The ability to create compounds which can exert a direct and specific therapeutic effect on cellular proteins involved in cancer is of great national importance. In order for a cancer drug to be effective, it must be specific for a protein involved in the disease process. In this grant proposal, we address both of the major problems associated with developing such useful drugs. The first concerns the selection of candidate compounds targeted against proteins involved in the disease process, and the second concerns techniques to derive structural information sufficient to produce a mimic which meets all the stringent requirements of a useful drug. We will simultaneously address both of these concerns by designing a process for the identification of ligand mimics based on the selection of constrained peptides by an immobilized target receptor, which can be integrated with a new high speed structural method. Specifically, we will identify and determine the structure of specific legend mimics for a cell surface protein involved in human cancers; human Notch. CuraGen is able to produce this solution by taking a novel interdisciplinary approach, combining expertise in NMR, molecular dynamics and molecular biology.