Microbiome Composition in Both Wild-Type and Disease Model Mice Is Heavily Influenced by Mouse Facility

Murine models have become essential tools for understanding the complex interactions between gut microbes, their hosts, and disease. While many intra-facility factors are known to influence the structure of mouse microbiomes, the contribution of inter-facility variation to mouse microbiome composition, especially in the context of disease, remains under-investigated. We replicated microbiome experiments using identical mouse lines housed in two separate animal facilities and report drastic differences in composition of microbiomes based upon animal facility of origin. We observed facility-specific microbiome signatures in the context of a disease model [the Ednrb (endothelin receptor type B) Hirschsprung disease mouse] and in normal C57BL/6J mice. Importantly, these facility differences were independent of cage, sex, or sequencing-related influence. In addition, we investigated the reproducibility of microbiome dysbiosis previously associated with Ednrb^-/- (knock-out; KO) mice. While we observed genotype-based differences in composition between wild-type (WT) and KO mice, these differences were inconsistent with the previously reported conclusions. Furthermore, the genotype-based differences were not identical across animal facilities. Despite this, through differential abundance testing, we identified several conserved candidate taxa and candidate operational taxonomic units that may play a role in disease promotion or protection. Overall, our findings raise the possibility that previously reported microbiome-disease associations from murine studies conducted in a single facility may be heavily influenced by facility-specific effects. More generally, these results provide a strong rationale for replication of mouse microbiome studies at multiple facilities, and for the meticulous collection of metadata that will allow the confounding effects of facility to be more specifically identified.