sampling of the search space optimally. The performance of any
software depends on how well it performs these three basic opera-
tions. The biggest challenge lies in exploring the vast space gener-
ated as a result of combinatorial explosion by the number of
element types and their arrangement possibilities in 2D space, i.e.,
topology. Furthermore, the problem gets more complex when a
single topology has variety of conformations in 3D space. Hence, it
makes it impossible for any software to perform an exhaustive
search. The ability of a software to reduce the chemical space
without sacrificing on possible leads including other algorithmic
decisions ultimately decides its quality of outcome results. A de
novo drug design program face following constraints while devel-
oping underlying algorithms.2.1 Design
Constraint: Primary
Target
The first design constraint to be encountered comes from the target
of interest. The type of input, specific for the biological target has to
be decided first [19]. This constraint is directly connected to the
quality assessment of the candidate fragments because the con-
straints extracted from fragments are used for the scoring of gen-
erated structures [20]. All the ligand–receptor-related information
contributes to the formation of primary target constraint. The
ligand–receptor-related information can be inferred by the explora-
tion of the receptor binding site to reveal possible ligand-binding
points or regions and extraction of complementarity information
between receptor and ligands. In doing so, the design process is
made biased toward specific region of primary target such as bind-
ing site based on the ligand–receptor interactions and key recep-
tor–ligand interactions. Analysis of electrostatic regions, hydrogen
bond (H-bond) donor and acceptor regions, hydrophilic and
hydrophobic regions,π–cation andπ–anion interactions as well as
noncanonical interactions is critical to identify key ligand interac-
tions. Receptor regions having H-bonding potential are of special
interest due to the directional nature of H-bond acceptor and
donor, which often form key interaction sites. Aromatic stack pair-
ing at the catalytic site of proteins has been shown to play critical
role for the selectivity and specificity of a protein target toward a
particular type of molecules [21]. Aromatic residues play important
role in enhancing the affinity by bothπ–πorπ–ionic interactions
and shape complementarity. These interactions allow the allocation
of ligand atoms with a defined orientation and with a complemen-
tarity within a small region of space. The accuracy of constraints
comes from the accuracy of the primary target structure. Therefore,
it is important to consider the most accurate target protein struc-
ture. In the absence of experimental structures, computationally
predicted structures of primary target can be used. However, there
remains greater scope for errors in computationally predicted struc-
tures; previous studies have reported its successful applications in
structure-based programs [22–24].126 Shashank P. Katiyar et al.