Protein network module mining method based on multi-target optimization

Abstract

The invention discloses a protein function module recognizing method based on a multi-target optimization algorithm. The protein function module recognizing method includes the steps that a network protein node is judged to be a nonoverlapping protein node and a candidate overlapping protein node, the problem of protein network module function mining is solved through hybrid coding, particle swarm initialization and particle swarm evolution. The monotony problem of protein network function module combinations can be solved, multiple module combinations are provided for a user to select, and therefore function module mining accuracy and effectiveness are improved.

Description

technical field [0001] The present invention relates to the technical field of complex protein network functional module identification, specifically a protein functional module identification method based on a multi-objective optimization algorithm, which describes a protein network as a binary group and utilizes an overlapping community detection algorithm of a multi-objective optimization algorithm To solve protein functional module identification. Background technique [0002] Thousands of proteins in organisms form a wide variety of functional protein modules at different times and in different spatial stages. In biologically meaningful cellular functions, protein functional modules are one of the most basic constituent units. The process of gene products plays a very important role. How to mine protein functional modules closely related to biological functions from protein interaction data has become an important breakthrough for people to uncover protein interactions ...

AI Technical Summary

Problems solved by technology

Such methods use community detection algorithms to mine modules in protein networks, which can quickly meet the needs of users. However, protein networks form protein modules with a wide variety of functions at different times and in different spatial stages. Such methods can only be used for a protein network. Mining out a set of protein modules cannot meet the diverse needs of users

Images