3.1 Diversity and community compositions
After quality filtering and reads control, a total of 1,522,719 quality sequences were obtained from 81 samples and clustered into 9797 OTUs at the cutoff of 97% sequence identity. Among the 9,797 OTUs obtained, 132 (1.3%) abundant taxa with 555,761 (36.5%) sequences and 8,182 (83.5%) rare taxa with 398,356 (26.2%) sequences were identified (Table S2). Both abundant (R2 = 0.097, P < 0.001) and rare (R2 = 0.617, P < 0.001) taxa showed significantly positive abundance-occupancy relationships (Fig. S1a). The abundant taxa were more widespread geographically than rare taxa, with 99% of the abundant OTUs being detected in more than 50% of sites. In contrast, only 4% of rare taxa existed in more than 50% of sites. Taxonomic analyses revealed that Proteobacteria (39%), Bacteroidetes (19%), Firmicutes (17%), and Actinobacteria (14%) were the most abundant bacterial phyla, together accounting for 89% of the total sequences of abundant taxa (Fig. S1b). The rare taxa were dominated by Proteobacteria (40%), followed by Actinobacteria (13%) and Bacteroidetes (12%). Some phyla, such as Acidobacteria, Chloroflexi, and Planctomycetes were present at high abundances in rare taxa.
Alpha diversity metrics, including the Shannon, Pielou’s evenness, and Faith’s PD, of rare taxa were significantly higher than those of abundant taxa (Wilcoxon rank sum test, P < 0.001, Fig. S2a-c). However, SES.MNTD was significantly higher for abundant taxa than for rare ones (Wilcoxon rank-sum test, P < 0.001, Fig. S2d), indicating that rare taxa were more closely clustered phylogenetically than abundant taxa.