Single-cell atlas of transcript usage remodelling in antiviral immune responses across human populations

Humans exhibit substantial interindividual variation in their immune responses to infection, yet the contribution of transcript usage --the relative abundance of gene isoforms-- to this variation remains poorly understood. Here, we generate the first single-cell atlas of transcript usage variation during early responses to influenza A virus and SARS-CoV-2 across 160 individuals of African and European ancestry. We show that viral stimulation induces widespread remodelling of transcript usage across all major immune lineages, with changes that are largely lineage-restricted and frequently undetected at the gene expression level. We further find that ancestry-associated effects on transcript usage are predominantly cell type-specific and contribute to population differences in antiviral responses. In addition, the genetic regulation of transcript usage during stimulation differs between influenza A and SARS-CoV-2, pointing to virus-dependent regulatory architectures. Together, our findings establish transcript usage as a dynamic regulatory layer shaping responses to viral infection across immune cell types and human populations, providing new insights into the molecular basis of variation in human antiviral immunity.