The availability of high-throughput technologies has greatly advanced biomedical research. The resulting high-dimensional data. require the development of novel computational and statistical methodologies. In this dissertation, I present such methods in Statistical Genetics, Bioinformatics and the combination of the two fields, i.e., "Genetical Genomics", to aid high-dimensional genetic and gene expression data analyses. The dissertation is organized as follows:;In Chapter 1, I introduce related background in genetic and gene expression data analyses. Chapter 2 focuses on gene mapping in animals. I introduce a more powerful statistical approach that efficiently utilizes available pedigree information in mapping complex traits in heterogenous stock of mice (Huang and Zhao, 2010). The main focus of Chapter 3 is gene expression data analysis. I propose bias-corrected diagonal discriminant rules (BDDRs) for high-dimensional gene expression classification and analytically show that BDDRs outperform the existing methods (Huang et al., 2010). In Chapter 4, I present an expression quantitative trait loci mapping study (Huang et al., 2007). Finally, I discuss the current challenges in the related fields and outline future research directions in Chapter 5.
展开▼
