Stochastic modeling of the impact of random forcing on persistent hepatitis B virus infection

We developed a computational methodology for the analysis of the impact of random forcing on the patterns of virus persistence in HBV infection. We examine the issue of robustness vs. sensitivity in models of chronic infections (which relates to a fundamental question in immunology). The practical details of the implementation of the stochastic ODE models in the analysis of spontaneous recovery are presented. These include the effect of sampling of the parameter space, the number of simulation runs needed for a robust estimation of the mean and the variance of the spontaneous recovery pattern, the impact of the noise intensity and type on the response of the models. The stochastic models indicate that the relative frequency of spontaneous recovery from the persistent HBV infection is rather low, less than 10%, for the random forcing intensity varying within three orders of magnitude. The persistent infection with low viral load/CTL abundance is driven to recovery more frequently than the high viral load/CTL number type once the intensity of the random perturbations exceeds a certain level.