Statistics
As the sequenced library size was variable among samples, ASV count was normalised by the Trimmed Mean of M values method in the edgeR package in R (Robinson et al., 2010). Bray-Curtis distance was calculated between samples to indicate the differentiation of microbiome composition. Based on Bray-Curtis distance, non-metric multidimensional scaling (NMDS) plot was generated to visualize the microbial community from the Vegan Package (Oksanen et al., 2007). Permutational multivariate analyses of variance (PERMANOVA) with maximum 999 permutations (adonis function from Vegan package) was conducted to test the effects of soil type, chickpea genotype, rhizocompartment, root segment, and their interactions on microbiome composition, and root segment and rhizocompartment were nested in soil and genotype. Shannon index of samples was estimated using iNEXT in R package (Hsieh et al., 2016). PERMANOVA and unconstrained principal coordinate analyses (PCoA) using Bray-Curtis distance were also conducted to analyse the composition of root exudate compounds. The Statistical Analysis of Metagenomic Profiles (STAMP) package was used to determine statistical differences of taxonomic groups between experimental treatments with Benjamini-Hochberg FDR method to correct for the P value (Parks et al., 2014). For the variables of Bray-Curtis distance, Shannon index, CSR, bacterial number, and root exudates, General Linear Model was applied to test the effects of experimental factors and interactions using Minitab (Minitab Inc., State College, PA, USA). Normal distribution of each of the variables was checked and confirmed. Tukey HSD post-hoc pairwise comparison test was performed to compare the treatments.