COS 61-2 - Predicting the zoonotic vector status of Dermacentor ticks: A machine learning approach

Wednesday, August 14, 2019: 1:50 PM
L011/012, Kentucky International Convention Center

ABSTRACT WITHDRAWN

Jessica Martin, Biology, Stanford University, Stanford, CA and Barbara Han, Cary Institute of Ecosystem Studies, Millbrook, NY
Jessica Martin, Stanford University; Barbara Han, Cary Institute of Ecosystem Studies

Background/Question/Methods

Ticks are second only to mosquitoes as vectors of zoonotic disease. In the U.S., cases of tick-borne diseases more than doubled over the past 13 years. Most research has focused on Ixodes ticks, including a recent study published by Yang and Han (2018). They utilized a machine learning technique, generalized boosted regression, to predict the zoonotic vector status of Ixodes ticks based on traits of the tick species (e.g. morphology, life history, geographic distribution, and host specificity). Dermacentor ticks have received comparatively little attention despite their role as vectors of zoonotic diseases.

Here, we assess which traits of Dermacentor tick species serve as predictors for zoonotic disease vector status. To address this question, we collected data from the primary literature on morphological characteristics (body size, etc.), life history characteristics, geographic distribution, habitat specificity, and host specificity. We then utilized the Global Infectious Disease and Epidemiology Network (GIDEON) database to determine which Dermacentor species serve as vectors of zoonotic diseases.

Results/Conclusions

In addition to constructing a similar database to Yang and Han’s study, we took a more in-depth approach by assessing habitat specificities (including biogeographic realms, ecotypes, and climate classifications) as well as host specificity (utilizing the Std index and di, a network analysis metric). Preliminary results indicate that there is significant variation in morphological characteristics, habitat specificity, and host specificity among Dermacentor ticks. Using the GIDEON database, we found that some species of Dermacentor ticks vector Borrelia (which causes Lyme disease), Rickettisa, and Hemorrhagic fever viruses. The GIDEON database included 14 species of Dermacentor ticks that are known vectors of zoonotic diseases.

In order to predict which traits of Dermacentor tick species are the most important for determining their vector status, we will next utilize a machine learning approach, generalized boosted regression. The results of our study will allow us to predict which recently described and little-studied Dermacentor ticks may serve as additional vectors of zoonotic diseases. Our results have the potential to inform which tick species should be closely monitored because of their vector status or potential, as well as the best management practices for Dermacentor tick species, and how these may differ from Ixodes ticks.