The following two tables display information about the most current database versions of SCOP and CATH. In particular they display their protein and domain content, their number of families, superfamilies, folds and classes (homologous superfamilies, topologies, architectures and classes respectively). The second table displays the content of the two hierarchies after mapping the domain definitions as described in the paper with a minimal overlap of 0.8.
The following two tables display the mapping of inner nodes from SCOP onto CATH nodes via the mapping method discussed in the paper. Only the best node from CATH according to the F-measure is mapped onto the query node in SCOP. The first table displays all mappable nodes, i.e. all nodes which have a F-measure > 0.0. In the second table we require a higher quality for a mapping, i.e. a F-measure > 0.8. While the mapped nodes now fit very well and their domain sets have a large "overlap", many nodes do not find a mapping partner in CATH which fulfils the quality criterion.
Those two tables display detailed data on the consistency checks of domain pairs from CATH via SCOP. The first table displays the cases (number of pairs and percentages summing up to 100% per column) being observed in the consistency checks. E.g. 474.399 pairs (2.604%) two domains from the same CATH superfamily are classified to be in different folds of the same class in SCOP. The second table displays the same type of information but this time shows how many distinct folds and superfamilies from SCOP and topologies / superfamiles from CATH account for the cases observed. E.g. we find that the 474.399 pairs accounting for the example described above origin from 69 fold and 91 superfamilies in SCOP (29 topologies in and 35 superfamilies in CATH respectively). Green (positive pairs) and Red (negative pairs) cells in the tables correspond to the pairs of domains selected for the benchmark sets extracted from our mapping between SCOP and CATH.
Those two tables display detailed data on the consistency checks of domain pairs from SCOP via CATH. The first table displays the cases (number of pairs and percentages summing up to 100% per column) being observed in the consistency checks. E.g. for 70.188 pairs (0.866%%) two domains from the same SCOP family are classified to be in different folds of the same class in CATH. The second table displays the same type of information but this time shows how many distinct folds and superfamilies from SCOP and topologies / superfamiles from CATH account for the cases observed. Therefore, we find that the pairs accounting for the example described above origin from 25 fold and 26 superfamilies in SCOP (54 topologies in and 77 superfamilies in CATH respectively). Green (positive pairs) and Red (negative pairs) cells in the tables correspond to the pairs of domains selected for the benchmark sets extracted from the mapping.
In order to allow for an interactive use of the mapping of SCOP and CATH computed we have implemented an interactive browser. You can choose any SCOP or CATH node as entry point and explore the relationships between SCOP and CATH in detail by clicking on a node (a SCOP or CATH set) in the displayed graphs to display its respective mapping in the other classification (CATH or SCOP, respectively).
In the table below we show all pairs of domains calculated on the mappable subset of all CATH domains sharing less than 50% sequence identity, which are inconsistently defined in SCOP and CATH. Only pairs where both SCOP domains belong to the classes 'a'-'d' are shown. The rows correspond to CATH levels, the columns to levels of the SCOP hierarchy. Clicking e.g. on the class row and the family column shows all pairs of domains which are classified to be in the same family in SCOP but to be in different classes in CATH.
Here we display the subgraphs of similar folds in SCOP which have been identified by a consistency check in CATH as described in the paper. We show the SCOP fold identifiers as well as the number of connections / links per fold in brackets. Click on the image link to view the fold graph in detail.
Here we display the subgraphs of similar topologies in CATH which have been identified by a consistency check in SCOP as described in the paper. We show the CATH topology ids as well as the number of connections / links per fold in brackets. Click on the image link to view the topology graph in detail.