Fig. 3: Relationships between resource utilization width and metabolic interaction potential or metabolic resource overlap in plant-associated bacteria.
a Distribution of experimental and genome-predicted resource utilization widths for 224 phyllosphere strains, and predicted widths for 3001 rhizosphere bacterial genomes. Gaussian nonlinear regression curve showing best-fit line. b Correlation between experimental and predicted resource utilization widths in phyllosphere strains. c Pairwise interaction simulations using narrow-spectrum and broad-spectrum resource-utilizing strains (with resource utilization widths ≤7 and ≥29) against all 224 phyllosphere strains, showing correlations between community metabolic interaction potential (MIP) or metabolic resource overlap (MRO) and average experimental or predicted resource utilization width. The line represents a linear regression fit. d Effect of phylogenetic distance on the strength of the correlation (R2) between community MIP/MRO and average experimental resource utilization width. The 224 phyllosphere bacterial models and the 3001 rhizosphere bacterial genomes were derived from the published data of Schäfer et al. and Dai et al., respectively40,41. Source data are provided as a Source Data file.
