Résumé
1. Within complex ecological systems, identifying animal species likely to play a key role in the emergence of infectious zoonotic diseases remains a major challenge. One approach consists of using information on current ecological and parasitological similarities among host species in order to predict the most likely pathways for future pathogen spillover.
2. Using field data acquired from 15 sympatric rodent species in various habitats in Thailand, Cambodia and Laos, we built networks based on shared parasites (17 helminth and 15 microparasite species) and shared habitats among rodent species and humans. We investigated the architectures of bipartite and unipartite networks using modularity, subgroups partitioning or node centrality, to assess the relative epidemiological importance of particular rodent species.
3. Our results showed that Rattus tanezumi, Bandicota savilei and R. exulans were consistently found to be members of subgroups that included humans in unipartite and bipartite networks on zoonotic agents and shared habitats. High values of centrality in shared zoonotic agents were found for the same three rodent species, whereas high values of shared habitats were observed for two of them. Although phylogenetically related rodent species likely shared both habitats and parasites, a lack of habitat specialization was associated with increased zoonotic parasite sharing.
4. Our results emphasize the disproportionate importance of these three rodent species, through their high degree of connectivity with humans, which may represent a high risk for direct zoonotic spillover. Moreover, due to its high centrality in habitats, R. tanezumi may also play a key role as a bridge host.
5. The recent discovery of new arenaviruses in rodents in South-East Asia, with associated disease in humans in Cambodia, provides an opportunity to test this empirically. The three rodent species identified using our network approach are some of the potential maintenance hosts for these new emerging arenaviruses.
6. Synthesis and applications. Our results on rodents and their pathogens in South-East Asia show that network analysis has a high potential to improve the surveillance of emerging zoonotic pathogens by targeting key host species and potential 'emerging' pathogen-rodent interactions in complex and heterogeneous landscapes.