DOI: 10.1111/irv.70328 ISSN: 1750-2640

A Mechanistic Viral Host Perturbation Score (VHPS) Reveals Shared and Virus‐Specific Transcriptional Programs Across Human Respiratory Infections

Sabri Saeed Sanabani

ABSTRACT

Background

Acute viral infections trigger overlapping but virus‐specific blood transcriptional responses; however, current analyses rarely quantify the mechanistic depth of activation across pathways, regulators, and immune cell compositions.

Methods

RNA‐seq data from GSE157240 were reanalyzed to infer transcription factor (TF) activity using DoRothEA/decoupleR, signaling pathway activity using PROGENy, and stress pathway and transcriptional perturbation breadth metrics. These four components were integrated into a viral host perturbation score (VHPS), a proof‐of‐concept framework designed to quantify the magnitude of host transcriptional perturbation. Differential expression analysis, Gene Ontology enrichment, module scoring, and CIBERSORTx‐based immune deconvolution were used to compare respiratory and systemic viral infections.

Results

VHPS‐stratified viruses along an axis of host transcriptional perturbation, with DNA viruses showing the highest median perturbation, followed by dengue and enterovirus/rhinovirus, whereas influenza A and other respiratory RNA viruses were intermediate and healthy controls were lowest. Upregulated genes revealed virus‐specific and shared signatures, and immune/stress GO terms showed stronger T cell and interferon enrichment for respiratory viruses. A 26‐gene “shared respiratory antiviral” module correlated with VHPS and TF/stress components but not PROGENy activity. Immune deconvolution revealed virus‐specific leukocyte remodeling, with marked Treg depletion and elevated neutrophils in respiratory viruses, monocyte skewing in dengue, and the lowest Treg fractions in DNA viruses.

Conclusions

Together, VHPS and immune deconvolution delineate shared and virus‐specific regulatory programs in acute viral infections, but the generalizability of VHPS requires evaluation in independent datasets.

More from our Archive