Tiling microarray informatics

Since the late 1990s, DNA microarrays have become an increasingly popular tool for the study of molecular genetics, This is due in part to several technological advances I hat have steadily increased the number of unique biochemical assays that can he carried out with a single DNA microarray. Today's microarrays can facilitate millions of such experiments in parallel. Since the human genuine contains only about 25 000 unique genes, this leaves a great deal of microarray real estate for designing experiments that target regions of the genome with no known function. Tiling microarrays use the available bandwidth for assaying the putatively inactive regions of the genome by considering evenly spaced units of the genome as individual regions' for study. This thesis deals with some of the computational challenges encountered when examining data produced by tiling microarrays. Areas covered include the identification of interesting regions of the genome from tiling microarray data, and the compensation for and use of cross-hybridization signals for biological discovery.