Résumé
High-throughput sequencing technologies now allow for rapid and cost-effective surveys of multiple pathogens in rodents, but it is currently unclear if the host organ chosen for screening influences the number and identity of bacteria detected. We used 16s rRNA metabarcoding to identify bacterial pathogens in the heart, liver, lungs, kidneys and spleen of 13 water voles (Arvicola terrestris) collected from two populations in Franche-Comté, France to determine if bacterial assemblages within organs are similar, if all five organs are necessary to detect all of the bacteria present in an individual animal, and if differences between the two host population’s bacterial assemblages can be detected by each organ. We identified 25 bacteria representing 17 genera; average bacterial richness for each organ ranged from 1.5 ± 0.4 (mean ± standard error) to 2.5 ± 0.4 bacteria/organ and did not differ significantly between organs. The average bacterial richness when organ assemblages were pooled within animals was 5.4 ± 0.7 bacteria/animal, and rarefaction analysis indicates that all five organs are required to obtain this. Organ type does not, however, influence bacterial assemblage composition in a systematic or predictable way (PERMANOVA, 999 permutations, pseudo-F4,51=1.34, p=0.12). Ordination and PERMANOVA analysis indicates that differences in pooled, liver, and lung assemblages map to host populations, but heart, kidney and spleen assemblages do not. Our results demonstrate that the number of organs sampled influences the power to detect bacterial pathogens and host-population trends in bacterial assemblage composition. These results can inform sampling decisions in public health and wildlife ecology.