High-performance computing tools for modeling evolution in epidemics

Abstract

We describe a series of stepwise refinements of a biological model resulting in a high-performance simulation system for individual-based models of the co-evolutionary dynamics associated with spatially explicit epidemic processes. Our model includes two competing host species, a macroparasite capable of serving as a vector, and the vector-borne microparasite. Genetic algorithms are used to simulate genetic change; we are particularly interested in the evolution of pathogen virulence. The simulation system employs cellular automata to track individual organisms distributed over a two-dimensional lattice. Our models are able to identify each individual's parentage, and to account for both biotic and abiotic spatial heterogeneity. Using the developed system we conducted a series of experiments to demonstrate how individual-based modeling and explicit representation of space, although computationally expensive, can produce qualitatively new biological results.