Study On Optimization Algorithms Of Semi-flexible Molecular Docking In Drug Design

With the continuous increase of the cost of a new drug development, pharmaceutical industry faces several huge challenges. They are eager to look for an effective optimization solution in drug development. The improvement of computer science makes the computer-aided drug design be true. Molecular docking that is a main method of computer-aided drug design has been used often.The thesis focuses on the semi-flexible molecular docking that the structure of receptor is no change, while ligand's changes during the docking process. The problems consist of a search algorithm for a conformation space and a design of scoring function.The thesis firstly studies the theory of molecular docking, several commonly be used conformation space search algorithms, scoring functions and some problems that is not solved at present. Secondly, in order to obtain an effective search for a conformation space, IPSGA algorithm (Improved Particle Swarm Genetic Algorithm) is proposed, Thirdly, in order to balance between the speed and the accuracy of scoring function, a scoring method based on force function and Gehlhaar potential function and explicit salvation model is proposed. Finally, the molecular docking algorithm is be proved by using a dataset of 77 crystal complexes to test its rationality and accuracy.