| cDNA microarray experiments represent a new class of biomedical research that allows monitoring the expression levels of thousands of genes simultaneously and thus enables a profound understanding of biological processes. Since microarray experiments generate huge amounts of complex data, sophisticated statistical approaches are necessary to properly address the problems under investigation. In this thesis, we examine an analysis method for gene expression data derived from cDNA microarray experiments. The understanding of the technical and biological concepts involved in the microarray experiments is essential to develop definitive statistical methodology. In this context, we first investigate the biological background and technical procedures behind microarray experiments. Second, we examine a number of published statistical methods that address related issues in the analysis of gene expression data, such as normalization and identification of differentially expressed genes. Next, we present a microarray experiment which is designed to determine the degree of transcriptional complexity of bone regeneration and to determine the temporal profile of genes expressed during fracture healing. Several interesting features from the bone regeneration data distinguish it from other commonly used microarray experiments. Using this data set, we proceed our work to develop proper methods for normalization and pattern identification. Finally, a simulation study is conducted to evaluate the performance of our methods for analyzing gene expression data. |
