Monte Carlo Procedure
Monte Carlo-based docking techniques used by applications such as ICM (Molsoft)7 start with an initial randomly generated ligand conformation that is subsequently modified via bond rotations or rigid body translations. As with genetic algorithm-based techniques, the resulting ligand conformations are scored using a pre-defined scoring function before the ligand is subjected to another round of modification. Energetically favorable conformations that exceed the defined scoring threshold are saved, while those that are energetically unfavorable are discarded. In this manner, multiple conformations can be sampled over the course of a docking run.
Due to its inherent speed, matching algorithm-based docking can be used to screen compounds rapidly. Matching algorithm-based techniques combine analysis of the topology of the binding site with parameters derived from pharmacophore modeling of the ligand and active site to identify putative binding conformations. To this end, the initial steps in a matching algorithm-based docking run involve generation of a ‘ligand-accessible’ molecular surface representation of the target receptor. Afterwards, the ligand is modeled into the identified binding site using pharmacophore-derived parameters as well as shape-fit algorithms.8 Matching algorithm-based techniques account for numerous factors such as hydrogen-bonding and van der Waals interactions in evaluating the favorability of the ligand conformation9-11 and are employed in applications such as Dock12 and SanDock.8