Current techniques for the diagnosis of genetic disease are limited by an inability to detect the majority of mutations and polymorphisms in DNA. DNA molecules which differ by one or a few base pairs can most efficiently be distinguished currently by denaturing gradient gel electrophoresis due to sequence specific differences in melting properties of the DNA. However, base changes can be detected only in domains of the molecule which melt first in these gels. We propose to improve this technique so that base pair changes can be detected in all domains of a DNA molecule. We will modify DNA probes in ways that will significantly alter the melting behavior of DNA hybrids, so that new regions of the DNA molecule will be included in the first melting domain. In addition, we will use experimental data to generate computer models to predict melting behavior of modified DNA molecules. These computer programs can then be used to design optimal probes for detecting sequence changes in any DNA molecule. This technology should be broad enough for application to any gene of interest, and could lead to the commercial development of diagnostic tests for a number of genetic diseases.