PB-PENTAdb home
PB-PENTAdb : database of pentapeptides
A database of pentapeptides was derived from SCOP 1.75 culled at 100% sequence identity (astral100 dataset) but also at lower sequence identity filtering. A sliding window of 5 residues is used to extract the coordinates of every pentapeptide contained in all domains from SCOP. The dihedral phi and psi angles are then calculated and mapped to a trained neural network of 16 cells each representing a local structure prototype (the protein blocks) named a, b, c, … , n, o, p according to the methodology developed by de Brevern et al (2000). Each pentapeptide hence extracted is assigned to the PB with which it has the lowest rmsda value (angular root mean square deviation). Information about the sequence, the PB, the dihedral angles, the secondary structure (from DSSP), solvent accessibility (from NACCESS) and backbone B-factors are stored in PB-PENTAdb database.
Basic search
The whole process for building the database is illustrated below :
|
The program takes as input a pentapeptide amino acid sequence in the input text box below. Hint : use the underscore sign “_” to designate any amino acid. Example DI_RG. |
As output, a table is provided with
- the pentapeptide sequence
- the PB
- the SCOP id of the domain where this pentapeptide is extracted
- the position of the first residue of the pentapeptide in the domain
- a series of psi and phi values
- the secondary structure assignment (from DSSP)
- a series of 5 values corresponding to the “all atoms” relative solvent accessibility for all 5 residues
- a series of backbone B factors