A global network of coexisting microbes from environmental and whole-genome sequence data

Microbes are the most abundant and diverse organisms on Earth. In contrast to macroscopic organisms, their environmental preferences and ecological interdependencies remain difficult to assess, requiring laborious molecular surveys at diverse sampling sites. Here, we present a global meta-analysis of previously sampled microbial lineages in the environment. We grouped publicly available 16S ribosomal RNA sequences into operational taxonomic units at various levels of resolution and systematically searched these for co-occurrence across environments. Naturally occurring microbes, indeed, exhibited numerous, significant interlineage associations. These ranged from relatively specific groupings encompassing only a few lineages, to larger assemblages of microbes with shared habitat preferences. Many of the coexisting lineages were phylogenetically closely related, but a significant number of distant associations were observed as well. The increased availability of completely sequenced genomes allowed us, for the first time, to search for genomic correlates of such ecological associations. Genomes from coexisting microbes tended to be more similar than expected by chance, both with respect to pathway content and genome size, and outliers from these trends are discussed. We hypothesize that groupings of lineages are often ancient, and that they may have significantly impacted on genome evolution.