Data Analyses

All statistical analyses were performed using MicrobiomeAnalyst (https://www.microbiomeanalyst.ca/) [33] and R software version 4.1.2. For taxonomic analysis, the alpha-diversity was evaluated using the Chao1 index with the “vegan” package (version 2.5-6, https://CRAN.R-project.org/package=vegan) [34]. Phylogenetic tree analysis of soil was conducted using cluster_agg from Usearch10. The alpha-diversity indices were compared by one-way analysis of variance (ANOVA) and Tukey's multiple comparisons test.

The overall soil treatment effect on bacterial communities was examined using principal coordinates analysis (PCoA) combined with multivariate PERMANOVA of Bray-Curtis distances based on ASVs level and computed with the “vegan” R package. Further, to explore the structure of bacterial community assembly processes by deterministic or stochastic processes, the β-nearest taxon index (βNTI) [35] was calculated with the “picante” package (version 1.8.2). The bacterial community assembly processes can be indicated by the fraction of communities with |βNTI|>2 and |βNTI|<2, respectively. The linear discriminant analysis effect size (LEfSe) method was used [36] to compare microbial compositions of the soil amendments. The correlation between the phylum and environmental variables was assessed by redundancy analysis (RDA) using the “vegan” R package. To evaluate their associations with the community structures, environmental factors were transformed and normalized for use in the Mantel test [37]. For soil chemical property values, group comparisons were performed using the Wilcoxon/Kruskal-Wallis test.