Abstract :
Computer methods can now be used on almost every stage of drug development, but the most common areas
of computers application are virtual screening and lead generation/optimization stages. Accurate prediction of the
protein-ligand binding affinities is a crucial step in the structure-based drug design approach. Current algorithms and
tools for binding energy calculation that are used upon the development of new drug candidates with an emphasize
on underlying principles, advantages and limitations, software and general considerations in the selection of specific
methods are discussed in the paper.
Four main classes of currently available physics-based computer methods (molecular docking, end point /
approximate free energy, relative binding free energy, and absolute binding free energy) are reviewed in details.
Molecular docking approaches are the method of choice to filter out compounds-nonbinders, but they are not accurate
enough to predict binding affinity. The end point methods are more physically rigorous and closer to real free energy
calculations, but they are more computationally-intensive and not predictive for some types of proteins. Relative
binding free energy methods take into account conformational and entropic contributions, thus offering more
accurate predictions. However, they have high computational requirements and can be used only to compare related
ligands or receptors. The extremely computational-dependent method of absolute binding free energy calculation is
the most powerful approach, giving predictions with good correlations to experimental binding affinities.
Keyword :
Binding Energy, In Silico Tools, Computer-Aided Drug Design, Free Energy Calculation, Ligand Binding Thermodynamics