Gene module analysis method

Abstract

The invention discloses a gene module analysis method. The gene module analysis method comprises the following steps: inputting a gene phenotype double-layer network, a gene function similarity network and a known disease gene set s0 related to a disease phenotype; increasing the connection relationship between genes and phenotypes in the gene phenotype double-layer network; taking genes which arein edge connection with disease genes in s0 and are not in s0 in the gene phenotypic double-layer network as candidate genes; calculating and selecting a candidate gene with the maximum sum of semantic similarity, topological similarity and phenotypic relevance, and adding the candidate gene into s0; and when the expanded candidate gene set does not significantly enrich the GO ontology function annotation and the biological pathway gene related to the disease phenotype at the same time and express the differential gene in the disease phenotype sample and the normal sample any more, recordingthe current algebra as m, and outputting m-1 as candidate genes expanded in the first generation s0 and known disease genes connected with the edges of the expanded candidate genes.

Description

technical field [0001] The invention relates to the technical field of data analysis, in particular to a gene module analysis method. Background technique [0002] At present, the detection methods of disease modules in human interaction networks can be roughly divided into three categories: [0003] The first type of research method is to detect disease modules based on gene expression data. Gene expression profile data is a common and effective data resource for the mining and research of disease modules. Because there are still a lot of errors, missing, and high-dimensional defects in the existing data, it is not possible to obtain good results only by using the previous clustering algorithm to mine the disease module structure. [0004] The second type of research method is to detect disease modules based on a single network, but the one-sidedness and errors of single network data will have a great impact on the mining of gene modules, resulting in a large deviation bet...

AI Technical Summary

Problems solved by technology

Because there are still a lot of errors, missing, and high-dimensional defects in the existing data, it is not good to use the previous clustering algorithm to mine the disease module structure.

[0004] The second type of research method is to detect disease modules based on a single network, but the one-sidedness and errors of single network data will have a great impact on the mining of gene modules, resulting in a large deviation between the final gene module and the actual gene module

Images