Epidemiology

The Right Time and Place

+ See all authors and affiliations

Science  21 Sep 2012:
Vol. 337, Issue 6101, pp. 1434
DOI: 10.1126/science.337.6101.1434-b
CREDIT: ADAPTED FROM O. G. PYBUS ET AL., PROC. NATL. ACAD. SCI. U.S.A. 109 15066 (2012)

Spatial analysis, such as that used in epidemiology, is vulnerable to many errors, including that caused by spatial dependence or autocorrelation. For transmission of many infections, individuals need to be neighboring, and because the environment and behavior of nearby surroundings tend to be more similar than distant ones, neighbors will tend to be statistically dependent. This phenomenon causes complications for accurate epidemiological modelling. However, pathogens can also be mapped in time, as well as place, and evolutionary biologists have developed phylogeographic methods to aid this sort of historical sleuthing. Given current concerns about the unexpected emergence of pathogens such as West Nile virus (WNV) in the United States, Pybus et al. have merged these concepts to develop an alternative, less error-prone approach. Taking data for WNV, they show how the diffusion coefficient and variation in the spatial spread of a pathogen can be estimated from genome data alone. This approach revealed that instead of a steady front of east-to-west dissemination of WNV, it progressed in fact by rare long-range movements, probably triggered by bird migration or anthropogenic transport of mosquitoes. These rare events leave a detectable phylogenetic footprint. Ignorance of this hitherto hidden heterogeneity has led in turn to considerable overestimation of the pathogen's basic reproductive number, R0, a key parameter for estimating the epidemic potential of a pathogen.

Proc. Natl. Acad. Sci. U.S.A. 109, 15066 (2012).

Navigate This Article