For drug design it is essential to know which ligands can reach the active site of a protein. These ligands are potential candidates that inhibit or activate the given protein, and thereby cure a disease. We will show how to use sampling-based motion planning to solve protein-ligand disassembly problems.
The protein-ligand docking problem is computationally expensive, because of the high dimensionality of the search space. We overcome this problem by partitioning the degrees of freedoms (DoFs) into three sets (1 DoFs of the ligand with respect to the protein, 2 the ligands internal mobility, and 3 the mobile side chains of the protein) and do a decoupled motion calculation for each set of DoFs. For the search we extend the Exploring/Exploiting Tree (EET) algorithm. Using EET, problems with complex ligands become manageable, because the internal mobility of the ligands is handled efficiently by the EET internal DoFs version. Relevant side chains can be selected and simulated using the EET side chains version, enabling us to solve protein-ligand interactions where side chains hinder the ligand from exiting the protein.