Juxtaposition between host population structures: implications for disease transmission in a sympatric cervid community

Abstract

Sympatric populations of phylogenetically related species are often vulnerable to similar communicable diseases. Although some host populations may exhibit spatial structure, other hosts within the community may have unstructured populations. Thus, individuals from unstructured host populations may act as interspecific vectors among discrete subpopulations of sympatric alternate hosts. We used a cervid-bovine tuberculosis (Mycobacterium bovis) system to investigate the landscape-scale potential for bovine tuberculosis transmission within a nonmigratory white-tailed deer (Odocoileus virginianus) and elk (Cervus canadensis) community. Using landscape population genetics, we tested for genetic and spatial structure in white-tailed deer. We then compared these findings with the sympatric elk population that is structured and which has structure that correlates spatially and genetically to physiognomic landscape features. Despite genetic structure that indicates the white-tailed deer population forms three sympatric clusters, the absence of spatial structure suggested that intraspecific pathogen transmission is not likely to be limited by physiognomic landscape features. The potential for intraspecific transmission among subpopulations of elk is low due to spatial population structure. Given that white-tailed deer are abundant, widely distributed, and exhibit a distinct lack of spatial population structure, white-tailed deer likely pose a greater threat as bovine tuberculosis vectors among elk subpopulations than elk.

Keywords: Cervidae; Discriminant Analysis of Principal Components; Mycobacterium bovis; community; disease transmission; population genetics; spatial Principal Component Analysis; sympatric populations.

Figure 1

Riding Mountain National Park (RMNP) and Duck Mountain Provincial Park and Forest (DMPP&F), Manitoba, CAN with spatial locations of bovine tuberculosis (bTB)-positive white-tailed deer (WTD; n = 11) and elk (n = 41) and from 1991 to 2010 illustrating apparent disease clustering in the Riding Mountain Region.

Figure 2

Discriminant Analysis of Principal Components for the 494 genotyped white-tailed deer from Riding Mountain National Park (2004–2006) revealed three genetically distinct (A) clusters on the landscape. However, as indicated by overlapping ellipses (B) clusters (red, green, blue) are sympatric.

Figure 3

Spatial information on genotyped white-tailed deer (A) analysis of global eigenvalue scores for the spatial principal component analyses (sPCA) performed on the 494 genotyped white-tailed deer (2004–2006). All plots are positioned according to their spatial coordinates. Shades of red represent genetic distance where the more related two individuals were the more similar their color; for example, an individual depicted as dark red is closer genetically to an individual that is a similar shade of red than an individual whom is depicted as light red. No clustering in shading revealed no discernible spatial patterning on the landscape. We contrast the absence of spatial structure in white-tailed deer against three elk clusters that have been extrapolated from radio-telemetry relocation data (n = 11 194 over n = 379 collared elk collected between 2002–2009) to illustrate the distribution of bounded subpopulations of elk ([B] calculated using 95% minimum convex polygons). Furthermore, spatial structure from radio-telemetry data correlates with spatial information on genotyped elk ([C] n = 312 at 30 microsatellites) from Riding Mountain National Park which occupy three spatially distinct clusters (Vander Wal et al. 2012).

Figure 4

Comparison of nearest-neighbor distances between bovine tuberculosis (bTB)-positive white-tailed deer and elk, illustrating some support that bTB-positive white-tailed deer may be more dispersed. Boxplots (A) (median, 25% and 75% quartiles, and 95% confidence intervals) illustrating that despite small sample size, white-tailed deer with bovine tuberculosis (n = 11) were more dispersed on the landscape than bovine tuberculosis-positive elk (n = 41). Nonoverlapping notches (‘> <‘) indicate that the two groups are likely significantly different (Chambers et al. 1983). This is confirmed by a Mann–Whitney U-test (χ = 4.94, P = 0.02, df = 1). A histogram (B) with smoothed density distribution (solid curve) illustrating that mean nearest-neighbor distances for white-tailed deer fall at the 90th percentile (dashed line) of 1000 iterations of an n = 11 bootstrap (without replacement) of bTB-positive elk mean nearest-neighbor distances.