Drivers of host-infectious agent community associations in seabirds from sub-Antarctic oceanic islands

Abstract

Understanding the patterns and drivers of infectious disease dynamics at different levels of life organization is a fundamental disease ecology challenge. In this study, we performed a multi-infectious agents (IA) screening of 1,983 individuals belonging to 18 seabird species, sampled between 2017 and 2022 across five sub-Antarctic islands in the Southern Indian and Atlantic Oceans. Our aim was to identify the drivers of IA community in these seabirds by considering intrinsic and extrinsic factors on three dimensions: distance between seabird metapopulations and within-island sites, macro-community (host functional traits based on feeding and breeding features) and intra-host IA community composition. Samples were screened for 24 DNA-based IAs using a high-throughput real-time PCR method, followed by subsequent hierarchical modelling of species communities. Campylobacter lari, Escherichia coli, Pasteurella multocida, Chlamydiaceae and Mycobacterium spp. were found to be present on all islands in some, but not all, species. However, only E. coli was detected in all species. According to our modelling results, the influence of year and within-island site had limited influence on IA presence/absence, accounting respectively for an average of 1.0% and 1.8% of the variation in IA community. Including host species as a response variable produced a better model fit than including host species functional traits, highlighting the difficulty in simplifying systems by focusing on high-level functional categories. Model outputs show how each species influences IA distribution depending on IA traits, but no clear trend has emerged. Interestingly, burrowing species were less frequently infected with directly transmitted IAs, probably due to the low frequency of transmission events associated with their breeding features and limited social contacts. Our findings pave the way for further multi-dimension studies that would combine complementary approaches to disentangle the complex processes at play in the dynamics of hosts and IAs interactions.