The Study On Sampling Algorithm For Protein-ligand Flexible Docking

Understanding the interaction between protein and ligand(P–L)is crucial for the pharmaceutical discovery.Accompany by the rate of 3D protein individual structure determination increases every year with the development of techniques such as X–ray and NMR,prompting computational algorithms have been developed that use the individual structure of the receptor and ligands to predict the high-definition 3D structure of their complex.In contrast to the traditional rigid docking,P–L flexible docking have to sample huge conformational space.Therefore,this dissertation analysis the key technologies for P–L flexible docking,mainly focus on the research of conformational space sampling algorithms:1.The analysis on the key technologies of P–L flexible docking.We introduce two central problems,conformational space sampling and complex conformation scoring,for P–L flexible docking.The two central problems are further subdivided several subproblems,include protein flexibility,ligand preparation,conformational sampling algorithm,scoring function,and analysis or post–processing of predictive results,with major emphasis being given about conformational space sampling algorithm.2.The study on the ensemble docking protocols using Rosetta Ligand protocol.We propose extensions to the Rosetta Ligand protocol with ensemble docking approach,which incorporate geometric distance constraints(GDC)and repacking receptor side–chain(RRSC)methods to generate and select representative protein structures as docking templates.The ensemble docking methods can produce and identify more correct docked complexes for proteins where satisfactory results can not be obtained using traditional Rosetta Ligand protocol.3.We extend the replica exchange Monte Carlo(REMC)sampling algorithm to P–L docking conformational prediction using Rosetta Ligand protocol.We use multi–objective optimization Pareto front information to facilitate the selection of replicas for exchange.An in–depth comparison between MC,REMC,multi–objective optimization–REMC(MO-REMC),and hybrid MO–REMC(HMO–REMC)sampling algorithms was performed.The experimental results show that the MO–REMC and HMO–REMC conformational sampling algorithms are powerful approaches for obtaining P–L flexible docking conformational predictions based on the binding energy of complexes using Rosetta Ligand protocol.4.We extend the adaptive temperature REMC(AT-REMC)sampling algorithm to Rosetta Ligand.The AT-REMC sampling algorithm use mean acceptance rate information of joint chains to “learn” the best temperature parameter values so that the algorithm can converge quicker while traditional REMC algorithm run.An integrative comparison between MC,adaptive temperature MC(AT-MC),REMC,and AT-REMC sampling algorithms was performed to illustrate the differences between the four sampling methods.The experimental results demonstrate that the AT-REMC sampling algorithm is efficient method for P–L flexible docking conformational prediction.The major contributions of this dissertation include four aspects: Firstly,in contrast to the traditional rigid docking,two central problems of key technologies for P–L flexible docking are outlined.Secondly,for protein flexibility,we propose extensions to the Rosetta Ligand protocol with ensemble docking incorporate GDC and RRSC methods to generate and select representative protein structures as docking templates.Thirdly,we originally propose MO-REMC and HMO-REMC enhance sampling algorithms to dramatically increase the efficiency of conformational space sampling.Fourthly,we extend original Rosetta Ligand protocol using AT-REMC sampling algorithm.All the experimental results show that these tactics and methods have played an important role in study on P–L flexible docking.The results and methods may have certain reference to following related studies.