Molecular docking is a structure-based drug design method that predicts the binding mode and affinity by studying the interaction of organic small molecule ligands with biological macromolecular receptors. Molecular docking methods have a wide range of applications in the fields of enzymology research and drug design. Since the Kuntz team at California State University, San Francisco developed the first molecular docking software DOCK in 1982, scientists have developed a variety of theoretical models and docking algorithms. The most important theoretical models and corresponding docking methods are:
1. Lock-and-key model, rigid docking;
2. Induced fit model (induced-fit), flexible docking (flexible docking) and semi-flexible docking (semi-flexible docking);
3. Conformation ensemble, ensemble docking.
The essence of molecular docking is the recognition process between two or more molecules, involving spatial matching and energy matching between
molecules. The docking software places small ligand molecules at the active site of the receptor target, and searches for ligands by continuously optimizing the position, conformation, dihedral angle of the rotatable bond, and the side chain and skeleton of the receptor amino acid residues. The best conformation for binding of small molecules to the receptor target, prediction of its binding mode and affinity.
Autodock Vina and DOCK6 are two widely used open source and free molecular docking software, with a very high accuracy. Nevertheless, it is still necessary to be cautious to perform accurate molecular docking calculations. Because the recognition process of biological macromolecules and small molecules is quite complicated, researchers cannot ignore the full understanding of the research system, such as the recognition of hot residues and key interactions, the removal and retention of water molecules in the pocket, and the treatment of metal ions. At the same time, the experimentally measured three-dimensional structures of biological macromolecules often contain various problems, such as low precision and missing atoms. These will seriously affect the accuracy of the molecular docking results. Therefore, pre-docking is particularly important. A large number of studies have shown that for compounds with large differences in the molecular structure of co-crystal ligands, or systems with larger (or smaller) receptor pockets or larger flexibility, using default parameters and operations often results in disappointing results. The common situation is that the binding modes of the derivatives of the same skeleton should be similar, but the conformations of different orientations are obtained; according to the docking score, the compounds with better activity are ranked lower, and the compounds with lower activity are ranked higher. In addition, the computing performance of the computer is also one of the factors restricting the docking of large-scale molecules. In these cases, more experience guidance or special treatment needs to be added.
The application of molecular docking mainly has the following two aspects:
Molecular docking can be used to study/predict the interaction between active small molecules and biological macromolecules, and to understand its mechanism of action.
Molecular docking can be used to quickly find small molecules with potential activity from chemical databases, save experimental costs, and improve hit rates.
If conditions permit, the former needs to examine the reproducibility of the docking software (method), that is, re-dock or self-dock, and the latter needs to examine the screening performance, including Enrichment rate, goodness of fit. Only through rigorous and scientific treatment can we ensure that the simple docking process gets accurate calculation results.