Architecture: 1.20

CATH is a tree-like organisation of nodes that begins with the class node (i.e. the v5.0 of the tree) and ends with the domain nodes (i.e. the leaves of the tree). Each node below the class have a parent that they belong to, e.g. the parent of the H level (Homologous superfamily) is the T (Topology). Additionally, each node above the domains have child nodes that belong to them, e.g. the child nodes of a given H (Homologous superfamily) level are the S35 families (domains clustered at > 35% sequence identity).

CATH is a tree-like organisation of nodes that begins with the class node (i.e. the v5.0 of the tree) and ends with the domain nodes (i.e. the leaves of the tree). Each node below the class have a parent that they belong to, e.g. the parent of the H level (Homologous superfamily) is the T (Topology). Additionally, each node above the domains have child nodes that belong to them, e.g. the child nodes of a given H (Homologous superfamily) level are the S35 families (domains clustered at > 35% sequence identity).

The simple answer is that some of the larger superfamilies in CATH comprise relatives which are very structurally diverse, having different secondary structure embellishments. Attempting to provide a global alignment of all domains in the superfamily would result in poor quality alignments. As a result, structural alignments in CATH are generated by first clustering together the structures that have at least some recognisable structural similarity using a normalised RMSD score (SIMAX = RMSDxNumber of residues in the largest domain/number of aligned residues) less than a given cutoff.

A Close Structural Cluster describes a cluster of two or more CATH domains that have been grouped together due to their close structural similarity (SIMAX score < 5Å where SIMAX = RMSD x Number of residues in the larger domain/number of aligned residues).

A Distant Structural Cluster describes a cluster of two or CATH domains that have been grouped together with a more relaxed structural similarity threshold to investigate more distant evolutionary similarities (SIMAX score < 5Å where SIMAX = RMSD x Number of residues in the larger domain/number of aligned residues).

The term "structural neighbour" is used here to describe two CATH domains that share a degree of structural similarity. The SSAP algorithm is used to calculate this similarity and works by generating a structure-based pairwise alignment of the two domains then scoring this alignment using a RMSD score which has been normalised by the length of the larger domain divided by the number of aligned residues (SIMAX).