About D2Check Values at the Protein
and Residue Levels
As the field of proteomics advances with an increasing number of measured and
proposed protein sequences and structures, the validation and interpretation of
said structures becomes increasingly important. The urgent need for scoring functions
arises, not only from the more latent need to verify experimentally derived structures,
but also from the increasing use of de novo and homology computer models constructed
from known sequence.
While several useful scores exist, this site provides yet another tool
to help the community.
The dihedral angle distribution within a protein structure remains
an important aspect in approaching this problem. So called Ramachandran
plots, which evaluate the &Phi - &Psi values of each
in any given protein structure,
for example, have been widely used in protein validation
research. Beyond
the correlations in the internal &Phi and &Psi values of one residue,
the correlations in
the &Psii-&Phii+1 values between adjacent
amino acids have been seen to also play an
important role in dihedral angle analysis.
Based on this observation , a new checking function recently suggested by
Hernandez et al.,
uses the aforementioned dihedral angle correlations to evaluate the
degree of structural compatibility of a
given structure with the nonredundant set of
proteins deposited in the
Protein Data Bank (PDB).
The analysis is based upon
standard entropic functions rooted in information theory
and gives rise to the so-called D1 and D2 values.
These checks serve to extend the substantial understanding that has
already been gained using dihedral angle analysis.
This web site, called D2CheckServer
automates the process for obtaining D2Check values.
Given a PDB ID or user-provided PDB file, it calculates and displays the
associated &Phi - &Psi (i.e. Ramachandran) and
&Psii-&Phii+1 plots.
Using this information the server proceeds to calculate
the D2Check value
and the results are presented in two complementary forms.
First, the D2Check value averaged over the entire protein is
displayed with respect to the distribution of all proteins sampled in the
PDB.
The value is defined such that it simply represents
the standard deviation
from the mean of the distribution of proteins in the PDB.
Therefore, larger D2Check values indicate that a
given protein is increasingly dissimilar to the ensemble of structures
in the PDB.
To further aid in the analysis, a more detailed representation that assigns
D2Check values at the residue level are also presented
in the form of a convenient color strip.
In this manner, it becomes easy to locate ``interesting'' sites
of the given protein.
For more details on the construction and interpretation of
D2Check, please refer to our recent
publication
on the subject.
Using D2Check you will have access to both
the &Phi - &Psi and &Psi - &Phi
values and plots for any user-defined or existing protein structure
in the PDB.
Likewise you will have access to the D2 score of that particular
structure providing a measure of how
reasonable the structure is with respect to the PDB as measured by
the new entropic checking function.
The overall distibution of D2 for all
the structures in PDB is given below:
|