Cross-Cohort Computational Inference of miRNA-mRNA Regulatory Programs from scRNA-seq in Convalescent Monocytes Associated with Prior COVID-19 Severity

Severe coronavirus disease 2019 (COVID-19) is characterized by acute immune dysregulation, with monocytes playing a central role in driving inflammation and disease severity. However, the transcriptional and post-transcriptional regulatory mechanisms underlying monocyte dysfunction in severe COVID-19 remain unexplored. In the study, an integrative analysis of paired bulk RNA-seq and miRNA-seq datasets was performed together with independent single-cell RNA-seq (scRNA-seq) data from convalescent individuals with a history of ICU or non-ICU COVID-19. Pooled cell proportions descriptively indicated a higher proportion of classical monocytes and lower proportions of non-classical monocytes, B cells and dendritic cells in individuals with a history of ICU disease; however, none of these differences was statistically significant in patient-level analyses after multiple-testing correction. Using the miRSCAPE framework, miRNA expression was inferred at single-cell resolution and identified distinct cluster-specific inferred miRNA expression patterns. Differential expression analysis of classical monocyte populations identified 284 nominally significant differentially expressed genes between convalescent ICU and non-ICU samples. Integration of miRNA-mRNA correlation analysis with experimentally validated interactions and independent assessment highlighted a focused regulatory network centered on ZMAT3, RHOB and HLA-DQA1. Host–pathogen interaction analysis identified database-supported SARS-CoV-2-host interactions involving ORF3a-RHOB and nucleoprotein-RHPN2, with additional host–host interactions connecting RHPN2, HLA-C and HLA-DQA1. Collectively, these findings provide a computational framework for investigating inferred miRNA associations of monocyte inflammatory pathways associated with prior COVID-19 severity and highlight regulatory interactions that warrant further experimental validation.