Program for searching a database of protein 3D structures for structural homology with a query protein

To improve speed, 3D-MatchDB uses an algorithm of fast alignment of secondary structure elements (helix, beta-sheet) and preprocessed PDB database, which has secondary structure elements mapped to 3D structures. Current version has 12,834 protein chains from PDB, cleared from redundant entries, so that their sequence homologies are not higher than 98%. 3D-MatchDB performs pairwise structural alignment of query protein with each database entry, calculates RMSD, Zscore, Aligned Size, and number of gaps for each alignment, and outputs a sorted list of entries that have structural homology to query protein with RMSD less than 5 angstrom and Zscore above 3.2. Then user can get atomic coordinates of structurally aligned pairs of proteins by picking one structure from that list and using 3D-Match program for refined alignment.
Parameters calculated by 3D-MatchDB (RMSD, Zscore, Aligned Size, and number of gaps) may slightly differ from those calculated by 3D-Match, as the former uses faster and slightly less accurate alignment algorithm.