Multi-Scale Analysis of Lyme Disease Ecology

Abstract

Lyme disease is a zoonotic infectious disease. Increased public interest in Lyme disease has caused increased efforts by researchers for its surveillance and control. The main concept for this paper is to determine the mammalian species composition of areas at high risk for Lyme disease utilizing GIS-based (Geographic Information Systems) techniques coupled with k-means clustering, random forest, and multinomial logistic regression. Cluster analysis results were similar to previous work involving maps that display areas where people are at high risk for developing Lyme disease. There were differences in which mammal species presence had associations with Lyme disease risk observed at the two different scales within this analysis, with some overlap observed between the national scale and the smaller regions, as well as some overlap between the Rocky Mountain and Southeast regions that was not found at the national scale. This is an investigative analysis to determine which species are needed for habitat suitability analyses in efforts to prioritize vaccine deployment locations. There has been limited research on vaccine deployment for Lyme disease. Increasing our understanding of not only the vaccine but also the interactions between the components of disease transmission is necessary to control this infectious disease successfully.

Keywords

• Lyme disease
• risk modeling
• species composition
• vaccine deployment
• wildlife vaccination