Abstract:Background Myocardial infarction (MI) is a major contributor to global mortality, and microRNAs (miRNAs) are important in its pathogenesis. Identifying blood miRNAs with clinical application potential for the early detection and treatment of MI is crucial. Methods We obtained MI-related miRNA and miRNA microarray datasets from MI Knowledge Base (MIKB) and Gene Expression Omnibus (GEO), respectively. A new feature called target regulatory score (TRS) was proposed to characterize the RNA interaction network. MI-related miRNAs were characterized using TRS, transcription factor (TF) gene proportion (TFP), and ageing-related gene (AG) proportion (AGP) via the lncRNA-miRNA-mRNA network. A bioinformatics model was then developed to predict MI-related miRNAs, which were verified by literature and pathway enrichment analysis. Results The TRS-characterized model outperformed previous methods in identifying MI-related miRNAs. MI-related miRNAs had high TRS, TFP, and AGP values, and combining the three features improved prediction accuracy to 0.743. With this method, 31 candidate MI-related miRNAs were screened from the specific-MI lncRNA-miRNA-mRNA network, associated with key MI pathways like circulatory system processes, inflammatory response, and oxygen level adaptation. Most candidate miRNAs were directly associated with MI according to literature evidence, except hsa-miR-520c-3p and hsa-miR-190b-5p. Furthermore, CAV1 , PPARA and VEGFA were identified as MI key genes, and were targeted by most of the candidate miRNAs. Conclusions This study proposed a novel bioinformatics model based on multivariate biomolecular network analysis to identify putative key miRNAs of MI, which deserve further experimental and clinical validation for translational applications.

Myocardial infarction unveiled: Key miRNA players screened by a novel lncRNA-miRNA-mRNA network model