Coronavirus sampling and surveillance in bats from 1996-2019: a systematic review and meta-analysis.

The emergence of SARS-CoV-2 highlights a need for evidence-based strategies to monitor bat viruses. We performed a systematic review of coronavirus sampling (testing for RNA positivity) in bats globally. We identified 110 studies published between 2005 and 2020 that collectively reported positivity from 89,752 bat samples. We compiled 2,274 records of infection prevalence at the finest methodological, spatiotemporal and phylogenetic level of detail possible from public records into an open, static database named datacov, together with metadata on sampling and diagnostic methods. We found substantial heterogeneity in viral prevalence across studies, reflecting spatiotemporal variation in viral dynamics and methodological differences. Meta-analysis identified sample type and sampling design as the best predictors of prevalence, with virus detection maximized in rectal and faecal samples and by repeat sampling of the same site. Fewer than one in five studies collected and reported longitudinal data, and euthanasia did not improve virus detection. We show that bat sampling before the SARS-CoV-2 pandemic was concentrated in China, with research gaps in South Asia, the Americas and sub-Saharan Africa, and in subfamilies of phyllostomid bats. We propose that surveillance strategies should address these gaps to improve global health security and enable the origins of zoonotic coronaviruses to be identified.

Exploring the Mosquito-Arbovirus Network: A Survey of Vector Competence Experiments.

Arboviruses receive heightened research attention during major outbreaks or when they cause unusual or severe clinical disease, but they are otherwise undercharacterized. Global change is also accelerating the emergence and spread of arboviral diseases, leading to time-sensitive questions about potential interactions between viruses and novel vectors. Vector competence experiments help determine the susceptibility of certain arthropods to a given arbovirus, but these experiments are often conducted in real time during outbreaks, rather than with preparedness in mind. We conducted a systematic review of reported mosquito-arbovirus competence experiments, screening 570 abstracts to arrive at 265 studies testing in vivo arboviral competence. We found that more than 90% of potential mosquito-virus combinations are untested in experimental settings and that entire regions and their corresponding vectors and viruses are undersampled. These knowledge gaps stymie outbreak response and limit attempts to both build and validate predictive models of the vector-virus network.

A minimum data standard for vector competence experiments.

The growing threat of vector-borne diseases, highlighted by recent epidemics, has prompted increased focus on the fundamental biology of vector-virus interactions. To this end, experiments are often the most reliable way to measure vector competence (the potential for arthropod vectors to transmit certain pathogens). Data from these experiments are critical to understand outbreak risk, but - despite having been collected and reported for a large range of vector-pathogen combinations - terminology is inconsistent, records are scattered across studies, and the accompanying publications often share data with insufficient detail for reuse or synthesis. Here, we present a minimum data and metadata standard for reporting the results of vector competence experiments. Our reporting checklist strikes a balance between completeness and labor-intensiveness, with the goal of making these important experimental data easier to find and reuse in the future, without much added effort for the scientists generating the data. To illustrate the standard, we provide an example that reproduces results from a study of Aedes aegypti vector competence for Zika virus.

Optimising predictive models to prioritise viral discovery in zoonotic reservoirs.

Despite the global investment in One Health disease surveillance, it remains difficult and costly to identify and monitor the wildlife reservoirs of novel zoonotic viruses. Statistical models can guide sampling target prioritisation, but the predictions from any given model might be highly uncertain; moreover, systematic model validation is rare, and the drivers of model performance are consequently under-documented. Here, we use the bat hosts of betacoronaviruses as a case study for the data-driven process of comparing and validating predictive models of probable reservoir hosts. In early 2020, we generated an ensemble of eight statistical models that predicted host-virus associations and developed priority sampling recommendations for potential bat reservoirs of betacoronaviruses and bridge hosts for SARS-CoV-2. During a time frame of more than a year, we tracked the discovery of 47 new bat hosts of betacoronaviruses, validated the initial predictions, and dynamically updated our analytical pipeline. We found that ecological trait-based models performed well at predicting these novel hosts, whereas network methods consistently performed approximately as well or worse than expected at random. These findings illustrate the importance of ensemble modelling as a buffer against mixed-model quality and highlight the value of including host ecology in predictive models. Our revised models showed an improved performance compared with the initial ensemble, and predicted more than 400 bat species globally that could be undetected betacoronavirus hosts. We show, through systematic validation, that machine learning models can help to optimise wildlife sampling for undiscovered viruses and illustrates how such approaches are best implemented through a dynamic process of prediction, data collection, validation, and updating.