The macroecology and evolution of avian competence forBorrelia burgdorferi

Abstract

12Most efforts to predict novel reservoirs of emerging infectious diseases rely upon information 13 about the infection status of host species rather than their competence, the ability to transmit 14 pathogens to new susceptible hosts or vectors. Although competence can be difficult to quantify, 15 tick-borne pathogens can provide a useful model system, as larval ticks can become infected 16 only when feeding on a competent host during their first bloodmeal. For tick-borne diseases, 17 competence has been best studied for Borrelia burgdorferi sensu lato (Bbsl), the pathogen that 18 causes Lyme borreliosis, an increasing public health priority across the northern hemisphere. 19Major reservoirs include several small mammal species, but birds may play an underrecognized 20 role in human risk of acquiring this tick-borne pathogen, given their ability to disperse infected 21 ticks across large spatial scales. Here, we provide a global synthesis of the ecological and 22 evolutionary factors that determine the ability of bird species to infect larval ticks with Bbsl. We 23 find strong phylogenetic patterns in competence, which clusters in thrushes and has evolved 24 multiple times across birds. With boosted regression trees, we distinguish competent birds with 25 76% accuracy and show competent species are more likely to fledge early, have low baseline 26 corticosterone, breed at high latitudes, and have more diverse migratory movements into their 27 breeding range. We use these trait profiles to predict over 100 likely but unsampled Bbsl-28 competent birds, including novel concentrations of avian reservoirs in the Neotropics and pelagic 29 zones. Beyond expanding our understanding of avian reservoirs for tick-borne pathogens and 30 generating new hypotheses, our findings also emphasize the need to consider competence in 31 efforts to predict novel hosts of emerging diseases and in the study of multi-host pathogens. 32

Biodiversity loss underlies the dilution effect of biodiversity

“…While many studies focus on measuring the diversity of host species in the context of disease, the structure of host communities can also be measured in the context of disease using characteristics of host species or host functional traits (Johnson et al ., 2013; Halliday et al ., 2019; Kirk et al ., 2019), resulting in trait‐based measures of host community competence (Stewart Merrill and Johnson, 2020). This approach, which has rapidly gained traction in disease ecology, suggests that host species that are the best able to spread diseases (i.e., the most competent hosts), often share particular suites of physiological traits (Huang et al ., 2013; Martin et al ., 2019; Becker and Han, 2020). Thus, host community competence can be linked to distributions of important host traits across host communities (Johnson et al ., 2015b; Liu et al ., 2017).…”

Optimizing predictive models to prioritize viral discovery in zoonotic reservoirs

“…Previous work has been highly successful in predicting zoonotic reservoirs using a combination of taxonomic, ecological, and geographic traits as predictors. This approach has been previously used to identify wildlife hosts of filoviruses 27,100 , flaviviruses 28,101 , henipaviruses 23 , Borrelia burgdorferi 26 , to predict mosquito vectors of flaviviruses 102 , and to predict rodent reservoirs and tick vectors of zoonotic viruses 37,103 . These approaches treat the presence of a specific virus (or genus of viruses) or a zoonotic pathogen as an outcome variable, with negative values given for species not known to be hosts (pseudoabsences), and use machine learning to identify the characteristics that predispose animals to hosting pathogens of concern.…”

“…Given such restrictions, modeling efforts can play a critical role in helping to prioritize pathogen surveillance by narrowing the set of plausible sampling targets 25 . For example, machine learning approaches have generated candidate lists of likely, but unsampled, primate reservoirs for Zika virus, bat reservoirs for filoviruses, and avian reservoirs for Borrelia burgdorferi 26–28 . In some contexts, models may be more useful for identifying which host or pathogen groups are unlikely to have zoonotic potential 29 .…”

Beyond Infection: Integrating Competence into Reservoir Host Prediction

“…As competence encompasses several stepwise processes, this trait is continuous [ 8 ]. For some purposes, it may be approximated as binary [ 9 ].…”