ISSN print edition: 0366-6352 ISSN electronic edition: 1336-9075 Registr. No.: MK SR 9/7 Published monthly

Integrative analysis of miRNA in COVID-19 using machine learning and in-silico approaches

Figen Güzelgül, Erkan Öner, Serap Yalın, Ali Erdinç Yalın, Tuncay Yığıt, Ahmet Cemal Pazarlı, Marwa Abdelmageed, and Hacı Ömer Ateş

Department of Biochemistry, Faculty of Pharmacy, Tokat Gaziosmanpasa University, Tokat, Turkey

E-mail: figenguzelgul@gmail.com

Received: 27 March 2026  Accepted: 31 May 2026

Abstract:

The complex pathophysiology of coronavirus disease 2019 (COVID-19) and the quest for effective therapeutic targets remain critical challenges. While miRNA dysregulation is documented in severe cases, expression profiles during the acute phase of mild infections, particularly in home-recovery settings, remain poorly understood. In this study, key SARS-CoV-2 target genes were identified via gene network analysis. Protein–miRNA interactions were simulated using molecular docking and molecular dynamics (MD) simulations to evaluate binding stability. Subsequently, RT-PCR was employed to quantify miR-451a and miR-21-3p expression in blood samples from severe (n = 23) and mild (n = 22) COVID-19 patients, compared to a healthy control group (n = 24). Molecular docking analysis revealed that miR-451a exhibits a markedly stronger binding affinity for the SARS-CoV-2 main protease (Mpro; 6LU7) compared to miR-21-3p, yielding significantly more negative docking scores of − 298.01 kcal/mol and − 203.11 kcal/mol, respectively. Further in-silico analysis indicated that miR-451a interacts with key cytokines and the MAPK pathway. Clinical quantification confirmed that miR-451a expression was significantly down-regulated in severe patients compared to the control group (p = 0.01). Notably, miR-451a also showed significant differential expression in mild patients recovering at home, whereas miR-21-3p expression did not reach statistical significance. Integrated molecular docking, MD simulations, and principal component analyses (PCA) consistently suggested that miR-451a acts as a potent inhibitor of SARS-CoV-2 Mpro and a negative regulator of inflammation. Unlike previous studies predominantly focused on hospitalized or critical cases, this investigation provides the first evidence of significant miR-451a down-regulation in mild COVID-19 patients during the acute phase of home recovery. Our multi-faceted approach underscores the potential of miR-451a as a crucial regulatory biomarker and a promising candidate for future viral inhibition strategies, effectively highlighting the study’s novelty and clinical relevance.

Keywords: COVID-19; miRNA; Molecular docking; Gene network analysis

Full paper is available at www.springerlink.com.

DOI: 10.1007/s11696-026-05140-6

Chemical Papers 80 (10) 12453–12469 (2026)