Tissue-specific cellular architecture shapes antiviral transcriptional programs in porcine PRRSV and ASFV infections

Background Porcine reproductive and respiratory syndrome virus (PRRSV) and African swine fever virus (ASFV) cause severe disease in pigs and represent major threats to global swine production. Bulk transcriptomic analyses have been widely used to characterize host responses to these infections; however, interpretation of bulk RNA sequencing (RNA-seq) signals remains limited because transcriptional programs arise from heterogeneous cellular compartments within infected tissues. In this study, we developed a cell-type–resolved transcriptomic framework integrating single-cell, bulk, and spatial transcriptomic data to improve the interpretation of host response programs in porcine viral infections. Results Using bronchoalveolar lavage fluid (BALF) single-cell RNA sequencing data, we constructed a weighted six–cell-type signature matrix representing epithelial cells, macrophages, dendritic cells, B cells, T cells, and natural killer (NK) cells. Application of this reference to bulk RNA-seq datasets from PRRSV and ASFV infections across lung and mandibular lymph node tissues enabled estimation of cell-type contributions and attribution of transcriptional programs to their underlying cellular compartments. Program-level analysis revealed that apparent similarities in bulk antiviral signatures mask distinct cellular architectures across tissues. While both infections exhibited interferon-associated transcriptional responses at the bulk level, cell-type–resolved attribution showed that these programs arise from different cellular compartments depending on tissue context. In lung tissues, transcriptional programs were primarily associated with epithelial and macrophage compartments, whereas lymphoid tissues showed stronger contributions from B cell and dendritic cell programs. Projection of the same signatures onto porcine Peyer’s patch spatial transcriptomic data recapitulated canonical intestinal immune microanatomy, including epithelial surface dominance and B cell–rich follicular regions, providing qualitative anatomical validation of the inferred cell-type programs. Conclusions These findings demonstrate that tissue-specific cellular architecture is a major determinant of antiviral transcriptional programs in porcine viral infections. Integrating single-cell, bulk, and spatial transcriptomic data enables cell-type–resolved interpretation of host response signals and provides a framework for improving biological interpretation of transcriptomic studies in veterinary infectious diseases.