Chemistry Reference
In-Depth Information
There will also be significant growth of interplay between theory and experiment
as computational chemistry will indicate the need for more experiments and vice
versa. It will also be necessary to have a simultaneous increase of users'
knowledge in physics, chemistry, biology, computation and statistics [487-491]. It
is a long road but we are on track.
DOCKING AND VIRTUAL SCREENING PROGRAMS
Molecular docking is a computational protocol that predicts conformation of a
small molecule binding. In almost every drug discovery project that identifies a
lead, molecular docking plays an important role. The procedure can also be used
to analyze lead-target interactions. It is commonly used to screen virtually
libraries of compounds in order to identify those compounds with potential to
bind to a specific target. Molecular docking protocols achieve their objectives by
computing fitness scores for generated binding conformations. The algorithm
searches the space and identifies the conformation with the highest score as the
predicted binding conformation. Compounds with highest scores are considered
candidate target-binding compounds. A good scoring function is an important
component and this is still a current challenge [492-546].
Drug discovery research often relies on usage of virtual screening (
via
molecular
docking) to identify active hits in compound libraries. Effectively, virtual
screening, using a variety of methods, has been successful in finding hits in drug
discovery. One- (1D) and two-dimensional (2D) ligand-based approaches are on
one end of the spectrum. They rely on knowledge derived form the connectivity
information (one or more active ligands). These include 2D fingerprints,
substructure matching and even text-based simulations. These are fast successful
methods. The results depend however, on the similarity of active database
compounds to the query. Docking is on the other end of the spectrum. These
methods rely on physics-based scoring of protein-ligand complexes. On one hand
it takes more computational time, but on the other hand it yields the opportunity to
find new and diverse actives. These can be unrelated to existing active
compounds. Between these extremes we have we have 3D ligand-based methods
(pharmacophore and shape screening). These protocols use the information from
active ligand and look for other ligands that match the 3D properties of a query
molecule. Combinations of methods can also be used.
