Conventional genome mapping and sequencing involves the analysis and processing of individual samples and pieces of experimental data. Although these methods work, it is quite clear that more efficient and less expensive methods are needed. Our top down physical mapping experiments have focused on the parallel processing of information from multiple samples at one time. This approach has aided the construction of genomic restriction maps and allowed us to assess the degree of large-scale conservation across wide regions of the human genome. The principles of parallel processing were applied in top down experiments that ordered an overlapping cosmid library from the 14-Mb Schizosaccharomyces pombe genome. This approach produced an eight-fold increase in efficiency in clone ordering over similar efforts. Recently, we have developed an enhanced sequencing by hybridization protocol that allows DNA sequence information to be collected on a large number of samples at once. Our current research focuses on applying parallel processing principles to make genome-wide comparisons between pairs of samples for analyzing disease states.