Biomarker identification and inhibitor discovery for Marburg virus using machine learning driven virtual screening and molecular simulations

Marburg virus (MARV) results in severe hemorrhagic fever, with high morbidity and lacks approved antiviral therapies. The present research identifies differential genes (DEGs) involved in immune response pathways against the MARV, using bioinformatics and molecular modeling to explore new therapeutic targets. Cases were analyzed intensively using microarray data sets GSE58287 and GSE226148. DEGs were screened, followed by functional enrichment of the response in those DEGs. Important regulators use the analyses of protein-protein interaction networks. Screened compounds were further validated using the Lipinski Rule and absorption, digestion, metabolism, excretion, and toxicity (ADMET) profiling. The prediction of inhibitors was performed using machine learning for these targets, which was validated by ADMET profiling. Molecular docking and dynamics simulations were used to calculate the affinities of candidates for binding to STAT1 . A total of 1,179 DEGs were identified, with most of them associated with immune responses and antiviral defense mechanisms. Functional enrichment analysis indicated that the DEGs are involved in innate immunity and viral pathways, including influenza A, COVID-19, and hepatitis C. Network analysis revealed STAT1, IRF7, and CXCL10 as central regulators of the immune response. The Random Forest models predicted several potential STAT1 inhibitors, which were confirmed through molecular docking and MD simulations. This study sheds light on the role of immune-related pathways and regulators in MARV infection. STAT1 was identified as a target, with inhibitors discovered through the use of machine learning, docking, and simulations. These findings provide insights into MARV pathogenesis and support the development of targeted therapies for MARV and related immune disorders.