Drug virtual screening method

Abstract

The invention relates to a drug virtual screening method. Screening objects comprise a plurality of candidate compounds. The method comprises the following steps that S1, information of all candidatecompound molecules is written into a database; S2, the record corresponding to one of the candidate compound molecules is taken out from the database; S3, the taken-out records serves as an input fileto be stored in a local storage of a computing node; S4, screening and analysis are conducted on the input file, and an analysis result is written into the local storage of the computing node; S5, the analysis result is read from the local storage of the computing node and is inserted into the database as a record; S6, the record corresponding to another candidate compound molecule which is not processed is taken out from the database, and S3 is repeated till the records corresponding to all the candidate compound molecules are processed. According to the method, the burden is transferred into the local storage of the computational node and a database server, the load of a metadata server is reduced, and the performance stability of a system is guaranteed.

Description

technical field [0001] The invention belongs to the field of data management, and in particular relates to a drug virtual screening method. Background technique [0002] Drug virtual screening refers to the pre-screening of compound molecules on the computer before the biological activity screening in the process of drug development, so as to reduce the number of actual screening compounds and improve the efficiency of lead compound discovery. During virtual screening, the screening program needs to analyze millions of candidate compounds sequentially to obtain a score for the compound. Among them, some screening programs will store each candidate small molecule in a separate file as one of the inputs; at the same time, the scoring results of the screening will also be stored in a separate file. Therefore, for each candidate compound molecule to be screened, at least two small files need to be managed. [0003] The current virtual drug screening is mainly carried out on hi...

AI Technical Summary

Problems solved by technology

[0004] The global shared storage file system used by clusters and supercomputers, such as the Luster file system, is not good at managing a large number of small files. Even if the candidate drug small molecules are divided into multiple groups, each group is screened in turn, which not only limits the screening. Due to the concurrency scale, only one group can be screened at the same time, and even if grouped, a large number of small files related to drug virtual screening will still be directly stored on the global file system, which will cause excessive load on the metadata server of the file system , resulting in a significant drop in file system performance, affecting the operation of clusters and supercomputers

Images