Diagnostic Value of Oxidative Stress-Related Features Mined by WGCNA and Machine Learning in Respiratory Syncytial Virus Infection

Abstract

Respiratory syncytial virus (RSV) is the leading cause of acute lower respiratory tract infections in infants and young children, with insufficient early diagnostic biomarkers and unclear molecular mechanisms. Oxidative stress (OS) is central to RSV-associated immune disorders and tissue damage. GEO datasets (GSE105450, GSE77087) were used to screen differentially expressed genes (DEGs); OS-related DEGs were identified via WGCNA. Four machine learning algorithms screened core diagnostic genes, with in vitro validation (A549/BEAS-2B cells) by qRT-PCR, shRNA-ID3 silencing, immunofluorescence and ELISA. ID3, MATK and ZMAT3 were potential diagnostic biomarkers (training AUC = 0.859, validation AUC = 0.725), upregulated in RSV-infected samples, correlated with immune cell infiltration, and involved in AKT/ERBB pathways. RSV upregulated ID3; ID3 knockdown promoted RSV replication and exacerbated OS. ID3, MATK and ZMAT3 are potential RSV diagnostic biomarkers; ID3 exerts host protection by regulating OS and inhibiting RSV replication, providing new targets for early diagnosis and intervention.