CSH Talk by Sirus Rasti (Visiting Student)

Zoonotic disease surveillance is inherently opportunistic: Documented host-pathogen associations largely reflect research and surveillance biases (favoring certain species, regions, and pathogens). This creates blind spots in our understanding of disease ecology, particularly for understudied wildlife host taxa and emerging pathogens. Building on a network derived from a systematic literature search of Austrian zoonotic disease research, we address this gap by applying three predictive modeling approaches to estimate the probability of undocumented links between hosts and zoonotic agents. We show that pathogen and host taxonomy are the strongest predictors of association, while ecological traits provide an additional but secondary signal. We predict previously undocumented associations in Austria, including novel links (i.e., not reported in any database), revealing hidden connectivity in the Austrian zoonotic Web. These findings demonstrate how predictive modeling can prioritize surveillance effort toward the most plausible yet unsampled host-pathogen combinations.

Please note that this is an internal talk. If you want to participate, please write an email to .