Dynamic virulence decomposition disentangles host and pathogen contributions over the course of infection

Decomposing virulence into host and pathogen effects is key to understanding how virulence is generated during infection, and how host and pathogen contributions evolve over time. Conventional virulence decomposition is, however, not designed to quantify how these contributions vary throughout an infection. Building on conventional virulence decomposition, we developed a dynamic framework that integrates disease space analysis to quantify these changing contributions. We applied this framework to existing data from house finches ( Haemorhous mexicanus ) experimentally infected with the emerging bacterial pathogen, Mycoplasma gallisepticum , examining two biologically meaningful infection phases: an early phase dominated by innate immune responses and a later phase in which the adaptive immune response is fully activated. Our analyses revealed that virulence components evolved differently across infection phases, with pathogen exploitation, per-parasite-pathogenicity (PPP), and host resistance each exhibiting distinct trajectories. These findings corroborated several known evolutionary patterns, while also revealing the evolution of increased PPP in this system. In contrast, the apparent evolution of host tolerance inferred from aggregated peak measurements was not supported by phase-specific analyses, demonstrating that aggregated and phase-specific measures can provide different mechanistic insights into virulence. By linking conventional virulence decomposition with disease space analysis, our dynamic virulence decomposition framework provides a general approach for quantifying how host and pathogen processes shape virulence over the course of infection.