Probalistic Methods In Genomic Data Analysis

In this thesis, three aspects of gene expression data analysis are discussed: Differential gene expression is addressed by a probabilistic method. Gene annotation enrichment analysis is discussed in the context of multiple hypothesis testing and the choice of null hypothesis. The possibility of inferring the activity of cellular signaling pathways from microarray data is explored. The methods developed are applied to various data sets. The method for differential gene expression is applied to aspects of B cell differentiation. The methods for annotation analysis and pathway activity inference are applied to data sets of breast cancer, colon cancer and leukemia.