Elucidation of potential miRNAs as prognostic biomarkers for coronary artery disease

Coronary Artery Disease (CAD) is a cardiovascular disorder characterized by narrowing of arteries due to metabolic dysregulations, which severely impedes blood flow through cardiac tissues. Genetic factors significantly contribute to the susceptibility of CAD and miRNAs play a crucial role in gene expression and regulation. In this study, we aim to identify highly specific miRNA-mRNA interactions and gene targets by employing machine learning approaches such as association rules mining (ARM) and singular value decomposition (SVD), followed by differential expression analysis of microarray datasets. For this, genes associated to CAD and its lethal sequelae (valvular heart disease, fibrosis, atherosclerosis, hyperlipidemia, oxidative stress and inflammation) were obtained from databases i-e., National Center for Biotechnology Information, Genetic Testing Registry (GTR). Highly conserved miRNAs were selected using bioinformatics repositories TargetScan, miRBase, and miRanda. Furthermore, ARM and SVD were utilized to discover significant association patterns and frequently occurring miRNAs. For the validation of hub miRNAs, differential expression analysis was carried out on two independent cohorts of miRNA expression datasets of cardiac patients. This integrated approach identified 3 hub miRNAs (miR-200a-3p, miR-32-5p and miR-92-3p). Functional enrichment analysis revealed their involvement in diabetes, cholesterol metabolism, inflammation, and atherosclerosis. Moreover, disease enrichment analysis showed their association with heart diseases, vascular diseases, and endothelial dysfunction. Conclusively, this is the first study that employed ARM and SVD approaches to identify hub miRNAs and novel gene targets involved in CAD. The identification of these miRNAs as putative biomarkers may lead to a more accurate prognostic score for early detection of CAD.