Additional file 10: Table S10. Overview of virulence genes detected in the 2114 MAGs using the virulence genes of interest in the six pathogenic genera (A) and in all taxa (B). From left to right: the percentage, E-value and bitscore of diamond BLAST comparisons, the VFdb identifier, the VFdb preferred gene name, the VFdb taxon, MAG name and alternative name, the metagenomics sample, the country and the study the MAG came from, the dRep cluster identified and the average abundance of the MAG in the 436 samples. The subtable to the right presents the gene prevalence based on occurrence (the times a VFdb gene gave a positive match to the dataset).

Additional file 1: Table S1. Shotgun metagenomic data used in this study. From left to right: accession and bioproject numbers, study, location, sample type, diet, age, farm id and size (whenever available). The table is also accessible at www.fecobiome.com/resources/data by entering the keyword “shotgun” in the “search our database” box. Samples can be further filtered by location or farm name (e.g. using the keywords «Theix» or «Saint-Gene-Champagnelle» will only list the samples sequenced in our study).

Additional file 1: Table S1. Shotgun metagenomic data used in this study. From left to right: accession and bioproject numbers, study, location, sample type, diet, age, farm id and size (whenever available). The table is also accessible at www.fecobiome.com/resources/data by entering the keyword “shotgun” in the “search our database” box. Samples can be further filtered by location or farm name (e.g. using the keywords «Theix» or «Saint-Gene-Champagnelle» will only list the samples sequenced in our study).

Additional file 10: Table S10. Overview of virulence genes detected in the 2114 MAGs using the virulence genes of interest in the six pathogenic genera (A) and in all taxa (B). From left to right: the percentage, E-value and bitscore of diamond BLAST comparisons, the VFdb identifier, the VFdb preferred gene name, the VFdb taxon, MAG name and alternative name, the metagenomics sample, the country and the study the MAG came from, the dRep cluster identified and the average abundance of the MAG in the 436 samples. The subtable to the right presents the gene prevalence based on occurrence (the times a VFdb gene gave a positive match to the dataset).

Additional file 2: Table S2. Summary of the contigs assembly. From left to right: sample name (accession number), number of contigs and number of base pairs, minimum, maximum and average contig length, N50 and mean coverage of contigs (estimated by mapping filtered reads to contigs using Bowtie2).

Additional file 2: Table S2. Summary of the contigs assembly. From left to right: sample name (accession number), number of contigs and number of base pairs, minimum, maximum and average contig length, N50 and mean coverage of contigs (estimated by mapping filtered reads to contigs using Bowtie2).

Additional file 3: Table S3. A: Summary of the assembled contig classification: Number and proportion of contigs assigned to each domain, for each sample. The absolute figures are shown at the left and the percentages at the right. The average percentage for each domain is also presented in the small subtable to the right. B: Archaeal sequences abundance estimate: Number of reads properly paired to the 155 RUGs previously presented [53] and the assembled contigs identified as archaea in our study. The absolute numbers and percentages are presented for each sample. Mean percentages and medians across all samples are presented to the right.

Additional file 3: Table S3. A: Summary of the assembled contig classification: Number and proportion of contigs assigned to each domain, for each sample. The absolute figures are shown at the left and the percentages at the right. The average percentage for each domain is also presented in the small subtable to the right. B: Archaeal sequences abundance estimate: Number of reads properly paired to the 155 RUGs previously presented [53] and the assembled contigs identified as archaea in our study. The absolute numbers and percentages are presented for each sample. Mean percentages and medians across all samples are presented to the right.

Additional file 4: Table S4. Overview and characteristics of the 2114 MAGs, the 74 Hungate and the 102 RefSeq genomes used to reconstruct phylogenies in our study. From left to right: the internal code (a derivative of the accession number the data came from and the autometa id); the alternative name, which we either generated by identifying the closest match following BLAST comparisons with RefSeq or the assigned strain name (for reference genomes downloaded from databases); the accession number the MAG came from (when relevant); country where the sample was collected; author; dRep clustering and checkm (estimated completeness and contamination) and mean contig coverage (estimated by mapping filtered reads to contigs using Bowtie2). The overall mean and median are presented as well. The MAGs with the same drep clustering id are >99% identical. The last column indicates the 1232 MAGs that were selected for the dereplicated dataset. The sequences of the 2114 MAGs are available here: doi.org/10.15454/UIJTJA . B: Mapping of filtered reads to contigs and the core MAGs. From left to right: Sample names, the total number of reads for each sample and the number of reads successfully mapped to contigs, the percentage of the number of reads mapped to contigs, the number of reads mapped to MAGs, and the percentage of reads mapped to MAGs. The smaller subtable to the right presents: the average percentage of reads successful mapping to contigs and to MAGs (top); the average percentage of reads mapped only to bacterial contigs (based on the classification from Autometa; NR database); the percentage of bacterial reads (based on the bacterial contig estimate) mapped to MAGs.

Additional file 5: Table S5. PERMANOVA tests examining the impact of the different study and sample type (in two independent studies) on the fifteen class-specific reconstructed MAG phylogenies. Dissimilarity matrices were created using AA alignments, built from the phylogenies presented in Sup. Res. 1. Two independent PERMANOVA tests were performed on each matrix, using the study identifier (author) and sample source type as explanatory variables.