Molecular Docking Based On Optimal Search Theory Research

Molecular docking technology is one of the main methods of computer-aided drug design right now. With the development of computer technology, medicinal chemistry and molecular biology, an increasing number ofsmall molecule compounds and target proteins are constantly being discovered.Experiments will be people in drug discovery in molecular docking encounter with the dual challenges of large-scale data to calculate the mass data storage.The vigorous development of computer and the Internet technology brings cloud computing technology which provides new ideas and solutions for us to resolve molecular docking of large-scale data computing and massive data storage difficult.Since the year of 2006, Hadoop has become an independent open source Apache project.Withits advantages of high-performance and low-cost, there are large data processing needs has been widely used in Hadoop.The key issue of this thesis for data storage in bulk molecular docking, and docking realization inquiry, set up a Hadoop cluster has five nodes and the deployment of the Hive framework on which to build a cloud data storage molecule docking database providing query and analysis capabilities. Starting establish metadata tables using MapReduce computing framework ligand molecule docking results files and file parsing, generating the corresponding data files stored in the cloud database needs from the perspective of the user’s query. In the background of Hadoop and Hive technology, we will study the parallelization of molecular docking processes related issues, and the main contentsof this thesis are as follows:1. Through the theoretical analysis of molecular docking algorithm, it shows that Autodock Vina implementation principle of molecular docking and efficiency, and the feasibility of the molecular docking problem based on cloud platform processing.2. Setting up a cluster of Hadoop as the experimental platform and building a cloud database for mass data storage based on Hive build, providing data query and analysis capabilities.3. Using MapReduce programming framework for parallel docking capabilities, researchingon the problems of AutoDock Vina software called in Hadoop, while designing the Map functions and Reduce functions, respectively, summarizing the distribution of molecular docking mission and docking results.4. Using the MapReduce framework to achieve docking summary of the results, filter out all the small molecule ligands and their corresponding values are stored in the scoring HDFS, Hive and import the database for the user query, analysis.5. By contrast to write a script that calls Vina Shell commands and MapReduce framework for parallel docking two ways to improve the cloud platform to test the efficiency of molecular docking, analyzingthe result of molecular docking by data loading, combination querying, multi-table join queryingand sorting querying and so on.The research work in this thesis can provide a demonstration in the field of molecular docking studies by the cloud computing technology.