Method for identifying biological essential protein

Abstract

The invention provides a method for constructing necessary protein recognition between multi-dimensional biological attribute information and PPI topological characteristics by using a deep neural network, and in order to solve the problem that gene expression data is not comprehensive, absent data is supplemented so as to improve robustness. And the convergence speed of the deep neural network is reduced by respectively constructing a PPI network topology structure, a Pearson's correlation coefficient and a homologous correlation coefficient. And finally, searching an optimal incidence relation of three features of the degree of the node, the Pearson's correlation coefficient and the homologous correlation coefficient through the deep neural network, thereby improving the recognition precision of the essential protein.

Description

technical field [0001] The invention belongs to the field of life sciences, and in particular relates to a method for identifying biologically necessary proteins. Background technique [0002] Protein is an essential element for body activity. In the process of life, proteins are closely connected to each other to complete a series of physiological activities, and finally form a protein network (protein-protein interaction, PPI). There are some essential proteins in the protein network. After the mutation is removed, the associated functions of the body will be lost, resulting in the body not functioning normally. Therefore, the prediction of key proteins based on the PPI network has a theoretical basis for the exploration of disease-causing genes and the development of drug targets. [0003] In the early days, identifying important proteins occurred mainly in biological experiments. Although bioassay techniques are highly accurate, such experiments are time-consuming and...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a method for identifying biologically necessary proteins to overcome the time-consuming training of the identification process of key proteins found and described in the current prior art, and the complex and cumbersome defects of model correctness verification

Images