Figure 1.
Absolute (A) and relative (B) bacterial, archaeal, fungal and ciliate protozoal marker loci copy numbers.
Absolute numbers are expressed per gram dry weight of rumen contents collected from a hay-fed cow and a pasture-fed sheep from which DNA was extracted in triplicate using nine different methods (Table 1). Relative numbers are shown as a proportion of bacterial locus copies. Values depicted are means and standard deviations of log-transformed data. The vertical bars indicate one standard deviation. Those that do not share a letter at the base of the bar are significantly different (p < 0.05, ANOVA, Scheffe post hoc test).
Figure 2.
Relative (A) bacterial and (B) archaeal community compositions in rumen samples.
The means (% of total community) and standard deviations, from the triplicate determinations, of the relative contribution of each microbial group in DNA obtained using the nine best extraction method are shown. The keys to the right indicate the major community components. The underlying data for bacteria, archaea, fungi and ciliate protozoa are provided in Table S3. Mbb, Methanobrevibacter.
Figure 3.
Consensus dendrogram illustrating the similarity of microbial communities obtained using different DNA extraction methods.
For each rumen sample and microbial group, a matrix of pair-wise Pearson similarities was constructed, tabulating the similarity of the community structure determined using each different DNA extraction method with the structure determined using DNA from every other method. These matrices were converted to distances to produce six matrices (bacterial genera [sheep and cow], archaeal mixed taxonomic ranks [sheep and cow], ciliate protozoal genera [cow only], and fungal subgenera [cow only]), which were used to generate six trees using the UPGMA algorithm [23], from which a consensus tree was formed using the CONSENSE program in PHYLIP [23]. The scale bar represents 1% community difference.
Figure 4.
Comparison of microbial community compositions following extraction of DNA with different methods.
Principal coordinate analyses of Bray-Curtis dissimilarities of A) bacterial communities (phylum level) and B) archaeal communities (mixed taxonomic ranks) in cow and sheep rumen samples are shown. The data from each of the individual triplicate extractions performed are plotted. The keys to the right indicate the different DNA extraction methods used. The values in parentheses give the amount of variation explained by each coordinate.
Figure 5.
Impact of rumen sampling method on the (A) bacterial and (B) archaeal rumen microbiota composition.
The apparent microbial community structure in parallel samples collected from 14 cattle using an oral stomach tube or through a rumen fistula was compared. The means (% of total community) and standard errors of the relative contribution of each microbial group are shown. The keys to the right indicate the major community components. The underlying data for bacteria, archaea, fungi and ciliate protozoa are provided in Table S7. Mbb, Methanobrevibacter. Differences between rumen sampling methods were assessed using dependent sample t-tests.
Figure 6.
Impact of rumen sample fractionation on the (A) bacterial and (B) archaeal rumen microbiota composition.
The apparent microbial community structure in samples collected from 16 cattle were separated into liquid, solid and total rumen sample fractions and the apparent microbial community structures were compared. The means (% of total community) and standard errors of the relative contribution of each microbial group are shown. The keys to the right indicate the major community components. The underlying data for bacteria, archaea, fungi and ciliate protozoa are provided in Table S8. Mbb, Methanobrevibacter. Differences between between total, liquid and solid rumen sample fractions were assessed using dependent sample t-tests.