Cluster Analysis of Microarray Gene Expression Data

Try to put well in practice what you already know. In so doing, you will, in good time, discover the hidden things you now inquire about. Rembrandt Harmenszoon van Rijn Preface As technique proceeds, the field of genetics is confronted with the possibility of obtaining, interpreting, analysing and evaluating a growing amount of data. One of the techniques used is the microarray technique, a high throughput genetic analysing procedure. In order to interpret and use those data, intelligent data analysing methods are necessary. Mathe-matical techniques will be used to assist in analysing data. Important techniques in this respect will be clustering techniques of several types. In the first place, these will be used to add unknown genes to clusters of known genes or to add new samples to existing clusters of samples. In the second place, they will be used to assist in creating functional clusters of genes or of samples without (necessarily) having knowledge about functionality or clus-ter structure. The goal of this thesis is to investigate the second application for cluster analysis and to compare some of the techniques with respect to their usability for gene clustering. Also these techniques will be used to cluster samples and identify genes that are most useful to differentiate between sample clusters. The relevance of gene clustering is to be able to identify presence or absence of functional relationships between certain genes in or between clusters respectively. Maybe also further insights can be developed in the genetic and biological pathways involved. The relevance of sample clustering is to identify genes important in the prediction of sample classification, in the first place and to be able to classify new samples into one of the sample clusters in the second place. I wish to thank