Table 1.
Origin and characteristics of the soil and wood samples.
Fig 1.
Taxon distribution (phylum or sub-phylum levels) in the eDNA and eRNA datasets.
Taxon distribution is visualized for the three studied habitat, forest soil, grassland soil and decaying wood. The “whole shared ds” corresponds to the distribution in the entire dataset (all habitats together). Note the differences in abundances for Glomeromycotina and Rozellomycota between the grassland soil RNA and DNA datasets.
Fig 2.
Fungal classes significantly differ from each other with respect to MOTU RNA:DNA read ratios.
For each MOTU (black dots), its log2-transformed ratio ([No. of RNA reads]: [No. of DNA reads] +1) in the whole dataset was plotted on a horizontal axis. Symbol size is proportional to the relative abundance (average reads number among the 24 samples) of the taxa in the dataset. For Agaricomycetes we separately considered symbiotic (symb, mainly ectomycorrhizal), saprophytic and undefined (others) MOTUs. Red circles correspond to the mean values and red bars to standard deviations. In the case of Agaricomycetes, the global mean value for this class (symbiotic + non-symbiotic species together) is indicated by a black dot with red margin. Identical letters above the mean values (a, b or c) indicate which of the distributions are statistically similar (P > 0.05; Kruskall-Wallis test and Dunn’s post hoc test).
Fig 3.
Correlations between DNA and RNA MOTUs read abundance in each of the three habitats.
Log2-transformed DNA and RNA reads abundances of each MOTU are plotted against each other. Linear coefficient of correlations (R2) and their level of significance, as well as linear trend-lines (dashed lines) are given.
Table 2.
Pairwise comparisons (two ways PERMANOVAs) between habitats using different MOTU datasets.
Fig 4.
Non-Metric Multidimensional Scaling (NMDS) ordination of soil and decaying wood samples (DNA and RNA).
Analysis was performed using MOTUs abundances in the “shared dataset” and Bray-Curtis indices. Convex hulls cluster samples according to habitat type. NMDS stress value = 0.048.
Fig 5.
Differential abundance analysis.
(A) Ternary plots illustrating the distribution of individual MOTUs (closed circles whose sizes reflect their abundance in terms or read numbers) in each of the three studied habitats. Plots were drawn separately for the DNA and RNA datasets to illustrate that the relative abundance of taxa in the three habitats varies depending on the nucleic acid used for metabarcoding. Taxa over-represented (following differential abundance analysis) in either forest soil, grassland soil or decaying wood are identified using a specific color code. (B) Pie charts that illustrate the taxonomic distribution (phylum level) and ecology (according to [Nguyen 2016 [66]]) of significantly over-represented MOTUs, identified at the species level, in each of the three habitats. (C) Mean relative abundance heatmap of enriched MOTUs in decaying wood (brown), forest soils (green) and grassland soils (yellow) across DNA and RNA samples from the different habitats. (D) Similarities between datasets was calculated using Spearman distances between samples and rows were clustered by MOTUs enriched in each habitat type. Displayed mean-relative abundance values were cut at 100 rarefied reads threshold.
Table 3.
Taxonomic origin and trophic modes of the differentially abundant MOTUs identified in the three habitats.