Abstract
BACKGROUND: Respiratory syncytial virus (RSV) is a leading cause of severe lower respiratory infections in children, yet biomarkers for assessing disease severity remain limited. Herein, we investigated the differential expression biomarkers between RSV infected hospitalized patients, healthy groups and RSV infected outpatients. METHODS: Two publicly available transcriptomic datasets (GSE77087 and GSE188427) were retrieved from the Gene Expression Omnibus (GEO) database. The GSE77087 dataset comprised peripheral blood samples from 81 children with confirmed RSV infection (61 hospitalized and 20 outpatient) and 23 healthy controls. The GSE188427 dataset included 147 RSV-infected children (113 hospitalized and 34 outpatient) and 51 healthy controls. Genes with |log2 fold change (logFC)| > 0 and false discovery rate (FDR) < 0.05 were considered differentially expressed. Overlapping DEGs between the two datasets were identified using the VennDiagram package. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted on the intersecting DEGs via the clusterProfiler package, with terms deemed significant at FDR < 0.05.The CIBERSORT algorithm was applied to estimate the relative proportions of 22 immune cell types in 228 RSV-infected samples. Potential drug interactions for hug genes were predicted using the Drug-Gene Interaction Database (DGIdb). Competing endogenous RNA (ceRNA) networks were constructed using the SpongeScan database to identify lncRNAs interacting with the target miRNAs. Networks were visualized using Cytoscape (v3.10.1).Finally, Machine Learning-Based Biomarker Selection and hub gene identification and validation. RESULTS: Differential gene expression analysis revealed 81 overlapping genes between hospitalized and outpatient RSV-infected children. Machine learning models, particularly SVM (area under the curve, AUC = 0.950), prioritized CD79A and GADD45A as key predictors of hospitalization. CD79A was significantly downregulated in severe cases, correlating with impaired B-cell responses and cytotoxic immunity, while GADD45A, upregulated in severe infections, linked to oxidative stress and neutrophil-driven inflammation. Immune cell profiling highlighted distinct infiltration patterns, with severe cases showing elevated naïve B cells and M0 macrophages versus activated NK cells and M1 macrophages in mild cases. Clinical validation in 92 children confirmed CD79A suppression and GADD45A elevation in severe RSV (p < 0.001), aligning with younger age, lower weight, and respiratory distress. Functional enrichment implicated endoplasmic reticulum stress and neutrophil extracellular traps in disease progression. Drug-target predictions and ceRNA networks further revealed therapeutic potential. CONCLUSION: These findings establish CD79A and GADD45A as clinically actionable biomarkers for RSV severity, offering insights into immune dysregulation and guiding personalized management strategies.