The Design And Implementation Of Distributed Data Management System For Large-Scale Virtual Screening

VS, short for Virtual Screening is an effective method in the field of drug discovery process, which attracts the close attention of both pharmaceutical companies and research institutions. VS aims to find the candidate compounds from millions of small molecules, which could dock on the targeted protein by shortening the research lifecycle and saving computing costs.VS process usually involves massive and millions of small molecules. In addition, VS process will produce corresponding results and a big amount of intermediate data. As a result, in order to guarantee the data integrity and effectiveness during the experiment, it is necessary to store and manage the massive data.With the development of computational methods and the improvement of computer’s performance, the computing speed of VS has been advanced. With the enhancement of the efficiency in VS, the data processing efficiency has become one of key factors corresponding to the VS efficiency.In this paper, we designed a distributed data management system for large-scale virtual screening. The process and storage of the large scale data was implemented. In order to improve the effectiveness of the data access, we proposed a data allocation strategy based on the performance of data nodes. As far as system scalability is concerned, we extended the data process and storage capability by increasing the process and storage nodes. In the end of the article, the test experiment for evaluating the whole performance of system was given, and the results show that the system is proved to improve the efficiency of data access and support a certain degree of system extensibility.