3.3.1 Topological features
MENs topological features (i.e. number of nodes, number of links, avgK, avgCC, geodesic distance and modularity) for the different reclamation plots are shown in Table 1. The five retrieved MENs were modular, especially for the reclaimed sites, which had modularity values above 0.6 (Table 1). The time after reclamation varied the topological features of the retrieved MENs (Table 1). A power-law model explained successfully (i.e. R2 > 0.8) the connectivity of the retrieved MENs (Table 1; Fig. S3, see supplemental material), showing that the five MENs in these microbial communities exhibited scale-free behavior. A longer time after reclamation led to higher avgK (Table 1), but the control had the highest avgK. In addition, the MEN retrieved for CK had the closest nodes (i.e. GD was the lowest; Table 1), being GD different for the CK with respect to the plots under reclamation (Table 1). The module number was not statistically different between reclaimed and control plots (Table 1), yet plots under reclamation had higher values for the module number.

3.3.2 Dominant microbial taxa

The five retrieved MENs had a distinct microbial structure and the nodes connectivity changed with the time after reclamation (Fig. 4, Table 2, Fig. S4, see supplemental material). However, each MEN had at least nine main phyla with a node degree above 1. The nine main phyla wereAcidobacteria, Actinobacteria, Bacteroidetes, Chloroflexi, Gemmatimonadetes, Planctomycetes , Proteobacteria, Thaumarchaeotaand Verrucomicrobia .
In the MEN belonging to r8 (Fig. 4A, Table 2), the phylumAcidobacteria , Actinobacteria , Chloroflexi ,Gemmatimonadetes and Proteobacteria had the highest node degree (i.e. node degree between 10 and 20) within this MEN. Within r8’s MEN, high degree nodes were primarily clustered into modules 2 and 3, and the phyla Acidobacteria (OTUs 2 and 178) was detected as the module hub. In the Fig 4A, the interactions between OTUs showed as positive, with few pink red color edges.
In the MEN belonging to r11 (Fig. 4B, Table 2), the nodes with the highest node degree (i.e. node degree between 14 and 25) belonged to the phylum Acidobacteria , Chloroflexi , Planctomycetesand Proteobacteria . These nodes were clustered into modules 1 and 2.
In the MEN belonging to r14 (Fig. 4C, Table 2), the node degrees were similar to those in r8 and r11, but smaller than in CK. The phylumAcidobacteria , Actinobacteria , Chloroflexi ,Nitrospirae , Planctomycetes and Proteobacteria had the highest node degree (i.e. node degree between 11 and 20). The nodes with the highest node degree were clustered into modules 3 and 4. The phyla Nitrospirae (OTU375) was detected as the module hub.
In the MEN belonging to r17 (Fig. 4D, Table 2), the node degrees (i.e. node degree between 22 and 37) were higher than in r8, r11 and r14. The phylum Acidobacteria , Bacteroidetes ,Gemmatimonadetes , Planctomycetes and Proteobacteria(OTU146) had the highest node degree. The nodes with the highest node degree were clustered into modules 2, 3 and 8. The phylaGemmatimonadetes was detected as the module hub, while the phylumAcidobacteria and Planctomycetes were detected as connectors within the network. In the Fig. 4D, although most interactions between OTUs showed as positive (light blue edges), more negatively interactions (pink red edges) than those in Figs. 4A, B and C.
In the MEN belonging to CK (Fig. 4E, Table 2), the top five nodes (i.e. node degree between 35 and 55) belonged to the phylumAcidobacteria , Planctomycetes and Proteobacteria . Nodes with higher node degree were clustered into modules 1 and 3, which included Acidobacteria , Bacteroidetes, Chloroflexi ,Planctomycetes and Proteobacteria . OTUs 128, 46 and 948 were detected as the module hub. In the Fig 4E, the interactions between OTUs became more complicated, with much more negatively interactions (pink red edges) than those in Figs. 4A, B, C and D.

