Modeling rapidly-evolving viral infections within hosts generates lifelike infection classes

Viruses, especially RNA viruses, are subject to high levels of mutation, leading to the generation of closely-related offspring ("quasispecies") in diseases ranging from polio to COVID-19. Viral diseases arise from a wide range of infection types, from opportunistic (e.g., CMV) to chronic (e.g., Hepatitis C) to acute (e.g., Flu). How can host-quasispecies interactions lead to different types of diseases? In this work we model within-host infections of many cells by quasispecies, subject to differing levels of host immunity, ease of viral entry (permissivity), and other pressures. Using simulations, we study immunity-permissivity space and find that it consists of three different phases separated by a transition of varying order and crossover regions. Examination of the steady-state and dynamic properties of the phases suggests a natural mapping of each phase onto a different class of infection. We also find interesting physical behavior near the transition, including the presence of multiple and distinct quasispecies, as well as glassy dynamics. Using perturbation theory, we derive steady-state probability distributions and viral loads for very low permissivity that validate our simulation results.