On Protein Sequence, Structure and Evolution: What can integration of the databases reveal?

In order to compare SCOP and CATH, the Astral RAF Sequence Map was used as a reference for PDB entry contents. Only the common subset of the two databses was compared, comprising 16948 PDB entries. The individual segments that form the domains were compared first and then the matches between the domains composed of these segments. A mismatch of up to five residues was allowed. The comparison of the allocation of domains across the two structural classifications was performed considering only the 30307 domains in CATH and SCOP that directly match each other. We performed two different independent analyses to identify the Pfam and PROSITE families that match SCOP domains. First, we used the PDB mapping provided by MSD in order to translate SCOP domains into SWISS­PROT sequences. Then for each SWISS­PROT sequence, the SCOP segments were compared with those of the Pfam and PROSITE families as provided by the InterPro database. For the comparison of SCOP and Pfam domains, a coverage of 70% was allowed, whereas the analysis of Prosite was restricted to those families which match entirely within the SCOP domains. In the second analysis we used rpsblast and pftools sequence search methods to match SCOP domain sequences onto Pfam and PROSITE families. The SCOP domains belonging to the three classes Low resolution protein structures, Peptides and Designed proteins were excluded in both analyses.