Machine learning-driven exploration of therapeutic targets for atrial fibrillation-joint analysis of single-cell and bulk transcriptomes and experimental validation.

Background: To explore new therapeutic targets and strategies for atrial fibrillation (AF) by analyzing gene expression profiles of AF patients using machine learning techniques combined with transcriptomic data, and to uncover the potential molecular mechanisms underlying AF. Methods: Transcriptomic datasets associated with AF were obtained from the GEO database. After batch effect removal and normalization, differential gene expression analysis was performed to identify differentially expressed genes (DEGs). Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Disease Ontology (DO) enrichment analyses were conducted to explore the functions and pathways of these DEGs. Three machine learning algorithms, Least Absolute Shrinkage and Selection Operator (LASSO), Support Vector Machine-Recursive Feature Elimination (SVM-RFE), and random forest (RF), were applied to screen key genes related to AF. A nomogram model was developed based on the identified key genes, and its diagnostic performance was evaluated. Single-cell transcriptome analysis was performed to investigate the cell-type-specific expression patterns of these key genes. Finally, Real-time PCR (RT-qPCR) and western blot (WB) analyses was performed on right auricular tissue from patients with atrial fibrillation and control samples. Results: A total of 64 DEGs were identified, including 27 upregulated and 37 downregulated genes. Enrichment analyses revealed that these genes were involved in biological processes such as positive regulation of muscular systemic processes, immune responses, and calcium signaling pathways. Three machine learning algorithms identified six key genes for AF. The nomogram model based on these six genes demonstrated excellent diagnostic performance with an AUC of 0.97. Single-cell transcriptome analysis showed specific expression patterns of these key genes in different cell types. Additionally, immune infiltration analysis indicated changes in the immune microenvironment in AF patients. qPCR and WB analyses also indicated that the differences in mRNA and protein expression levels of these six molecules between the control group and the atrial fibrillation group were consistent with the results of transcriptome analysis. Conclusion: This study provides new insights into the molecular mechanisms of AF and offers potential non-invasive biomarkers for AF diagnosis. The identified key genes and constructed model may facilitate the development of targeted therapies for AF.