Phylogenetic Estimation of Ecologically Important Traits Illuminates Microbial Community Responses to Change in Natural and Agro-Ecosystems.

Many types of environmental change - from global-scale changes in climate to field-scale changes in management - alter the composition of microbial communities. However, determining whether similar "types" of microbes are selected by similar changes, or by the same change across locations, remains difficult, limiting synthesis across studies. By allowing communities to be described in terms of the ecological types of taxa they contain, estimating trait values using 16S rRNA sequence-based phylogenetic placements could provide a way forward. We developed a reference tree of 1475 high-quality microbial genomes and used it to estimate values of three ecologically important traits: rRNA gene copy number (indicator of maximum growth rate), genome size (indicator of metabolic versatility), and anaerobic tolerance. Traits were estimated for soil microbial communities from two experiments: (1) a grassland experiment in which supplemental late-season precipitation and NPK fertilizer were added to adjacent high- and low-productivity soils; and (2) a long-term agricultural experiment comparing three management systems: organic (C+N inputs from cover crops and composted manure), conventional (N inputs from mineral fertilizer) and no input. Weighted average trait values for each community were calculated using OTU relative abundances. In the grassland, addition of nutrients resulted in higher copy numbers, larger genomes, and higher anaerobic tolerance, potentially selecting for faster-growing microbes with higher metabolic versatility. Supplemental precipitation produced the opposite set of responses. In the agricultural experiment, the organic system had higher copy numbers, smaller genomes, and higher anaerobic tolerance than the no input system, with the conventional system falling in the middle. Notably, although inputs to both the grassland and agricultural system selected for high copy numbers (potentially faster growth), the grassland nutrient addition increased average genome size (potentially higher metabolic versatility) whereas the agricultural carbon + nutrient inputs decreased it (potentially lower metabolic versatility). This difference may result from reduced soil habitat complexity, lower carbon substrate diversity, or other differences in the agricultural system compared to the grassland. In sum, community-wide estimation of microbial traits holds promise for ecologically meaningful prediction of environmental change responses, facilitating cross-system insight by providing a common currency for comparison.