基于生物信息学与机器学习筛选肥胖症关键基因OA

Objective:To explore biomarkers of human obesity diagnosis and their correlation with immune infiltration based on bioinformatics and machine learning methods.Methods:Human adipose tissue microarray data set was downloaded from Gene ex-pression omnibus(GEO)database.After analyzing GSE25401,GSE88837 and GSE94752 data sets in GEO and screening out differ-entially expressed genes(DEGs),KEGG pathway analysis,GO functional enrichment analysis were adopted,LASSO logistic regres-sion algorithm and support vector machine(SVM)algorithm were adopted to further screen key genes.An immune cell infiltration assay was performed to evaluate infiltration properties of 22 types of immune cells in human obesity patients and their association with hub genes.Receiver operating characteristic(ROC)curve was used to analyze diagnostic effect of characteristic genes.Bioinformatics analysis was conducted using R language(version 4.2.2),and threshold of significance was P<0.05.Results:A total of 190 DEGs were selected by limma software package,and 5 characteristic genes(PALLD,TF,CCL3,C6 and SCIN)were selected by LASSO re-gression and SVM algorithm.Through bioinformatics analysis,we discovered key role of these genes in immune microenvironment.ROC curves showed that the above 5 characteristic genes had good predictive and diagnostic effects on obesity.Conclusion:PALLD,TF,CCL3,C6 and SCIN are potential key genes and potential diagnostic biomarkers of human obesity,which may provide new targets for diagnosis and treatment of obesity.