![]() |
3 Dee - Database of Protein Domain Definitions |
List of all PDB files and the corresponding domain definitions that are included in the 3Dee database. Currently 3Dee is in line with the version of the PDB at the 21st July 1998. Note that domains which consist of segments from more than one chain, so called 'multi-chain domains', appear more than once in the list, as the 3Dee setup requires identical multi-chain domain definitions for each of these chains.
When sequence families are divided into similar domain organisation families, domains with significant sequence similarity may be separated. Therefore the representatives of the domain families do not make up a non-redundant set of sequences.
Derived after clustering the representatives of the domain families by their sequence similarity; this is a non-redundant set of domain sequences.
The representatives from the domain sequence families are clustered at different thresholds of structural similarity calculated with the STAMP program. The representatives of the resulting domain structure families are a set of structurally non-redundant domains.
pdb1gdh.ent 1gdha-1-AS { A 2 _ to A 100 _ A 290 _ to A 321 _ } 1gdha-2-AS { A 101 _ to A 289 _ } 1gdhb-1-AS { B 2 _ to B 100 _ B 290 _ to B 321 _ } 1gdhb-2-AS { B 101 _ to B 289 _ }The domain identifier consists of a four letter PDB code, a chain identifier, the initials of the database annotator who has defined the domain and a revision number, if more than one revision exists. For example 2mysa-4-AS.2 refers to the fourth domain of the 'a' chain of protein 2mys; it is the second revision of the domain definition produced by AS. The annotator's initials may be replaced by the terms DOMAK or AUTO that represent automatic domain definition methods. DOMAK locates domains for a given set of co-ordinates, while AUTO is using sequence similarity to chains with known domain boundaries.
The domain definition format consists of the following elements:
[chain id.] [res. number1] [ins. code1] to [chain id.] [res. number2] [ins. code2].
![]() |
|
Comments to
Uwe Dengler EMBL-European Bioinformatics Institute Wellcome Trust Genome Campus, Hinxton Cambridge, CB10 1SD, UK |