Hypertension Gene Data Analysis Based On Hypergraph Clustering

Hypertension is a cardiovascular disease that seriously endangers human health.There are many factors that induce hypertension,which have not yet been fully elucidated.Among the many factors that induce hypertension,the most common cause of essential hypertension is genetic inheritance.With the increasing mortality of hypertensive patients,the decreasing awareness rate,and the trend of younger age in recent years,researchers pay more and more attention to hypertensive patients.In recent years,due to the great progress in high-throughput,short-read,low-cost testing technologies such as gene chips and DNA microarrays,many methods for analyzing genome expression data have emerged.Activity,to explore the pathogenesis of some diseases from the perspective of genes.The application of many analysis methods such as ensemble algorithm,clustering algorithm and weighted gene co-expression network analysis in this field provides a new way to analyze a large amount of gene data.Therefore,we need more methods to study the pathogenesis of hypertension and unearth the key genes that affect its occurrence.With the continuous development of gene chip technology and the reduction of sequencing costs,a large amount of biological data has been generated,which not only brings opportunities but also brings huge challenges to the research of biological data mining.At present,the main problems in the field of research on key factors of hypertension are: primary hypertension is complex and changeable,and the main mechanism of inducing hypertension is not yet clear;the important factors affecting the incidence of hypertension are not fully understood.The etiology of hypertension is complex,and the mechanism of inducing hypertension is not clear;the key factors affecting the occurrence of hypertension have not been fully determined.Therefore,more research methods are needed to explore the key factors affecting the incidence of hypertension.This paper makes innovations in the following aspects:（1）Drawing on the concept of weighted gene co-expression network and combining with hypertension genes,a hypergraph-based weighted gene expression super network is constructed.（2）Analyze the shortcomings of the traditional clustering algorithm,use the hypergraph clustering algorithm to mine hypertension pathogenic genes,and improve the hypergraph clustering algorithm,and finally compare the effects of the traditional clustering algorithm.（3）For the disease-causing genes mined by hypergraph clustering,a hypertensive gene classification prediction based on Ada Boost was designed to determine and classify diseased hypertensive samples and normal hypertensive samples.The hypertension gene data studied in this paper comes from the Gene Expression Database（GEO）.By constructing a hypertensive gene expression super network,a hypergraph clustering algorithm is proposed to mine the pathogenic genes related to hypertension.The clustering effect visualization and clustering evaluation criteria were compared,and it was proved that the clustering effect of the improved hypergraph clustering was better than the traditional clustering algorithm.A strong classification model for the judgment of hypertension was designed.Hypertension samples were classified and predicted with an accuracy of 0.778.It shows that the method in this paper can dig out the biologically significant hypertension pathogenic genes,achieve accurate prediction of pathogenic hypertension genes,and have important significance for the study of the molecular mechanism of hypertension and the prevention and treatment of hypertension in clinical medicine.