Application Of Comprehensive Bioinformatics Analysis To Explore New Biomarkers Of Parkinson’s Disease

Objective:Application of comprehensive bioinformatics analysis to explore new biomarkers of Parkinson’s disease(PD).Methods:Download three PD data sets(internal training sets)in geo database and merge and preprocess them.Differentially expressed genes(DEGs)were obtained by differential analysis,significant module genes were obtained by weighted gene coexpression network analysis(WGCNA),and the intersection genes were obtained by the intersection of the genes obtained by the two analysis methods.The enrichment of go and KEGG was analyzed by intersection gene to understand their biological functions.The key genes were selected by stepwise regression analysis of intersection genes,and then the diagnostic prediction model was constructed by multivariate logistic regression analysis.Then the internal training set was analyzed by single sample gene set enrichment analysis(ssgsea),and the correlation between key genes and immune cells and immune function was further analyzed.Targeted prediction of key genes mi RNA(micro RNA)and transcription factor(TF).Finally,the external verification set verifies the results.Results:A total of 405 DEGs were obtained by difference analysis.In WGCNA analysis,8gene modules were identified,and 19 intersection genes were obtained by intersecting the genes in the most significant module with DEGs.The functional enrichment results of intersection genes are mainly concentrated in the regulation of nuclear transcription m RNA poly(a)tail shortening,the regulation of nuclear transcription m RNA catabolic process,the metabolic process of glutathione and its derivatives,the activity of glutathione transferase,the drug metabolism of cytochrome P450 and the metabolism of exogenous substances.Five key genes were screened by stepwise regression analysis,namely pnma3,aebp1,PABPC1,GCA and gstm2.The diagnostic prediction model constructed by multivariate logistic regression analysis has good resolution,and the correction curve,C index and AUC value are well verified.Ssgsea analysis shows that B lymphocytes(bcells)and plasma cell like dendritic cells(p DCs)may be potential core immune cells,and histocompatibility complex I(mhcclassi)may be potential core immune function.GCA has a significant negative correlation with p DCs,and pnma3 also has a significant positive correlation with bcells and p DCs.Mi RNA ‐ TF ‐ m RNA regulatory network analysis found that GCA was targeted by three mi RNAs and three TFs.Pnma3 is targeted and regulated by seven TFs.Jun and NFIA can regulate pnma3 and GCA at the same time.Conclusion:The diagnostic prediction models established by pnma3,aebp1,PABPC1,GCA and gstm2 have good diagnostic prediction ability.B cells and p DCs may be the potential core immune cells of PD,and MHC class I may be the potential core immune function of PD.GCA was negatively correlated with p DCs,pnma3 was positively correlated with B cells and p DCs.GCA was targeted by 3 mi RNAs and 3 TFs,pnma3 was targeted by 7 TFs,and Jun and NFIA could regulate pnma3 and GCA at the same time.These key genes,mi RNA and TF may become new biomarkers of PD in the future.