Emergent function, not microbial conformity: functional redundancy and the limits of taxonomic inference in microbiome genomics.

Microbiome genomics has achieved remarkable resolution of community structure, yet composition alone remains an unstable basis for inferring host-relevant biology. That instability reflects a broader interpretive problem in which taxonomically distinct communities can converge on similar outputs, while superficially similar communities can diverge in behaviour because of strain variation, gene content, regulatory state, ecological context, spatial organization and host physiology. Functional redundancy is therefore better understood not as a reserve of interchangeable organisms but as a distributed functional architecture through which host-relevant outputs can persist across variation in membership. The central question for microbial genomics is not whether composition matters, but when community structure can be expected to predict function, host consequence or recovery. A more rigorous framework must distinguish membership from encoded capacity, realized activity, ecological interaction and host-relevant effect, while also recognizing that host physiology and spatial context shape which microbial functions become possible and which outputs are ultimately encountered. Progress will depend first on matching the evidentiary layer to the claim and then on selecting proportionate additions, from strain-resolved genomics and pathway-level interpretation to targeted metatranscriptomic, metaproteomic, metabolomic, spatial, perturbation-recovery or host-response measurements. In that framework, reproducibility may reside less in recurring taxa than in conserved biological outputs, and restoration less in compositional resemblance than in recovery of the functions and host-facing consequences that were actually disrupted.