PreDom consists of one database and three prediction services.
View domain-domain interface
Domain-domain interface residues are defined by the difference of Accessible Surface Area (ASA) between the whole protein structure and the individual domain structure (Δ ASA > 1 Å2). In our database, a domain-domain interface residue is classified as core interface residue, which is buried (relative ASA < 0.1) in the whole protein structure and exposed in the domain structure, and peripheral interface residue, which is exposed in both the whole protein structure and the domain structure.
Predict domain-domain interface
PreDom:Interface predicts domain-domain interaction residues from the 3D structure of domain. Domain interface residues are predicted by two algorithm; KIP and IP.
- KIP (Known Interface residue Positions): Domain interface residues are predicted using the positions of known domain interface residues in homologous domains(in SCOP or CATH).
- IP (Interface Propensity): Domain interface residues are predicted using the prediction score calculated from interface propensity, which means usage preference in domain-domain interfaces compared to the other part of domain surface. To increase the specificity of prediction, two interface propensities, "core" and "peripheral", are used for the score calculation.
Predict 3D structure of multi-domain protein
PreDom:Structure predicts 3D structure of multi-domain protein from 3D structure of each domain.
- If whole 3D structure of a homologus multi-domain protein to the query protein is available, PreDom:Structure predicts the whole 3D structure of the query protein by homology modeling using MODELLER
- Otherwise, PreDom:Structure predicts the 3D structure of each domain by homology modeling. If two 3D structures of domains in the query protein can be obtained, PreDom:Structure predicts their relative orientation as follows
- Candidate structures of a multi-domain protein are generated by ZDOCK from individual domain structures.
- Domain-domain interface residues of each domain are predicted by the algorithm used in PreDom:Interface.
- The distance between the C-terminus of first domain and the N-terminus of second domain is estimated using the number of residues in the domain linker.
- Candidate structures are sorted by DINE score that is a linear sum of docking score, interface score and end-to-end distance score.
Predict functional sites
PreDom:Evtrace predicts functional sites of a protein on the basis of phylogenetic and structural information.
- Homologous proteins of the target protein are collected by homology search with BLAST.
- A phylogenetic tree is constructed from the multiple alignment of the sequences of the homologous proteins by the NJ method.
- Sequence clustering is performed on the phylogenetic tree, and then conserved residues and class-specific residues (TR residues) are extracted.
- Surface regions in which the TR residues are concentrated are identified as canditates of functional regions.
Related Web sites
- Targeted Proteins Research Program
- ToSY: Predict Docked 3D structure of two proteins/domains
- ESPER: Evaluating system for protein complex with evolutionary relationships
We started the deveopment of this web site supported by the Targeted Proteins Research Program of the Ministry of Education, Culture, Sports, Science, and Technology, and continue the development supported by Platform Project for Supporting Drug Discovery and Life Science Research（Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from Japan Agency for Medical Research and Development (AMED).