Table 1.
Bacterial composition of mock communities (in percentage of each taxona).
Table 2.
Food and mock sample names and sequencing dataset nomenclature.
Fig 1.
Rarefaction curves obtained from16S rDNA and gyrB amplicon sequencing of the three repeats for each food sample and the five mock communities.
The x-axis represents the sequencing depth in number of reads while the y-axis represents an estimation of the OTU richness detected. Samples are presented separately, with each panel representing data from one food type or the mock communities. Rarefaction curves for the MC5 community are specifically indicated on the far-right panel.
Table 3.
Comparison of bacterial richness between 16S rDNA and gyrB amplicon sequencing.
Fig 2.
Relative abundance of gyrB and parE sequences assigned to the fifty most-abundant genera recovered in the food samples and mock communities.
The limits of the uncertainty intervals correspond to the upper and lower standard deviation of the average proportion of gyrB reads obtained across the various samples in which each taxon was detected.
Fig 3.
Unrooted phylogenetic tree constructed with the partial gyrB and parE sequences extracted from the genomes of bacterial species present in the mock communities.
Branch length scale indicates the number of nucleotide substitutions per site. The tree is divided into three main branches: parE genes from Proteobacteria, gyrB genes from Proteobacteria, and parE/gyrB genes from Firmicutes. The two sub-branches discriminating between parE and gyrB genes from Firmicutes are equally distant from the gyrB branch of Proteobacteria.
Fig 4.
Principal coordinates analyses (PCoAs) based on Bray-Curtis distances among communities recovered by16S rRNA and gyrB gene sequencing of (A) mock communities and (B) food samples.
All OTUs within the communities were identified to the level of genus. The first and second axes of the PCoA performed for MC samples explained a high degree of the influence of OTUs on communities, with respectively 50.8% and 31.3% of the total variance. Similarly, the first and second axes of the PCoA performed on food sample communities explained, respectively, 43.9% and 26.3% of the total variance.
Fig 5.
Composition plots of relative abundances of OTUs generated by 16S rRNA and gyrB sequencing (A) within Bacteria at the phylum level, (B) within Firmicutes at genus level, and (C) within Proteobacteria at genus level.
Samples from a given food product are presented together (subpanels within A-C) and the two marker analyses for each sample are presented next to each other. Within each panel, the ratio of each taxon was estimated from the sum of all taxa (within all phyla, within Firmicutes, or within Proteobacteria, respectively).
Fig 6.
Comparative analysis of quantification of bacterial species by qPCR versus the estimated absolute number of reads obtained by amplicon sequencing.
(A) Results obtained with 16S rDNA and gyrB amplicon sequencing are shown in the upper plots (in pink and cyan, respectively) with data separated by phylum: Firmicutes (left) and Proteobacteria (right). (B) Boxplot showing the deviation from the linear regression model of the quantification obtained with 16S rDNA (pink) or gyrB (dark cyan) for several bacterial species in the food samples. For the purpose of quantification of the 16S rDNA data, species were merged to genus or broad infra-genus phylogenetic clades because the obtained OTUs could only be assigned to these levels (see main text on subspecies-level bacterial richness).
Fig 7.
Heatmap showing the ability of (A) 16S rDNA and (B) gyrB amplicon analysis to estimate intraspecies population levels.
Samples are ordered from left to right according to the sample type. The scale on the right of the heatmap depicts the color palette associated with the relative numbers of reads of the various OTUs. OTUs are labeled with their cluster number and the taxonomic assignment at the species level. The gyrB/parE OTUs associated with the strains used in the mock communities are labeled with (#). Boxes are drawn around the main phylogenetic clades that are detailed in the text.
Fig 8.
Unrooted neighbor-joining phylogenetic tree showing the evolutionary relationship between gyrB or parE sequences from OTUs in the current study and those from the published genomes of sequenced strains.
OTUs are labeled with their cluster numbers; sequences extracted from published genomes are labeled with the strain name followed by the species name. Strains used in the mock communities are indicated in bold type and tagged with (#). The tree nodes where OTUs and strain sequences cluster are indicated by an open circle (O). Brackets drawn on the right of major gyrB/parE clades indicate the identities of the corresponding 16S phylogenetic clades.