Soil bacterial communities exhibit strong biogeographic patterns at fine taxonomic resolution

Bacteria have been inferred to exhibit relatively weak biogeographic patterns. To what extent such findings reflect true biological phenomena or methodological artefacts remains unclear. Here, we addressed this question by analysing the turnover of soil bacterial communities from three datasets. We applied three methodological innovations: (i) design of a hierarchical sampling scheme to disentangle environmental from spatial factors driving turnover; (ii) resolution of 16S rDNA amplicon sequence variants to enable higher resolution community profiling; and (iii) application of the new metric zeta diversity to analyze multisite turnover and drivers. At fine taxonomic resolution, rapid compositional turnover was observed across multiple spatial scales. Turnover was overwhelmingly driven by deterministic processes and influenced by the rare biosphere. The communities also exhibited strong distance decay patterns and taxon-area relationships, with z values within the interquartile range reported for macroorganisms. These biogeographical patterns were weakened upon applying two standard approaches to process community sequencing data: clustering sequences at 97% identity threshold and/or filtering the rare biosphere (sequences lower than 0.05% relative abundance). Comparable findings were made across local, regional, and global datasets and when using shotgun metagenomic markers. Altogether, these findings suggest that bacteria exhibit strong biogeographic patterns, but these signals can be obscured by methodological limitations. We advocate various innovations, including using zeta diversity, to advance the study of microbial biogeography.