Computational Drug Discovery and Design

2.1 Programs
in the Cyscore
Package

Cyscore is an empirical scoring function for protein–ligand binding

affinity prediction. It is a combination of hydrophobic free energy,

van der Waals interaction energy, hydrogen-bond energy, and the

ligand’s entropy (seeNote 3). A curvature weighted surface area

model was introduced in Cyscore to improve the prediction of

hydrophobic free energy. Cyscore is an indicator of the shape

complementarity of protein–ligand binding.

CurvatureSurface is used for generating molecular solvent

accessible surface grid. It outputs a PDB format file with the

grid data in the ATOM records. Their “temperature factor” in

standard PDB format is replaced by curvature value at each grid

point. The output PDB file can be visualized by molecular visuali-

zation software such as Pymol or RasMol. If the surface grid is

colored by “B factor” or “temperature factor,” the concave regions

will be highlighted by warm colors while the other regions will be in

cold colors. This tool helps to identify binding pockets and validate

the protein–ligand shape complementarity.

RotaBond is used for counting rotatable bonds of ligands,

which are defined as any single nonring bond, bounded to nonter-

minal heavy (i.e., nonhydrogen) atom. The number of rotatable

bonds is correlated with the entropy loss upon protein–ligand

binding process. It is based on the assumption that each rotor in

the unbound state associates with a discrete number of low-energy

conformations but ‘freezes’ into a single conformation upon bind-

ing [29–31]. Thus the number of rotatable bonds is proportional

to a certain amount of conformational entropy. RotaBond outputs

the number of rotatable bonds as well as the list of the atoms that

form the bonds.

AddH is a tool for adding hydrogen atoms to a protein. It

predicts the coordinates of hydrogen atoms by stereochemical

Fig. 1Cyscore versus RMSD on a 106 protein–ligand docking set

Protein-Ligand Docking by Cyscore 235