Introduction
Influenza infections differ greatly among individuals, with controlled human infection studies showing wide variation in viral load, immune activation, and symptom profiles. However, the biological and quantitative mechanisms underlying these differences remain unclear. Understanding how host and viral factors interact to drive this heterogeneity could provide insights relevant to both natural and experimental infections.

Methods
We used a data-driven, mechanistic modeling framework that integrates viral replication dynamics with immune responses in participants infected with influenza. The approach used digital twin and dimensionality reduction techniques to identify and simulate patterns of infection and immune control across individuals.

Results
The analysis identified distinct infection clusters arising from multivariate influences, and quantified the contribution of variability in inoculum size, virus infectivity, and baseline immunity. Important tradeoffs, such as between cell efficacy and expansion, were revealed, and host—pathogen interactions remained relatively consistent between primary infection and reinfection scenarios. Additional analyses of symptom data illustrated the model’s predictive value and revealed potential subjectivity that was independent of viral strain.

Conclusion
These findings illustrate the importance of mechanistic modeling in disentangling the complex determinants of influenza infection outcomes and suggest that individual-level variability may lead to shared patterns of disease resolution.