Klinglmüller, F. (2008). Statistical methods for class - discovery and analysis in gene expression profiling studies [Diploma Thesis, Technische Universität Wien]. reposiTUm. http://hdl.handle.net/20.500.12708/183526
DNA microarray technology is an important tool in the field of life sciences. The ability to measure the amount of transcription for thousands of genes at the same time has led to new insights into the science of living cells. Seemingly in parallel this has led to the development of a large array of new statistical methods to cope with the high dimensionality of the data. This diploma thesis is going to present some of the methods currently used in the analysis of DNA microarray data and will finally show their application and results demonstrated on a dataset that I analyzed in the course of my cooperation with the DNA Microarray Facility of the Medical University of Vienna. This dataset came from a survey of 24 MALT lymphoma. We show the application of a recently proposed resampling framework for cluster algorithms, that allows for an estimation of the optimal number of clusters and provides an intuitive representation of the results for inspection of the cluster stability. The second part of the analysis is concerned with recovering the essential parts of the expression signature underlying the cluster partition, by using methods that test for the significance of genes and pathways.
