Table 1.
Summary of 33 soil metagenomes used in this study.
Figure 1.
Taxonomic distribution of the soil microbial communities.
A) Distribution at the phylum level; B) Distribution at the family level. Labels show the taxonomic units with average relative abundance >2% in at least one of five soil habitats: desert, forest soil, grassland, Arctic soil, and mangrove sediment.
Figure 2.
A nonmetric multidimensional scaling (NMDS) plot showing diversity of soil ecosystems.
A Bray-Curtis distance similarity matrix was calculated based on the pairwise taxonomic profiles of 33 soil samples and used to generate NMDS coordinates of each sample. The distance linking two samples is shorter, indicating higher similarity between these samples. Samples from five soil sites were illustrated by different symbols and colors.
Table 2.
Community structure similarity of the soil metagenomes within a habitat or between habitat pair.
Figure 3.
Taxonomic composition of soil microbial community based on the metagenomes from five soil habitats.
A) Taxonomic cladogram showing all detected taxa (relative abundance ≥0.5%) in at least one sample. Taxonomic clades with more than five samples ≥0.5% abundance were used as inputs for LEfSe. Seven rings of the cladogram stand for domain (innermost), phylum, class, order, family, genus, and species (outermost), respectively. Enlarged circles in color are the differentially abundant taxa identified to be metagenomic biomarkers and the circle color is corresponding to the individual soil habitat in which the taxon is the most abundant among 5 soil ecosystems (Green for forest soil, red for grassland, purple for Arctic soil, blue for mangrove sediment, and orange for desert). B) The histograms of relative abundances of family-level biomarkers in each sample. Bacterial families significantly differential among all pairwise comparisons were illustrated. The average abundance of each family in the individual soil habitat was denoted by the horizontal line.
Figure 4.
Metagenome-level metabolic reconstruction of the soil microbial community.
KEGG BRITE hierarchical structures that are illustrated by the innermost four rings were used to cluster metabolic modules. The outermost ring composed of circles denotes KEGG functional modules detected in at least one of 30 metagenomes from three soil sites. Differentially abundant modules were inferred by LEfSe and illustrated by the enlarged circles in distinct colors: green stands for the modules most abundant in the forest soil, red for the grassland, and orange for the desert. The outermost rectangles denote core and differentially covered modules among three soil sites: ≥90% coverage stands for presence and ≤10% coverage for absence.
Table 3.
Core metabolic modules shared by grasslands, deserts, and forest soils.
Figure 5.
A global microbial interaction network of the soil microbial community.
The network captured all significant associations (multiple corrected p-value <0.05) among the abundances of phylotypes at or above the genus level in the soil microbial community within and across the three soil sites. Phylotypes were illustrated by nodes (light blue for grasslands, blue for forest soils, and yellow for deserts) and edges denote significant correlations between phylotypes: positive correlation colored in green means co-occurrence whereas negative correlation in red means mutual exclusion. The border of nodes was colored according to taxonomic affiliations at the class level.