Fig 1.
Scheme of the MG-RAST analysis, showing the levels of hierarchy.
Refseq database was selected for taxonomic distributions (Phylum, Class, Order, Family, and Genus) and SEED Subsystems database was used for the genetic functional analysis (Levels 1 to 4). Total number of variables obtained for each level is in parentheses. “Virulence, Disease and Defense” and its subsequent levels are shown as an example.
Fig 2.
The top 20 most relatively abundant phyla.
The top 20 most abundant phyla by sample day for each study group (CON = control, ENR = enrofloxacin, TUL = tulathromycin).
Fig 3.
Linear discriminant analysis of the 50 most abundant phyla.
Changes in phyla composition over time (P-value < 0.05). An ellipse indicates the 95% confidence region to contain the true mean of the variable (day).
Fig 4.
Relative abundance of genera Bacteroides, Blautia, Coprococcus and Desulfovibrio by sample day represented by least square means.
The independent variables study group (CON = control, ENR = enrofloxacin, TUL = tulathromycin), sample day (4, 14, 56, and 112) and interactions were included as fixed effects in all models. Day 0 was included as a covariate in the model. The effects block and individual animals nested within block were controlled in the models as random effects. Asterisks indicate significant differences (P-value ≤ 0.05) between day and study group. Error bars represent the standard error of the least square mean.
Fig 5.
Linear discriminant analysis of the 28 microbial functions from Level 1.
Changes in microbial genetic function over time (P-value < 0.05). An ellipse indicates the 95% confidence region to contain the true mean of the variable (day).
Fig 6.
Relative abundance of genes with “Clustering-based subsystems” function by sample day represented by least square means.
The independent variables study group (CON = control, ENR = enrofloxacin, TUL = tulathromycin), sample day (4, 14, 56, and 112) and interactions were included as fixed effects in all models. Day 0 was included as a covariate in the model. The effects block and individual animals nested within block were controlled in the models as random effects. Asterisks indicate significant differences (P-value ≤ 0.05) between day and study group. Error bars represent the standard error of the least square mean.
Fig 7.
Relative abundance of genes with “Transposable elements” function by sample day represented by least square means.
The independent variables study group (CON = control, ENR = enrofloxacin, TUL = tulathromycin), sample day (4, 14, 56, and 112) and interactions were included as fixed effects in all models. Day 0 was included as a covariate in the model. The effects block and individual animals nested within block were controlled in the models as random effects. Asterisks indicate significant differences (P-value ≤ 0.05) between day and study group. Error bars represent the standard error of the least square mean.
Fig 8.
Boxplots of genes with “Resistance to antibiotics and toxic compounds” function.
By sample day and study group (CON = control, ENR = enrofloxacin, TUL = tulathromycin).
Fig 9.
Relative abundance of genetic functional categories within “RATC” by sample day represented by least square means.
The independent variables study group (CON = control, ENR = enrofloxacin, TUL = tulathromycin), sample day (4, 14, 56, and 112) and interactions were included as fixed effects in all models. Day 0 was included as a covariate in the model. The effects block and individual animals nested within block were controlled in the models as random effects. Asterisks indicate significant differences (P-value ≤ 0.05) between day and study group. Error bars represent the standard error of the least square mean.
Fig 10.
Relative abundance of genetic functional categories at level 4, within “RATC”, by sample day represented by least square means.
“DNA gyrase subunit A (EC 5.99.1.3)” and “rRNA adenine N-6-methyltransferase (EC 2.1.1.48)”. The independent variables study group (CON = control, ENR = enrofloxacin, TUL = tulathromycin), sample day (4, 14, 56, and 112) and interactions were included as fixed effects in all models. Day 0 was included as a covariate in the model. The effects block and individual animals nested within block were controlled in the models as random effects. Asterisks indicate significant differences (P-value ≤ 0.05) between day and study group. Error bars represent the standard error of the least square mean.