Fig. 6: Interactions between wheat root metabolites and rhizosphere microbiome and N cycling functional genes.

a, b Bipartite networks visualizing the interactions between individual molecules of root metabolite and rhizosphere microorganisms (genus level) in each management system. The orange nodes represent bacterial genera, the green nodes represent fungal genera; and nodes of other colors depict different categories of root metabolites. The edges represent significant Spearman correlations (|ρ| > 0.6, p < 0.05, two-sided test). Subnetworks extracted from the main network illustrate the interactions between different categories of root metabolites and bacterial (c, e) and fungal (d, f) biomarker families. Each category of root metabolites is depicted with unique colors. The size of each node corresponds to the number of connections between each microorganism or root metabolite. The thickness of connecting edges between root metabolites and microorganisms indicates the number of observed connections, with thicker lines representing a higher number of significant connections. Red edges indicate predominantly positive correlations (where positive correlations account for over 50% of the total), while blue edges denote mainly negative correlations (where positive correlations account for less than 50% of the total). g, h Effects of root metabolism quantity and quality on N cycling related genes using Redundancy analysis (RDA) analysis. PCoA1 and PCoA2 represent root metabolism compositional changes; the Shannon index indicates root metabolism chemodiversity; specific root exudation rates (TOC, mg C g⁻¹ root biomass h⁻¹), and different categories of root exudates. **p  < 0.001, ** p  < 0.01, * p  <  0.05 (two-sided, unrestricted permutation test, n = 999). i Contribution of root metabolism quantity and quality to six classes of N cycling genes in assimilatory nitrate reduction to ammonium (ANRA), dissimilatory nitrate reduction to ammonia (DNRA), denitrification, nitrification, N mineralization, and assimilation using Random Forest analysis. PCoA1 and PCoA2 represent root metabolism compositional changes; the Shannon index indicates root metabolism chemodiversity; specific root exudation rates (TOC, mg C g⁻¹ root biomass h⁻¹). Statistical significance of predictor importance was assessed using Random Forest with permutation tests (two-sided, 1000 permutations, ntree = 2000). Reported p values are exact values without adjustment for multiple comparisons. Yellow bars represent factors significant at p < 0.05, and the dark gray column indicates the factor that has no significant effect. ***p  <  0.001, ** p  <  0.01, * p  <  0.05. Source data are provided as a Source Data file.