Uncovering key biomarkers, potential therapeutic targets and development of deep learning model in heart failure

Heart failure (HF) represents a significant public health concern, characterized by elevated rates of mortality and morbidity. Recent advancements in gene sequencing technologies have led to the identification of numerous genes associated with heart failure. By utilizing available gene expression data from the Gene Expression Omnibus (GEO) database, we conducted a screening for differentially expressed genes (DEGs) related to heart failure. Key genes were identified through intersection analysis in conjunction with weighted gene co-expression network analysis (WGCNA). Following this, we pinpointed four essential genes (ITIH5, ISLR, ASPN, and FNDC1) by employing functional enrichment analyses, machine learning approaches, protein-protein interaction (PPI) assessments, gene set enrichment analysis (GSEA), and immune infiltration evaluations. Additionally, a novel diagnostic model for heart failure was successfully developed using a deep learning convolutional neural network (CNN), and its diagnostic performance was validated within public datasets. Analysis via single-cell RNA sequencing further indicated stable up-regulation patterns of these genes across various cardiomyocyte types in HF patients. Moreover, the exploration of drug-protein interactions revealed two potential therapeutic drugs targeting the identified key genes, with molecular docking offering a feasible pathway for this connection. In conclusion, we identified four potential key biomarkers closely related to HF and two possibly effective small molecules, which provide significant insights into the molecular mechanisms underlying heart failure and the search for new therapeutic targets.