2.3 Statistical analysis
Two-way ANOVAs were used to analyze the effects of varied tree ages,
soil depths, and their interaction on the soil properties and microbial
alpha diversities (number of OTUs and Simpson’s diversity and
Shannon-Wiener indices). A Pearson correlation analysis assessed the
association between microbial alpha diversity and environmental factors.
A value of p < 0.05 was considered significant. The
heterogeneity of the variance was tested, and the original data were
normalized by log-transformation or standardization prior to analysis
when necessary. Using non-metric multidimensional scaling (NMDS), the
dynamics of microbial community during the increase of L.
gmelinii plantations age and then assessed through a permutational
multivariate analysis of variance (PERMANOVA). The correlations between
environmental parameters and microbial community profiles were then
assessed through Redundancy analysis (RDA). A Mantel test was used to
assess the correlations of microbial communities and environmental
variables using the Vegan package. A partial Mantel test in the Vegan
package was used to control the covarying effects of various factors. A
classification random forest (RF) analysis was used to identify
microbial taxa associated with stand age in L. gmeliniiplantations. The analysis aimed to test which microbial taxa are
possible predictors of forest age change. Additionally, the functional
profiles of soil prokaryotes and fungi were separately analyzed using
FAPROTAX (Louca et al., 2016) and FUNGuild (Nguyen et al., 2016). To
estimate species coexistence across different L. gmeliniiplantations, metacommunity co-occurrence networks consisting of all the
members of the three microbial groups were constructed. To reduce rare
OTUs in the dataset, we removed OTUs with a relative abundance
<0.5%. Robust correlations with Spearman’s correlation
coefficients (ρ) > 0.7 and false discovery rate-corrected
p-values <0.05 were used to construct networks (Jiao et al.,
2020; Shi et al., 2020). Then, the networks were constructed in R with
the igraph package, and the related network indices were calculated.
Finally, the networks were visualized in Gephi (version 0.9.3;
https://gephi.org/). The node size was proportional to the number of
connections (i.e., degrees), and the node color represented the
microbial taxonomy. The link between each pair of nodes represented a
positive (in green) or a negative (in red) correlation.