Background: Ischemic cardiomyopathy (ICM) with high mobility and mortality is closely linked to immunology, oxidative stress, inflammatory response and so on. Early diagnosis counts for the effective treatment of ICM. However, there are still no distinctive diagnostic signatures. This research aims to investigate effective signatures and build the diagnostic model for ICM.
Methods: The Gene Expression Omnibus was used to retrieve the microarray data of GSE9800 and GSE580, which were obtained from tissue biopsy samples. Differentially expressed genes (DEGs), GO, and KEGG analyses were then carried out on the microarray data. The PPI network was constructed via STRING database. Following that, CIBERSORT techniques in conjunction with the LM22 feature matrix were used to assess the immune infiltration of the samples. The expression of a few chosen genes served as the predictor variable, and the occurrence of ICM served as the responder variable, in the construction of the best subset stepwise regression model.
Results: A total of 28 DEGs were found. And according to the GO and KEGG studies, numerous biological processes were enriched. Patients with ICM and their normal counterparts had considerably distinct immune cell types infiltrating. For the construction of the PPI network, the top 20 most significant DEGs were selected and were used to build the original regression model. The optimal subset screened using stepwise regression analysis contained three pivotal genes (SCD, SNX25, WNT7B) and the area under the curve (AUC) values in this model was 0.891.
Conclusion: We identified several possible hub genes, including SCD, SNX25, and WNT7B, which may be strongly related to the development of ICM. Based on the three genes, the logistic regression model could be used to accurately diagnose ICM patients.
Keywords: ischemic cardiomyopathy, ICM, immunology, inflammatory responses, diagnosis, differently expressed genes, bioinformatic analysis, the optimal subset stepwise regression