Moving toward targeted arbovirus control: using the virome to determine Aedes dispersal and population structure

PROJECT SUMMARY / ABSTRACT 3.9 billion people are estimated to be at risk of dengue and other viruses transmitted by Aedes aegypti mosquitoes. Without effective therapeutics, disease management relies on mosquito control. Optimizing the deployment and efficacy of mosquito control requires understanding key drivers of population structure and dispersal. Population structure is known to affect the spread of both beneficial (e.g., gene drives) and detrimental genes (e.g., insecticide resistance alleles), and to affect the probability of reintroduction or resurgence following local elimination. Because Ae. aegypti is a weak flyer, dispersal is comprised of natural movement across very fine spatial scales (~100-200m) punctuated by long-distance dispersal events due to human movement. However, current approaches used to characterize mosquito population structure and movement are limited. Genomic approaches (e.g., mtDNA) can estimate connectivity between populations on evolutionary timescales and across large spatial scales. In contrast, while close-kin mark recapture methods (CKMR) can characterize movement on fine-spatial scales, this method cannot logistically be scaled up to even intermediate spatial scales (e.g. city districts). In this application, we propose a novel, highly innovative approach that will yield a unified framework to estimate mosquito population structure and dispersal rate across varying spatial scales to inform control. Recent studies have shown that Ae. aegypti is host to more than 27 unique viruses. We aim to leverage genetic variation in insect specific viruses to identify the relevant drivers of Ae. aegypti dispersal and population structure at regional (AIM 1) and local scales (AIM 2). We will estimate population structure and dispersal at each scale using genetic diversity within the virome or within individual viruses and will compare our results with current gold standard techniques that estimate mosquito connectivity from variation in the mosquito genome. This work will be done on St. Kitts where Ae. aegypti is abundant island-wide and has driven substantial arbovirus outbreaks. To determine Ae. aegypti population structure at regional scales (AIM 1), we will sample Ae. aegypti at 25 sites across five different landcovers monthly in two consecutive months each season, using metrics common to landscape genomics to quantify population structure based on variation within the Ae. aegypti genome and virome. In addition, we will develop phylogeographic models of four target viruses that are commonly found in Ae. aegypti. We will then compare our results to predictions from theoretical models of dispersal on the island. To estimate dispersal on local to intermediate scales (AIM 2), we will compare CKMR methods to our approach that estimates dispersal using phylogeographic models based on target viruses in mosquitoes sampled from a transect across an urban center, Basseterre. We predict Ae. aegypti dispersal will be dominated by human-mediated dispersal at the regional scale, with natural dispersal occurring between nearby, environmentally similar sites. We believe that the use of phylogeographic models of target viruses will provide a single method capable of estimating dispersal and population structure across multiple spatial scales.