and the inhibitory constant (K i ) are related using the above expressions. In the
process of identification of correct poses, the ranking is based on correlation with
corresponding K A values of the training compounds.
Steric, electrostatic, hydrogen bonding, inhibitor strain, and enzyme strain are
some of the important factors affecting the predictive accuracy of EI complex.
The binding energy is calculated from electrostatic (E coul ) and van der Waal’s
(E vdW ) interaction energies (Eqs. 5 and 6).
E coul r
ð Þ ¼
X N A
i¼1
X N B
j¼1
q i q j
4pe 0 r ij
ð5Þ
where N is the number of atoms in the molecules A and B, q is the charge on each
atom, r is the distance separating the two point charges, and e 0 is vacuum
permittivity.
The other contributing factor to the potential energy calculation is the van der
Waal’s contribution. The treatment of the non-bonded interactions is done by
implementing Lennard-Jones 12-6 function.
E vdW r
ð Þ ¼
X N A
j¼1
X N B
i¼1
4e
r ij
r ij
12
À
r ij
r ij
6
"
#
ð6Þ
where
e is the well depth of potential energy.
r is the collision diameter of the atoms i and j, respectively (Fig. 1).
Generally, molecules will be represented as a function of potential energy. Other
two ways of representation are surface and grid methods, of which surface representation was pioneered by Connolly [36, 37]. This surface-based docking is more
suitable for protein–protein docking to a good extent. Grid representation is also a
standard method for protein–protein docking [38].
There are three methods of treating the conformation freedom during docking.
(1) Systematic methods.
(2) Random or stochastic methods.
(3) Simulation methods.
Systematic method is a stepwise or incremental search. The algorithm tries to
explore all the degrees of freedom. The search can be done in two ways. First,
fragments of molecule are docked into the active site region and then joined to each
other. Second, the ligand is divided into rigid (core fragment) and the flexible part
(side chain). The rigid portion of the ligand is first docked into the grid region to
which the flexible regions are attached in an orderly manner.
In random search method, the search was accomplished by implementing
random changes to either ligand or a cluster of ligands. The pre-defined probability
Structure-Based Drug Design…
277
process of identification of correct poses, the ranking is based on correlation with
corresponding K A values of the training compounds.
Steric, electrostatic, hydrogen bonding, inhibitor strain, and enzyme strain are
some of the important factors affecting the predictive accuracy of EI complex.
The binding energy is calculated from electrostatic (E coul ) and van der Waal’s
(E vdW ) interaction energies (Eqs. 5 and 6).
E coul r
ð Þ ¼
X N A
i¼1
X N B
j¼1
q i q j
4pe 0 r ij
ð5Þ
where N is the number of atoms in the molecules A and B, q is the charge on each
atom, r is the distance separating the two point charges, and e 0 is vacuum
permittivity.
The other contributing factor to the potential energy calculation is the van der
Waal’s contribution. The treatment of the non-bonded interactions is done by
implementing Lennard-Jones 12-6 function.
E vdW r
ð Þ ¼
X N A
j¼1
X N B
i¼1
4e
r ij
r ij
12
À
r ij
r ij
6
"
#
ð6Þ
where
e is the well depth of potential energy.
r is the collision diameter of the atoms i and j, respectively (Fig. 1).
Generally, molecules will be represented as a function of potential energy. Other
two ways of representation are surface and grid methods, of which surface representation was pioneered by Connolly [36, 37]. This surface-based docking is more
suitable for protein–protein docking to a good extent. Grid representation is also a
standard method for protein–protein docking [38].
There are three methods of treating the conformation freedom during docking.
(1) Systematic methods.
(2) Random or stochastic methods.
(3) Simulation methods.
Systematic method is a stepwise or incremental search. The algorithm tries to
explore all the degrees of freedom. The search can be done in two ways. First,
fragments of molecule are docked into the active site region and then joined to each
other. Second, the ligand is divided into rigid (core fragment) and the flexible part
(side chain). The rigid portion of the ligand is first docked into the grid region to
which the flexible regions are attached in an orderly manner.
In random search method, the search was accomplished by implementing
random changes to either ligand or a cluster of ligands. The pre-defined probability
Structure-Based Drug Design…
277