Abstract: The diversity and interactions of soil fungal community are the key to maintain the diversity and stability of ecosystem. In this study, we examined the structure, diversity and co-occurrence networks of fungal community in rhizosphere and non-rhizosphere soils of planted and natural Picea asperata forests using high-throughput sequencing technique and bioinformatic methods. The results showed that Inocybaceae and Sebacinaceae were dominant family in soils of planted and natural forests, respectively. At the genus level, Inocybe was dominant one in soils of planted and natural forests. There were significant differences in β-diversity of fungal communities between rhizosphere and non-rhizosphere soils in both planted and natural forests. There were no significant correlations between environmental variables and the relative abundance and α-diversity of fungal communities. Herb layer coverage, soil water content, total organic carbon concentration, and plant species richness played important roles in explaining the variations of β-diversity of fungal communities. Results of the network analysis showed that the negative correlations were dominant among soil fungal communities in natural forest, suggesting that the competition of different groups in natural forest. Moreover, there were more negative correlations in non-rhizosphere soils than in rhizosphere soils, which indicated that fungal communities in non-rhizosphere soils comprised more competitive network structure than in the rhizosphere soils. Biomarker species were identified based on differential abundance analysis. Sebacinaceae was the single shared keystone species in the fungal network which had significant differences among rhizosphere and non-rhizosphere soils of planted and natural forests. Therefore, it is suggested that the variation of differential species in the soil fungal communities between the planted and natural forest might had limited influence on the stability of the community of planted and natural forests.

Key words: subalpine forest, Picea asperata, fungal community, high-throughput sequencing, molecular ecological network