Cross-biome metagenomic analyses of soil microbial communities and their functional attributes

Abstract

For centuries ecologists have studied how the diversity and functional traits of plant and animal communities vary across biomes. In contrast, we have only just begun exploring similar questions for soil microbial communities despite soil microbes being the dominant engines of biogeochemical cycles and a major pool of living biomass in terrestrial ecosystems. We used metagenomic sequencing to compare the composition and functional attributes of 16 soil microbial communities collected from cold deserts, hot deserts, forests, grasslands, and tundra. Those communities found in plant-free cold desert soils typically had the lowest levels of functional diversity (diversity of protein-coding gene categories) and the lowest levels of phylogenetic and taxonomic diversity. Across all soils, functional beta diversity was strongly correlated with taxonomic and phylogenetic beta diversity; the desert microbial communities were clearly distinct from the nondesert communities regardless of the metric used. The desert communities had higher relative abundances of genes associated with osmoregulation and dormancy, but lower relative abundances of genes associated with nutrient cycling and the catabolism of plant-derived organic compounds. Antibiotic resistance genes were consistently threefold less abundant in the desert soils than in the nondesert soils, suggesting that abiotic conditions, not competitive interactions, are more important in shaping the desert microbial communities. As the most comprehensive survey of soil taxonomic, phylogenetic, and functional diversity to date, this study demonstrates that metagenomic approaches can be used to build a predictive understanding of how microbial diversity and function vary across terrestrial biomes.

Fig. 1.

Differences in alpha diversity levels across 16 soils. X axis shows taxonomic richness of the bacterial communities (number of phylotypes out of 118,000 amplicon reads per sample). Y axis shows the functional gene richness (number of functional gene categories identified from 688,000 annotated shotgun metagenomic reads per sample). See Table S2 for additional information on diversity levels across the 16 soils.

Fig. 2.

Ordination plots derived from principal coordinates analyses of Bray–Curtis distances between bacterial community composition (Upper, based on amplicon 16S rRNA gene data) and metagenome composition (Lower, based on annotated shotgun metagenomic data).

Fig. 3.

Relative abundances of major categories of functional genes in the shotgun metagenomes obtained from the desert soils (both cold and hot deserts) versus the other, nondesert, biomes. Asterisks and bold type indicate those categories with significantly different relative abundances in desert and nondesert soils (Bonferroni corrected P values <0.05, uncorrected P values <0.002).