In order to take advantage of techniques such as PCR, vector replication, etc. it is necessary to reassemble DNA fragments based on data resulting from the error-prone biological process of hybridization. We have developed new algorithms which extract reliable information about a "physical map" of a genome. By providing explicit, detailed information about the accuracy of the maps, for the first time biologists can make more efficient use of the maps in later stages of the DNA sequencing process. Furthermore, these algorithms can be used to pinpoint especially noisy sections of the data and thereby tell the biologists which additional experiments will be most fruitful. Biologists at the Whitehead Institute/MIT Center for Genome Research and at the Los Alamos Genomic Center are working with us to ensure the effectiveness of our approach and the usefulness of our software library.