Figures and legends
Fig 1 Explanatory factors associated with features of the fungal-plant network in individual (a& b ) and aggregate (c ) (Only the top three factors are shown to be formally significant). (a ) Random forest analysis showing influences on number of fungal OTUs linked to each plant species, number of edges and modularity of plant-fungi network. (b ) The heatmap of spearman correlation coefficients between the network features and environmental factors. (c ) The explained variation values were estimated with the multiple regression on distance matrices analysis. The results showed that plant species richness is the most important predictor for plant-fungi network topological features. MAP, mean annual precipitation; MAT, mean annual temperature; CNratio, carbon: nitrogen; SM, soil moisture; STP, soil total phosphorus; SOC, soil organic carbon. *P < 0.05, **P < 0.01, ***P < 0.001.
Fig 2 Relationship between plant species richness per site and plant-fungi network features (a, c& d ) or the ratio of fungal richness to plant richness (b ) in each site (vertical axis). (b )All OTUs detected in soil are included, not only those linked to plant species. (P=< 0.001)
Fig 3 Relationship between plant species richness per site and plant-bacteria network features (a, c& d ) or the ratio of bacterial richness to plant richness(b ) in each site (vertical axis). (a ) there is no significant trend between plant richness and number of bacterial OTUs linked to each plant species(P>0.05). (b ) All OTUs detected in soil are included, not only those linked to plant species. (P=< 0.001)
Fig 4 Observed plant-microbiota network structure for each main vegetation type sampled in this study. In each network of (a ) desert steppe (DS) (8 plants and 25 fungi), (b ) alpine meadow (AM) (21 plants and 118 fungi), and (c ) alpine steppe (AS) (28 plants and 138 fungi), (d ) desert steppe (DS) (13 plants and 139 bacteria), (e ) alpine meadow (AM) (21 plants and 84 bacteria), and (f ) alpine steppe (AS) (6 plants and 15 bacteria), fungi (yellow) and bacteria (purple) are linked with their host plants (green). The size of circles represents the degree of microbiota or plants in each network. A blue edge indicates a negative correlation, and a red edge indicates a positive correlation.
Fig 5 Random Forest analysis and linear regression showing influences on the network-level topological features of fungi-only network(left) and bacterial-only network(right). Random forest shows that STP is the most important environmental factors for the fungi-only network (a ), while SM is the most important for the bacteria-only network (b ). MAP, mean annual precipitation; MAT, mean annual temperature; CNratio, carbon: nitrogen; SM, soil moisture; STP, soil total phosphorus; SOC, soil organic carbon. *P < 0.05, **P < 0.01 and ***P < 0.001.