3.2 Soil quality index
PCA results showed that the first three components with eigenvalues ≥1.0 explained 87% of the total variation of soil properties (Table S1). According to the Norm value, SOC, BD, DOC, TN, NH4+, C/N, ACPT, ACP and ALP were the highly weighted variables in the PC1 and highly correlated with each other. Thus, SOC was selected in the PC1 with the highest eigenvector. Similarly, GLS and TP were selected for PC2 and PC3, respectively. In short, SOC, GLS and TP were the three important indicators selected for MDS2. The SQI was calculated after scoring and weighting the four indicators (Table S2). The SQI values derived with the MDS method ranged from 0.16 to 0.92. After conversion, the SQI of Forest decreased sharply by 65%. Within four plantations, the highest values were observed in the Peach, the lowest in the Oil (Fig. 2).
3.3 Diversity, composition and functional guilds of the fungal community
After quality filtering, we obtained 1,252,350 sequences for downstream analysis (range, 83,490 ± 369 reads per sample) from a total of 1,251,620 raw sequencing reads. A total of 3,933 OTUs were clustered with ≥ 97% identity from good-quality sequences. Rarefaction curves demonstrated that the sequencing depth was sufficient to capture species richness of the community (Fig. S1). Following forest conversion, Observed species, Shannon-Wiener and Chao1 indices of α-diversity increased by 3.5%, 6.5% and 2.9%, respectively. All indices were higher in Berry and Peach than the other three plantations (p < 0.05; Fig. 3a, 3b, 3c). Moreover, Chao1, Observed species, and Shannon-Wiener indices increased with pH, UR and Mg (p <0.05; Table S3). According to NMDS analysis, the dissimilarity of the fungal community structure changed substantially after conversion (Fig. 3d).
In total, 15 fungal phyla and 696 genera were identified but only 4 phyla and 20 genera were recorded with relative abundances of over 1%. Ascomycota, Basidiomycota, Glomeromycota and Zygomycota were dominant phyla, accounting for approximately 87% of the fungal sequences (Fig. 4a). After conversion, phyla such as Ascomycota, Zygomycota, and Glomeromycota increased, but Basidiomycota and Rozellomycota decreased by 57% and 32%, respectively (Fig. 4a). Ascomycota increased with TK, but decreased with NO3-(p < 0.05; Table S5). Basidiomycota declined with SOC, DOC, C/N, NH4+, GLS and ACP, and Rozellomycota fell with NO3-, UR, GLS and Ca, while Glomeromycota increased with BD (p<0.05; Table S5). At the genus level, relative abundance ofPseudochaetosphaeronema and Russula decreased by approximately 99%, while Pseudophialophora, Rhytisma andLepidostroma increased after conversion (Fig. 4b; Table S4).
Most fungi were classified as saprotrophs (35%), undefined saprotrophs (29%) or ectomycorrhizal fungi (13%). Comparisons within the top 35 guilds revealed that three functional guilds (ectomycorrhizal, wood saprotroph, fungal parasites) were more abundant in Forest than in plantations, whereas undefined saprotrophs, arbuscular mycorrhizal, bryophyte parasite and plant pathogen fungi were enriched in plantations (Fig. 5).
The majority of the top 50 OTUs were either broadly dispersed across five sites (32%) or only occurred in the plantations (32%) (Fig. 6). Specific OTUs for Forest, Oil, Peach, Berry and Fir plantations were six, five, one, four and two, respectively. Specifically,Pseudochaetosphaeronema OTU_4 and Russula OTU_7 were distinct from others for Forest. Biscogniauxia OTU_55 andInocybe OTU_25 were special to Oil. Cistella OTU_100 was unique to Peach. Simplicillium OTU_48, SubulicystidiumOTU_125, Dendrochytridium OTU_90 and MelanogasterOTU_88 were four particular taxa for Berry. Lepidostroma OTU_38 and Tomentella OTU_17 were specific for Fir (Fig. 6). LDA analysis showed that the Agaricomycetes class and Russulaceae family were identified as biomarkers between Forest and Oil, Berry, respectively. Compared with the Forest, Russulales order (belonging to Basidiomycota) was identified as a biomarker in Peach and Fir plantations (Fig. S2).