Indentifying and modeling the distribution of cryptic reservoirs of Ebola virus using artificial intelligence

Abstract

Fruit bats (Megachiroptera) have been found to be the principal reservoirs of Ebola virus (EBOV) to humans. However, bats do not appear to be the primary reservoir in the environment and between outbreaks. The cryptic reservoir species of EBOV and its distribution have not been identified. The purpose of the study was to identify the most likely cryptic reservoir species of EBOV and the probable distribution of cryptic reservoir species where EBOV could be maintained in Sierra Leone. The Bioagent Transport and Environmental Modeling System (BioTEMS) was used to analyze mammals, arthropods, plants and protists in order to identify the most likely species to be the cryptic reservoir for EBOV. ArcGIS and BioTEMS were used to determine the probable distribution of cryptic reservoir species. BioTEMS identified free-living pathogenic amoebae (FLPA) as the probable cryptic reservoir species (Test Performance = 93.3). Diptera in the order Chrysops were also identified as possible secondary reservoirs and mechanical vectors of EBOV. Distribution of likely hot spots for FLPA and phytotelmata/tree-holes were identified in several regions of Sierra Leone, primarily in the southeast and are similar to those predicted by other authors, but at a much higher resolution (15 m for BioTEMS verses up to 5 km in other studies). Water-filled cavities (phytotelmata), specifically tree-holes, were identified as the most likely sites for the cycle of transmission to occur among FLPA and susceptible secondary reservoirs. Free-living pathogenic amoebae are not only pathogenic to humans and animals but they serve as reservoirs and Trojan horses of infection as well. Identifying what and where cryptic reservoirs of EBOV persist between outbreaks provides an opportunity for the first time to conduct environmental epidemiologic surveillance to mitigate outbreaks and to test anti-microbial delivery systems such as the ProVector® to reduce EBOVand FLPA.Keywords: Filovirus, Amoeba, Epidemiology, Machine Learning, Vector, Disaster Management