Fig 1.
Schematic representation of the sampling and processing protocol for 16S amplicons libraries.
Fig 2.
Rarefaction curves (A), alpha (B) and beta diversity (C) in 16S amplicon libraries. Alpha diversity (A) rarefaction curves were calculated on unfiltered data, while alpha diversity indices (B) were calculated using the Shannon and Simpson index. Letters above boxplots indicate statistical significance using a pairwise Wilcoxon test (adjusted with the Bonferroni method). The p-value of a PERMANOVA test on beta diversity is indicated at the bottom of (C). Coloured circles on the NDMS are ellipses of confidence (at 95%) for all media and CIA conditions. Abbreviations: CIA: culture-independent approach; media used in the CDA: RF: rice flour, NFB: nitrogen-free medium, NGN: Norris glucose N-free medium, TSA10 and TSA50: tryptone soya agar at 10 and 50%.
Table 1.
Comparison of diversity in culture-dependent (CDA) and culture-independent (CIA) approaches, using ASV or OTU analysis.
Fig 3.
Taxonomic binning at phylum (A, top 30), class (B, top 25), order (C, top 25) and genus (D, top 25) levels (with most abundant taxa in bold), and a circular phylogenetic tree of ASV with the class-rank distribution among CIA and CDA (E).
Fig 4.
Venn diagrams of the diversity between CDA and CIA.
Venn diagrams were produced at the ASV level (A), genus level (B), and between culture media used for the CDA (C). Specific genera obtained on a given culture medium are listed in (C).
Table 2.
List of the top 50 most abundant bacterial genera in the culture-independent approach (CIA), and their occurrence in media of the culturable-dependent approach (CDA).
Table 3.
Distribution and mean relative abundance of the top 20 genera detected in the CDA.
Fig 5.
Predicted enriched pathways in amplicon libraries.
Nitrogenase enrichment prediction in 16S amplicon libraries (A) and dot-plot of predicted enriched Metacyc pathways (B) in CIA vs CDA cross comparisons, non-selective media (TSA, RF) versus others, and nitrogen-free media versus others. Enzyme and pathway enrichment were predicted with PICRUSt2 and a differential analysis was performed with a Kruskal-Wallis test. “diff” is the abbreviation for differential analysis (see Materials & Methods).