Second, the pharmacophore models are ranked based on a measure of sensitivity and specificity and the top models are returned. The pharmacophore models are enumerated and then the selectivity is estimated based on a Genetic Function Approximation GFA) model. The GFA model for the selectivity of a pharmacophore is built from a training set of 3285 pharmacophore models. This set is used for searching the CapDiverse database in DS. The logarithmic values of the number of database search hits are 
used as the targets. The number of total features in pharmacophore models and the feature-feature distance bin values are used as the descriptors for training the GFA model. Molecular docking studies were carried out using GOLD 5.1 program from Cambridge Crystallographic Data Center, UK. GOLD uses a genetic algorithm for docking ligands into protein binding sites to explore the full range of ligand conformational flexibility with partial flexibility of protein. Molecular docking was performed to generate the bioactive binding poses of inhibitors in the active site of enzyme. Protein coordinates from the Niltubacin crystal structure of chymase co-crystallized with N7O, determined at a resolution of 1.8A ? were used to define the active site. All the water molecules present in the protein structure were removed and hydrogen atoms were added. The active site was defined with a 10 A ? radius around the ligand present in the crystal structure. Ten docking runs were performed per structure unless three of the 10 poses were within 1.5 A ? RMSD of each other. All the hit compounds as well as training set compounds were docked into chymase binding site. The GOLD fitness score is calculated from the contributions of hydrogen bond and van der Waals interactions between the protein and ligand, intramolecular hydrogen bonds and strains of the ligand. The interacting ability of a compound depends on the fitness score, greater the GOLD fitness score better the binding affinity. The protein – ligand interactions were examined by DS. Hit molecules which showed the expected interactions with the critical amino acids present in the active site of the protein, and comparable high binding scores than the bound ligand, were selected as potent inhibitors of chymase. The experimentally known and highly active chymase inhibitors with substantial structural diversity which were used for the common feature pharmacophore generation were selected for DFT calculations. Moreover, four final hits KM09155, HTS00581, HTS0589, and Compound1192 retrieved from databases by the selected pharmacophore models, which showed important results with respect to all Z-VAD-FMK Caspase inhibitor properties like key molecular interactions with the active site components, calculated drug-like properties, and high GOLD fitness score, were also designated for DFT study. Various quantum-chemical descriptors such as LUMO, HOMO, and locations of molecular electrostatic potentials were computed. In this study, four different 3D structures of chymase bound with its inhibitors such as 3N7O, 1T31, 3SON, and 2HVX were selected as input for structure-based pharmacophore generation. The generated four pharmacophore models along with their excluded volume spheres and geometrical constrain are illustrated in Figure 4. The excluded volume spheres presented in our models provide an insight regarding the disallowed regions in the binding site.