Extended Data Fig. 4: Precipitation legacy effects in the soil microbiota are resilient to short-term water- and host-related perturbations.

a. Schematic representation of the experimental design used to evaluate the resilience of precipitation legacy effects to perturbations and their functional importance to plant drought response. Six soils spanning the Kansas precipitation gradient (“legacy phase”), were either left unplanted or planted with seedlings of the native grass species Tripsacum dactyloides (eastern gamagrass) and subjected to either drought conditions or regular watering in a factorial design (“conditioning phase”). To evaluate how soil precipitation legacy affects plants, and to disentangle the role of the microbiota from possible effects of co-varying abiotic soil properties, we then used the experimentally-conditioned microbial communities to inoculate a new generation of T. dactyloides and Z. mays plants. These “test phase” plants were divided between water-limited conditions and well-watered control conditions. b. Alpha diversity of bacterial communities was not affected by the different conditioning treatments (drought or well-watered, with or without host). For each treatment, at least 12 biological replicates (individual soil pots) were analysed, resulting in a total of 156 bacterial community profiles generated. Alpha diversity was assessed using the Shannon Diversity Index. In the boxplots, the horizontal line represents the median; box edges indicate the interquartile range (25th-75th percentiles); and whiskers extend to the smallest and largest values within 1.5× the interquartile range. Individual data points, including outliers, are overlaid as dots. Differences among groups were tested with one-way analysis of variance (ANOVA; p = 0.16), which revealed no statistically significant variation. Tukey’s test further confirmed the absence of significant pairwise differences, as indicated by the shared letters. c. Phylograms show that the different conditioning treatments (drought or well-watered, with or without host) did not impact the relative abundance profiles of main bacterial phyla. d. Constrained ordination of metagenome taxonomic composition in response to conditioning phase treatments. Statistics are from permutational MANOVA (PERMANOVA) with 9999 permutations. PERMANOVA R² and p-values for the effects of drought/water and host conditioning are shown. The bar on the left indicates the percentage of variance explained by the experimental variables, as estimated by PERMANOVA.