Soil microbiomes are invaluable for ecosystem functions, sustainable agroecology and land-atmosphere climate feedbacks. Harnessing multi-omic data towards mechanistic and predictive understanding of the soil microbiome will aid in managing microbial phenotypes to benefit humanity. Yet bridging from microorganisms to biome-scale science to enable accurate modeling remains challenging due to the massive complexity and diversity of the soil microbiome. We hypothesized that, through targeted enrichments of a starting soil inoculum, we could obtain ‘functional module’ communities that are low-diversity, reproducible and predictable. These communities represent unique sets of microbial taxonomy and/or expressed functions and together encapsulate a significant extent of soil phylogenetic diversity while also enriching for underrepresented soil taxa.
To generate functional module communities, we cultured a soil inoculum in minimal medium with specific modifications designed to select for metabolic processes of interest. Emerging functional modules were then queried using both amplicon analysis (taxonomic composition) followed by metatranscriptomic analysis (expressed functions). This multi-omic data was then integrated to reveal how the soil microbiome function changes in response to variations in nutrients, growth conditions and stresses related to our field environment.
Results/Conclusions
We found that of the 241 unique taxa in the soil core microbiome, 90 were found in at least one functional module core, collectively encapsulating approximately 37.3% of soil phylogenetic diversity. We also found that several hundred OTUs were found in one or more of our functional modules but not in the soil core microbiome, showing that this approach can enrich for rare taxa that may be missed through an analysis of the native soil. Metatranscriptomic analysis revealed functional patterns that varied significantly among module treatments. Expression patterns were most uniform for simple substrates and increased in variability with substrate complexity. Bray-Curtis distances for transcriptomic and taxonomic profiles both showed consistent enrichment patterns by module. Further, a Mantel’s test revealed that functional dissimilarity was correlated with taxonomic dissimilarity. Comparing patterns of enriched transcripts showed that each of the modules had its own distinct biogeochemical reaction pattern on the collective microbial metabolic map, highlighting the potential for combining separate modules into a more complete biochemical repertoire, and ultimately reconstructing the full biochemical capacity of the soil microbiome. Our results provide a powerful glimpse into the functional expression of the soil microbiome and the potential to scale the microbiome into discrete functional units to generate a better predictive understanding of how the soil microbiome can influence ecosystem functions.