Identifying the next Zika: An analysis of zoonotic potential in Flaviviridae

      Purpose: The viral family Flaviviridae contains several viruses known to cause human disease and recent widespread epidemics (e.g. Zika, West Nile, Dengue, and Yellow Fever viruses). Yet, many viruses in the family are little studied and the ecological, viral, and genomic factors that contribute to zoonotic potential are unknown.
      Methods & Materials: Through an extensive review of the literature we compiled all known vertebrate hosts and arthropod vectors for each of the 63 ICTV-recognized (2015) viruses in the family Flaviviridae. For each host and vector species, we compiled ecological and life history traits and geographic range information where available. Viral genomes for 40+ available viral species were downloaded and phylogenies were reconstructed for the following genes with potential involvement in expanded host range and zoonotic potential: envelope (E), premembrane (prM), capsid (C), NS1, NS2AB, NS3, NS4AB, and NS5. A novel approach utilizing Generalized Additive Models and a combination of ecological and phylogenetic traits was used to identify ecological and genetic factors that predict whether or not a virus is zoonotic.
      Results: We found 40/63 (64%) of ICTV recognized Flaviviridae species are zoonotic -- with evidence of infection in humans and at least one other vertebrate host -- with a mean of 21 (range 1-413) known host species per virus. Including a control for research effort, the phylogenetic breadth of non-human hosts was a significant predictor of whether or not a virus is zoonotic. We map the natural geographic ‘hotspots’ of flavivirus diversity based on the distribution of known hosts and also show non-random patterns in the distribution of zoonoses across the each gene specific viral phylogeny.
      Conclusion: These analyses can be used to target future surveillance for flaviviruses to the host species and geographic regions most likely to harbor them. We show that viral and ecological factors can predict zoonotic potential in this important group of human pathogens, and identify specific genetic factors of interest for further characterization to better understand zoonotic potential.