Fig. 3

The bacterial communities and their predicted functional potentials differed between the samples collected until the 35th day post planting (early samples) and those collected later (late ones). The graphs in panels A, B, C, D, E, and H show (top to bottom) leaves, roots and soils. The black outlines indicate early samples, while the red outlines indicate late samples. The significance of differences in means (Wilcoxon’s test) in panels A, D and F is shown: ***p < 0.001, **p < 0.01, *p < 0.05, ns – not significant. (A) The alpha diversity, evenness, and species richness of ASV sets. (B) dbRDA analysis of d05 generalized UniFrac distance matrices (results of a permutational test (ANOVA) are shown on graphs). (C) dbRDA analysis of Morisita-Horn distance matrices calculated for PICRUSt2-predicted functions sets (results of a permutational test are shown on graphs). (D) Alpha diversity, evenness, and richness of predicted functions sets. (E) Ecological processes governing maintenance of soil, root, and leaf bacterial communities. DR—drift, HD—homogenizing dispersal, DL—dispersal limitation, HoS—homogenous selection, HeS—heterogenous selection. (F) Nestedness. (G) Ecological processes governing transfer of organisms from soils to roots and from roots to leaves. Processes are marked as in panel E. (H) Taxonomic structure at the genus level. Differentially represented (DR) taxa were identified with DESeq2 and ALDEx2 (taxa identified by both methodologies were considered DR), and the upper indices next to the taxon names indicate the kinds of samples in which there was a difference: s—soils, r—roots, and l—leaves.