Antagonistic coinfection in rock-paper-scissors models during concurrent epidemics

TL;DR

Addresses how two concurrent epidemics interact within a spatial rock-paper-scissors framework by introducing antagonistic coinfection governed by mortality-reduction factors $\gamma_1$ and $\gamma_2$ under a Moore-neighborhood May-Leonard dynamic. The study shows that increasing antagonism lowers mortality for coinfected hosts, increases cure probability, and boosts species densities while shrinking the characteristic length scale of spatial domains. Mobility restriction further lowers infection and selection risks and can boost organisms' life expectancy by up to $54\%$ under total antagonism, indicating a synergistic potential between behavioral and epidemiological interventions. Together, these findings provide mechanistic insights for managing concurrent epidemics in complex ecological networks.

Abstract

We investigate the dynamics of dual disease epidemics within the spatial rock-paper-scissors model. In this framework, individuals from all species are equally susceptible to infection by two distinct pathogens transmitted via person-to-person contact. We assume antagonistic mortality, where the simultaneous occurrence of coinfection reduces the probability of host mortality due to complications arising from either coexisting disease. Specifically, we explore two scenarios: global antagonism, where the presence of one pathogen inhibits the progression of the other in coinfected hosts, and uneven antagonism, where only one pathogen affects the development of the other. Using stochastic simulations, we show that the characteristic length scale of the spatial patterns emerging from random initial conditions diminishes as antagonism becomes more significant. We find that antagonism enhances species population growth and reduces the average probability of healthy organisms becoming infected. Additionally, introducing individuals' mobility restrictions significantly decreases both organisms' infection risk and selection pressures. Our results demonstrate that combining mobility restrictions with antagonistic coinfection can increase organisms' life expectancy by up to $54\%$. Our findings show that integrating antagonistic coinfection and mobility restriction strategies into ecological models may provide insights into designing interventions for managing concurrent epidemics in complex systems.

Figures (10)

• Figure 1: Illustration of the rock-paper-scissors model. Selection interactions are represented by arrows denoting the dominance of organisms of species $i$ over individuals of species $i+1$.
• Figure 2: Snapshots of the rock-paper-scissors model with antagonistic coinfection. These snapshots capture the spatial organisation of organisms on a lattice with $300^2$ grid sites, evolving over 3000 generations. Figure \ref{['fig2a']} presents the initial random conditions, while Figs. \ref{['fig2b']}, \ref{['fig2c']}, and \ref{['fig2d']} showcase the spatial distribution of individuals at the end of Simulations A ($\gamma=0.0$), B ($\gamma=0.5$), and C ($\gamma=1.0$), respectively. Purple, yellow, and pink dots represent individuals of species $1$, $2$, and $3$, while light blue dots indicate empty spaces.
• Figure 3: Temporal variation of species densities in Simulations A, B, and C, depicted in the snapshots shown in Figs. \ref{['fig2b']} (brown line), \ref{['fig2c']} (cyan line), and \ref{['fig2d']} (red line).
• Figure 4: Autocorrelation function for organisms of the same species for various coinfection antagonistic factors. The results were averaged over 100 simulations conducted on lattices with $500^2$ grid sites until $t=5000$ generations. The error bars show the standard deviation. Brown, cyan, and red lines correspond to the results for $\gamma=0.0$, $\gamma=0.5$, and $\gamma=1.0$, respectively. The inset figure illustrates the characteristic length scale as a function of the coinfection antagonistic factor. The horizontal dashed line indicates the threshold assumed to calculate the characteristic length scale shown in the inner panel
• Figure 5: Cure probability as a function of the antagonistic factor. The results were averaged over $100$ simulations running in lattices with $500^2$ grid sites until $t=5000$ generations. The error bars show the standard deviation. The orange line illustrates the scenario where the antagonistic coinfection impacts the mortality of both diseases. In contrast, the purple line represents the asymmetric case, where only the mortality of disease $1$ is influenced.