Study on the Computational Modeling of the Flexibility of Protein Structure
|Course||Applied Computer Technology|
|Keywords||protein structure flexibility modeling simultaneously perturbation pointmutant GPCR-Ligand docking|
The three-dimensional structure of the protein determines its biological and cellfunctionalities. The flexibility of protein structure means that several equilibriumstructures exist for a native protein because of the subtle and local perturbationson the structure. To most of proteins, such flexibility is critical and thus plays animportant role in protein functioning. The current biochemical experimental way tocapture the flexibility is time-consumed and expensive job. The even worse situationis that such experimental ways can only be applied to water-soluble proteins. Thisthesis focuses on the computational way to modeling the flexibility of proteins in orderto provide some effcient helps for many subsequent applications.Backrub is a simple motion model, based on the observed uctuations in highresolution crystal structure, to describe the flexibility of proteins. Based on Backrubmodel, this thesis presents two parallel modeling methods to model the flexibility ofprotein structures, by fusing the multiple conformations after parallel perturbationand simultaneously perturbing a single conformation. The two modeling methodsare simultaneously perturbing protein backbone and side-chain so that a variety ofdisturbances can simultaneously happen at the same time. Therefore the methods canmimic the real protein conformational ?uctuations in nature.To validate the eflectiveness of the proposed models, both methods are appliedto predict point-mutation protein structure. Compared to the sequential movementmodel, our parallel models improve the performance in both the accuracy of the proteinside-chain and the predicting time. The second model which simultaneously perturbsa single conformation is further applied to provide flexible conformers for G protein-coupled receptor-Ligand docking problem. Such conformers help docking to producehigher-quality complexes.The modeling techniques for protein flexibility can simulate the real movement ofnatural proteins in agreement with some certain biological insights. In the applicationsof point-mutation protein structure prediction and GPCR-Ligand Docking, the meth- ods improve the accuracy of protein prediction and generate better complex structures,which make supplements to the fields such as protein structure prediction, moleculardocking and drug design.