Extended Data Fig. 1: Precipitation level and mineral nutrient content correlate with soil microbiota composition.

a. Soil alpha diversity (estimated using the Shannon diversity index) did not differ among sampling locations. Locations are ordered by increasing precipitation (Fig. 1a) (N = 6 independent soil samples per location). Significance was determined using analysis of variance (ANOVA, F_5,28 = 2.29, p = 0.073); letters correspond to a two-tailed Tukey post-hoc test (α = 0.05) of pairwise contrasts. Points represent the estimated marginal means with 95% confidence intervals. b. Phylogram showing the relative abundance profiles of the main bacterial phyla across soils exposed to different precipitation levels. c. Principal component analysis showing the ionomic profiles of the six focal soils (N = 6 per soil). The bar on the left denotes the percentage of the variance explained by the predictor variables. d. Heatmap showing the standardized concentration (z-score) of each mineral nutrient (rows) in the collection of soils exposed to different precipitation levels. The values were clustered according to the ion concentrations and the regions were ordered from low to high precipitation. e. Bar graph showing the Pearson correlation coefficient between each mineral nutrient abundance and the level of precipitation across the collection of soils used. Coloured bars indicate statistically significant correlations (q < 0.05). f. Pairwise correlation analysis between soil mineral nutrient dissimilarities and soil microbiome composition dissimilarities in soils representing a gradient of precipitation (N = 6 per soil). Each point represents one pair of soil samples. Panel shows the one-tailed Mantel test r statistic and its p-value (p = 1e-4).