The CDD (48) contains over 28 000 PSI-BLAST-derived Position Specific Score Matrices representing domains taken from the Simple Modular Architecture Research Tool (Smart) (52), Pfam (53), TIGRFAM (54) and from domain alignments derived from COGs and Entrez Protein Clusters. In addition, CDD includes 3100 superfamily records, each of which contains a set of CDs from one or more source databases that generate overlapping annotations on the same protein sequences. The NCBI Conserved Domain Search (CD-Search) service locates conserved domains within a protein sequence, and these results are available for all proteins in Entrez through the ‘Conserved Domains’ link. Wherever possible, protein sequences with known 3D structures are included in CDD alignments, which can be viewed along with these structures using Cn3D. Cn3D is also equipped with advanced alignment-editing tools that use variants of PSI-BLAST and threading algorithms. The Conserved Domain Architecture Retrieval Tool allows searches of protein databases on the basis of a conserved domain and returns the domain architectures of database proteins containing the query domain. CD alignments can be viewed, edited or created de novo using CDTree. CDTree
