Protein fold recognition through application of residual dipolar coupling data

Residual dipolar coupling (RDC) represents one of the most exciting emerging NMR techniques for studying protein structures. However, solving a protein structure using RDC data alone is a highly challenging problem as it often requires that the starting structure model be close to the actual structure of a protein, for the structure calculation procedure to be effective. We report in this paper a computer program, RDC-PROSPECT, for identification of a structural homolog or analog of a target protein in PDB, which best matches the 15N-1H RDC data of the protein recorded in a single ordering medium. The identified structural homolog/analog can then be used as a starting model for RDC-based structure calculation. Since RDC-PROSPECT uses only RDC data and predicted secondary structure information, its performance is virtually independent of sequence similarity between a target protein and its structural homolog/analog, making it applicable to protein targets out of the scope of current protein threading techniques. We have tested RDC-PROSPECT on all 15N-1H RDC data (representing 33 proteins) available in the BMRB database and the literature. The program correctly identified the structural folds for approximately 80% of the target proteins, significantly better than previously reported results, and achieved an average alignment accuracy of 97.9% residues within 4-residue shift. Through a careful algorithmic design, RDC-PROSPECT is at least one order of magnitude faster than previously reported algorithms for principal alignment frame search, making our algorithm fast enough for large-scale applications.