3.3.3 Keystone species: Nodes and hubs

The five retrieved MENs had peripheral species that occupied more than 97% of the total nodes (Fig. 5, Files S1-S5, see supplemental material). Three module hubs were detected for r8 (i.e. OTU 178, OTU2, OTU5380) and r11 (i.e. OTU225, OTU47, OTU7) which belonged to the phylum Acidobacteria and Bacteroidetes, respectively. However, four module hubs were detected for r14 (Acidobacteria ,Bacteroidetes , Nitrospirae and Planctomycetes ), r17 (Bacteroidetes , Gemmatimonadetes andPlanctomycetes ) and CK (Acidobacteria , Chloroflexi ,Planctomycetes and Proteobacteria) (Fig. 5). With regard to connector species, over 16 connectors were detected, being CK the treatment that presented the highest number (Fig. 5). Both module hub and connector species had a wider distribution across microbial species. No network hubs were identified (Fig. 5).

3.3.4 Relationship between MEN and key soil attributes

The connectivity of module 1 was significantly positively correlated with OP (r=0.59, p < 0.05) in r8, while module 5 was negatively correlated with PPO (r=-0.59, p < 0.05) in this plot (Fig. 6A). Module 4 was positively correlated with AK (r=0.63, p<0.05) in r11 (Fig. 6B), while module 5 was positively correlated with URA (r=0.58, p<0.05). However, no positive significant correlations were observed between modules and soil attributes in r14, whereas the negative correlations were found between SOM and module 6, 13 (r=-0.57, p < 0.05; r=-0.65; p < 0.01) (Fig. 6C). In r14, module 3 had the negative correlation with OP (r=-0.53, p < 0.05), while module 10 with PPO (r=-0.55, p < 0.05) (Fig. 6C). AK was negatively correlated with module 1, 11 and 12 (r=-0.57, p < 0.05; r=-0.53, p < 0.05; r= -0.55, p < 0.05) (Fig. 6C). In r17, significant correlations were observed between module 2 and SOM, DHG and URA (r=-0.56, p < 0.05; r=0.71, p < 0.01; r=0.73, p < 0.01) (Fig. 6D). In addition, Module 3 had a positive correlation with pH and DHG (r=0.59, p < 0.05; r=0.85, p < 0.001) (Fig. 6D) in r17, and module 5 was positively correlated with SOM and TN (r=0.61 and 0.63; p < 0.05) but negatively correlated with DHG (r=-0.71; p < 0.01) (Fig. 6D). Moreover, Module 6 had a positive correlation with DHG and URA (r=0.63 and 0.64; p < 0.05), whereas module 7 had a negative correlation with DHG (r=-0.68; p < 0.01) (Fig. 6D). Module 8 was positively correlated with pH, DHG and URA (r=0.58, p < 0.001; r=0.87, p < 0.05; r= 0.76, p < 0.01), while negatively correlated with SOM, AK and TN (r=-0.79, p < 0.001; r=-0.57, p < 0.05; r=-0.56, p < 0.05) (Fig. 6D). Module 9 had a negative correlation with SOM, OP and AK (r=-0.73, p < 0.01; r=-0.56, p < 0.05 and r=-0.61, p < 0.05), but a positive correlation with PPO significantly (r=-0.57, p < 0.05) (Fig. 6D). In the control plot (CK), significant correlations were found between modules 1, 3, 4, 7, 8 and soil pH (r=0.67, p < 0.01; r=-0.78, p < 0.001; r=0.87, p < 0.001; r=-0.57, p < 0.05; r=-0.76, p < 0.05), while modules 1, 3, 4, 7 were correlated with the OP value significantly (r=-0.55, p < 0.05; r=0.64, p < 0.05; r=-0.54, p < 0.05; r=0.57, p < 0.05) (Fig. 6E). Module 5 was negatively correlated with SOM and FDA (r=-0.67, p < 0.01; r=-0.57, p < 0.05) (Fig. 6E), while modules 6 and 7 were also significantly related with DHG, respectively (r=-0.53, p < 0.05; r=0.57, p < 0.05) (Fig. 6E).