Statistical analysis

The analysis of alpha diversity adopted four indices, including ACE, Chao1, Shannon, and Pielou evenness (Chao & Yang, 1993; Chao, 1984; Shannon, 1948). These indices were calculated using QIIME (Version 1.9.1) and visualized with R package ggplot2 (Version 2.15.3). One-way ANOVA was used to determine significant differences with a P-value of 0.05. If significant differences were detected, the outlier data were identified using Duncan’s post hoc test. The differences in species complexity among samples were derived from the beta diversity analysis. Moreover, the beta diversity on Bray-Curtis was obtained using QIIME software (version 1.9.1). The R language labdsv package was employed to calculate the indicators of each species in each treatment in the comparison group (Roberts & Roberts, 2016). Cross-validation was used for statistical testing to obtain the P value, which is displayed in a bubble chart.