Niche dimensionality drives microbial community structure

Abstract

Niche dimensionality links environmental complexity to ecosystem structure. Although niche theory is often invoked in investigations of microbiomes, most models assume very high-dimensional coexistence, effectively sidestepping the role of dimensionality. However, direct estimates of niche dimensionality of microbiome have been lacking. Here we use joint species distribution modeling (JSDM) to infer niche dimensionality from relative abundance data alone. In paired 16S rRNA-metabolomics datasets, inferred dimensionality closely tracked the complexity of the metabolic environment, validating our abundance-only approach. Across nearly 200 human gut microbiome studies, lower dimensionality coincided with greater inter-species competition (metabolic niche overlap) and reduced biodiversity, consistent with the macroecological niche dimensionality hypothesis. Consumer/resource simulations reproduced these empirical relationships when both species-intrinsic metabolic tradeoffs and species-extrinsic environmental tradeoffs were imposed, with the latter dominating. Dimensionality was reduced in stressed or diet-simplified microbiomes and correlated broadly with ecological stress markers; negatively with prevalence of generalists and positively with microbial load. Together, these results establish niche dimensionality as a measurable and previously overlooked driver of microbial community structure.