The ability to conduct whole genome analysis of variation is slowly becoming a reality with both improvements in biotechnology and advancements in data analysis. However, large-scale de novo sequencing still remains a formidable task for complex plant and mammalian genomes. Although not providing resolution at the sequence nucleotide level, physical maps convey useful information that can be leveraged to discover biological events not possible with sequencing technologies. Optical mapping, a novel restriction mapping technology, is able to produce complete genome-wide physical maps both quickly and cheaply. These maps serve not only as invaluable aids for de novo sequencing, but can be used directly to make valuable inferences regarding the underlying genome itself. However, in order for optical mapping to be useful as a tool for genomic analysis, both computational and statistical questions must be addressed. In this thesis, we explore some of the issues involved with analyzing optical mapping data. Specifically, we explore various statistical models and their implications for optical mapping data. We also develop a new scoring function for the alignment of optical maps using dynamic programming. A strategy for comparing optical mapping data against a clone-based sequencing strategy for a genome is examined. Finally, methods for assembling optical mapping data into a complete genome-wide physical map are presented.