetic experiments in animal models employed crosses between two inbred strains. New genetic populations are being developed for rodents and other model organisms to serve as community resources for systems genetics studies. These multiple parent populations provide informative new features that are not available in traditional two-parent crosses. Multi-parent populations pose new analytical challenges, including haplotype reconstruction, the treatment of the multiple founder alleles and the need to account for kinship structure in genetic mapping. Large-scale molecular and clinical phenotypes collected on multi-parent populations present additional challenges and opportunities. In this project we will develop efficient and practical statistical methods to meet these diverse challenges. We will develop modular, extensible software, including tools for interactive data visualization that empower researchers to explore systems genetics data on multi-parent populations. With these general goals, our specific aims are to (1) Develop statistical methods for the genetic analysis of multiparent cross designs, including methods for haplotype reconstruction and genome imputation, parsimonious genetic models for the large number of possible genotypes in such crosses, and mixed models for polygenic effects to account for varying degrees of relationships among individuals; (2) Develop statistical methods for genetic analysis of high-dimensional phenotypes, including methods for RNA-Seq data, integration of disparate data types, and network modeling; and (3) Develop next-generation QTL mapping software for high-dimensional data, including functionality for interactive data visualization and multi-parental crosses.
Genetic studies in model organisms, such as the mouse, are an important complement to direct studies of humans, with many advantages, including the ability to control the environment and diet and to obtain deep physiological trait information, including molecular profiling of disease-relevant tissues. Populations formed by the inter-mating of multiple strains are supplanting traditional studies, based on crosses between a strain that is susceptible to disease and one that is resistant. Through this research effort, we will develop improved statistical methods and software for the analysis of systems of molecular and clinical traits in such multi-parent populations of model organisms.
Showing the most recent 10 out of 54 publications
