Gene expression microarray data have great potential in helping researchers to understand the biological mechanisms of disease and hence their diagnosis. How to utilize and analyze these large-scale data to extract useful information is the major challenge of bioinformatics field. In this dissertation, we propose a rank-based framework for the statistical analysis of expression microarray data. We first explore the rank-invariant property of various microarray preprocessing methods, then propose a rank-based classifier called Top-scoring Triplet (TST), and finally we present a maximum entropy model of distribution on ranks. |