Abstract
Data on the gut microbiota (GM) of wild animals are key to studies on evolutionary biology (host-GM interactions under natural selection), ecology and conservation biology (GM as a fitness component closely connected to the environment). Wildlife GM sampling often requires non-invasive techniques or sampling from dead animals. In a controlled experiment profiling microbial 16S rRNA in 52 house mice (Mus musculus) from eight families and four genetic backgrounds, we studied the effects of live- and snap-trapping on small mammal GM and evaluated the suitability of microbiota from non-fresh faeces as a proxy for caecal GM. We compared CM from individuals sampled 16-18 h after death with those in live traps and caged controls, and caecal and faecal GM collected from mice in live-traps. Sampling delay did not affect GM composition, validating data from fresh cadavers or snap-trapped animals. Animals trapped overnight displayed a slight but significant difference in GM composition to the caged controls, though the change only had negligible effect on GM diversity, composition and inter-individual divergence. Hence, the trapping process appears not to bias GM profiling. Despite their significant difference, caecal and faecal microbiota were correlated in composition and, to a lesser extent, diversity. Both showed congruent patterns of inter-individual divergence following the natural structure of the dataset. Thus, the faecal microbiome represents a good non-invasive proxy of the caecal microbiome, making it suitable for detecting biologically relevant patterns. However, care should be taken when analysing mixed datasets containing both faecal and caecal samples.