Characterizing host-virus associations is critical due to the rising frequency of emerging infectious diseases originating from wildlife. Past analyses have evaluated zoonotic risk as binary, but virulence, transmissibility, and death burden can vary dramatically. Recent work suggests bats harbor more viruses with high virulence in humans than other taxa. However, it remains unknown whether all bats harbor viruses of equal viral epidemic potential. We used phylogenetic factorization to flexibly identify clades of mammals (at any taxonomic level) associated with low or high viral epidemic potential. We found that virulence, transmissibility, and death burden only cluster within specific bat clades, often composed largely of cosmopolitan families. Mapping the geographic distributions of these bat clades with anthropogenic footprint data suggests high viral epidemic potential in coastal South America, Southeast Asia, and equatorial Africa. Our results deepen understanding of the host-virus network and identify clades to prioritize for viral surveillance, risk mitigation, and future studies characterizing mechanisms of viral tolerance.
Due to the rising frequency of emerging infectious disease outbreaks in humans, there is an urgent need to characterize associations between viruses and their hosts. Over 70% of human emerging infectious diseases are caused by zoonotic pathogens, which originate in animals and adversely impact human health and economies. Statistical models can improve our capacity to preempt and mitigate zoonotic transmission by identifying the kinds of wildlife species most likely to harbor zoonotic pathogens. For example, past work identified rodent and avian species likely to be undetected reservoirs of zoonotic pathogens. Species predicted to carry high-impact zoonotic pathogens can be prioritized for surveillance and prevention measures.
Zoonotic risk is often evaluated as a binary variable: whether a pathogen can or cannot infect humans. However, zoonotic pathogens differ in virulence (i.e., severity of disease) and transmissibility (i.e., capacity to spread in human populations after emergence). For example, within the genus Betacoronavirus, the case-fatality rates (CFRs) of viral species vary, with 5.9% in severe acute respiratory syndrome coronavirus (SARS-CoV) and 34.4% in Middle East respiratory syndrome coronavirus (MERS-CoV). Further, while the former efficiently spreads from person to person, the latter is much less efficient at human-to-human transmission. Virulence and transmissibility influence death burden (i.e., the total human disease-induced mortality), which collectively can describe a virus's epidemic potential. To reduce zoonotic emergence and optimize time and resource allocation for public health measures, zoonotic pathogens with the highest viral epidemic potential should be prioritized for surveillance and transmission prevention.
Recent work has demonstrated that bats (order: Chiroptera) harbor more viruses with high virulence in humans than other mammalian or avian orders. Bats are recognized as "special hosts" because they exhibit high viral diversity (hosting more viruses compared to most other taxa) and appear to tolerate many viruses without exhibiting clinical signs of infection. The high virulence of some of these viruses can also be at least partially explained by the evolutionary distance between bats and humans: humans are less likely to be pre-adapted to cope with these novel infections and show resistance, increasing morbidity and risk of mortality. However, as the distance between a vertebrate host and humans increases, a virus is also less likely to be pre-adapted to overcome human host defense mechanisms.
As the only flying mammals, bats evolved distinct immune adaptations to cope with the metabolic demands of flight, which likely has enabled them to tolerate otherwise virulent viruses. However, bat immune systems are diverse, and bats also do not uniformly interact with viruses. As the second most speciose order of mammals (1500 species), bats exhibit diverse ecological traits that may influence their viral communities, including their wide geographic distribution, use of torpor, heterogeneous diets, gregariousness, high mobility, long lifespan to body size ratio, and deep coevolutionary histories with viruses. Thus, the presumption that all bat species equally harbor a large number of virulent viruses may be inaccurate, and this has important implications for both public health and bat conservation. In particular, bats have received negative media attention given their identity as reservoir hosts of several high-profile viruses (e.g., SARS-CoV, Nipah virus), leading to retaliation against bats; this, in turn, can actually increase spillover risk. Clarifying the distribution of viral virulence across the order Chiroptera could provide the public with a more complete picture of viral epidemic potential, thereby improving public perceptions of bats and informing outbreak prevention efforts.
As other measures of viral epidemic potential, transmissibility in humans and high death burdens have had weaker connections to bats specifically. Recent work identified primates as the only order connected with unusually high onward transmission in humans, owing to their small phylogenetic distance from human hosts. Additionally, no taxonomic orders have been associated with high human death burdens. Corroborating which, if any, host groups are associated with high transmissibility and death burden in humans is further relevant to zoonotic risk assessment.
In this study, we test whether viruses with high epidemic potential in humans are uniformly distributed across mammals or if particular clades demonstrate high epidemic potential. We focus on bats to determine whether the entire order organically exhibits equal virulence or if only specific clades of bats exhibit high virulence. Given prior results, we also test whether mammalian host groups at finer taxonomic scales than order differ in transmissibility and death burden. We hypothesize that these measures of viral epidemic potential cluster within clades of bats, based on the unique coevolutionary histories between different bats and viruses. First, to assess the phylogenetic distribution of viral epidemic potential, we quantify phylogenetic signal of CFRs, onward transmission, and mean death burden across mammals and within bats specifically. Next, we use a flexible graph-partitioning algorithm to identify clades of mammals with unusually high or low viral epidemic potential without the need to a priori assume a specific phylogenetic scale. Finally, we mapped the geographic distributions of any identified bat clades with unusually high viral epidemic potential in tandem with anthropogenic footprint, visualizing hotspots of potential zoonotic spillover risk to guide outbreak surveillance and prevention efforts.