Identification of immune-related genes for the diagnosis of ischaemic heart failure based on bioinformatics and three machine learning models

Abstract

BackgroundThe role of immune cells in the pathogenesis of ischaemic heart failure (IHF) is well-established. However, identifying key diagnostic candidate genes in patients with IHF remains a challenge. Therefore, this study aimed to use bioinformatics and machine learning algorithms to identify potential diagnostic genes for IHF. MethodsTwo IHF datasets were obtained from the GEO database, and key genes for IHF were identified using Limma and WGCNA. Functional enrichment analysis was performed to explore the potential mechanisms of IHF. Next, we used three machine learning algorithms, namely LASSO, RF, and SVM-REF, to identify immune-related diagnostic genes for IHF. ssGSEA enrichment analysis was also completed. To assess the diagnostic value of the identified genes, we developed nomogram and validated them on additional GEO datasets. Finally, an immune infiltration analysis was conducted using the CIBERSORT algorithm to explore immune cell dysregulation in IHF. ResultsOur analysis yielded a total of 92 key genes associated with IHF. Enrichment analysis revealed that the mechanisms underlying IHF are mostly associated with immunity and inflammation. Using the machine learning algorithms, we identified four IHF diagnostic genes, namely RNASE2, MFAP4, CHRDL1, and KCNN3. We constructed nomogram and validated the diagnostic value of these genes on additional GEO datasets. The results showed that these four genes had high diagnostic value (AUC value of 0.961). Furthermore, our immune infiltration analysis revealed the presence of three dysregulated immune cells in IHF, namely Macrophages M2, Monocytes, and T cells gamma delta. ConclusionIn summary, our study identified four potential diagnostic candidate genes for IHF by using bioinformatics and machine learning algorithms. We also explored the potential molecular mechanisms of IHF and the immune cell infiltration environment of the failing myocardium. These findings provide new insights into the pathogenesis, diagnosis, and treatment of IHF.