Recent advances in de novo design strategy for practical lead identification.

De novo design programs such as LEGEND, LUDI, and LeapFrog can identify novel structures that are predicted to fit the active site of a target protein. However, in the conventional de novo design strategy, the output structures obtained from the programs can be problematic with regard to synthetic accessibility and binding affinity prediction. Thus it has been practically difficult to obtain novel lead compounds that are appropriate for medicinal chemists through the de novo design strategy. Since the late 1990s, several new strategies for lead identification have been reported and the successful examples have been disclosed. One of the strategies is validation of small fragments, which can be substructures of de novo ligands, by using NMR, X-ray, or MS spectra. Another method is prioritization of output structures obtained from de novo design programs by chemical accessibility. This review describes these new strategies with practical applications and future perspectives of de novo design. Copyright 2003 Wiley Periodicals, Inc.