Abstract
SSDB is a database that defines a mapping relation from disulfide bond patterns to protein structures. This database contains 4740 disulfide proteins (or 7880 chains) of known structures that contain at least two disulfide bonds. All the disulfide proteins are from Protein Data Bank. The database hierarchically classifies disulfide proteins according to their respective disulfide number, disulfide connectivity and disulfide patterns. There are 531 disulfide-pattern classes, starting from 2 to 19 disulfide bonds. In the database, proteins belonging to each disulfide pattern class correspond to a single fold only, despite possible low sequence identity. This database established a one-to-one mapping relationship from disulfide patterns to folds. Hence, given information of disulfide connectivity, SSDB can find possible folds for sequence of unknown structure without resorting to sequence alignment. SSDB also provides information of basic disulfide patterns common to proteins of similar folds. The database can be queried by using disulfide connectivity, or sequences in FASTA. DSSP also provides the CHIME interface for interactively viewing three-dimensional structures and links to other structural databases such as SCOP, CATH and CE. Our findings shed new light on the relationship between the protein conformations and their disulfide patterns and, and suggest an alternate way of finding structure homologues in homology modeling. The database can be accessed from http://because-1.life.nctu.edu.tw/ssbond.