Asymmetric disease dynamics in multihost interconnected networks

TL;DR

This paper develops an analytic framework to understand how asymmetries in infection probabilities across different hosts influence epidemic spread in multilayer networks, revealing complex interactions affecting outbreak risk and size.

Contribution

It introduces a novel analytic model for multihost epidemic dynamics that accounts for asymmetry in infection probabilities and validates it with simulations.

Findings

Outbreak probability depends on source and inter-host infection asymmetry.

Outbreak size is mainly influenced by non-focal host to focal host infection probability.

Asymmetry in infection probabilities significantly shapes multihost disease dynamics.

Abstract

Epidemic spread in single-host systems strongly depends on the population's contact network. However, little is known regarding the spread of epidemics across networks representing populations of multiple hosts. We explored cross-species transmission in a multilayer network where layers represent populations of two distinct hosts, and disease can spread across intralayer (within-host) and interlayer (between-host) edges. We developed an analytic framework for the SIR epidemic model to examine the effect of (i) source of infection and (ii) between-host asymmetry in infection probabilities, on disease risk. We measured risk as outbreak probability and outbreak size in a focal host, represented by one network layer. Numeric simulations were used to validate the analytic formulations. We found that outbreak probability is determined by a complex interaction between source of infection and between-host infection probabilities, whereas outbreak size is mainly affected by the non-focal host to focal host infection probability alone. Hence, inter-specific asymmetry in infection probabilities shapes disease dynamics in multihost networks. These results expand current theory of monolayer networks, where outbreak size and probability are considered equal, highlighting the importance of considering multiple measures of disease risk. Our study advances understanding of multihost systems and non-biological systems with asymmetric flow rates.