Mammalian intestinal microbiota remain poorly understood despite decades of interest and investigation by culture-based and other long-established methodologies. Using high-throughput sequencing technology we now report a detailed analysis of canine faecal microbiota. The study group of animals comprised eleven healthy adult miniature Schnauzer dogs of mixed sex and age, some closely related and all housed in kennel and pen accommodation on the same premises with similar feeding and exercise regimes. DNA was extracted from faecal specimens and subjected to PCR amplification of 16S rDNA, followed by sequencing of the 5′ region that included variable regions V1 and V2. Barcoded amplicons were sequenced by Roche-454 FLX high-throughput pyrosequencing. Sequences were assigned to taxa using the Ribosomal Database Project Bayesian classifier and revealed dominance of Fusobacterium and Bacteroidetes phyla. Differences between animals in the proportions of different taxa, among 10,000 reads per animal, were clear and not supportive of the concept of a “core microbiota”. Despite this variability in prominent genera, littermates were shown to have a more similar faecal microbial composition than unrelated dogs. Diversity of the microbiota was also assessed by assignment of sequence reads into operational taxonomic units (OTUs) at the level of 97% sequence identity. The OTU data were then subjected to rarefaction analysis and determination of Chao1 richness estimates. The data indicated that faecal microbiota comprised possibly as many as 500 to 1500 OTUs.
Citation: Hand D, Wallis C, Colyer A, Penn CW (2013) Pyrosequencing the Canine Faecal Microbiota: Breadth and Depth of Biodiversity. PLoS ONE 8(1): e53115. https://doi.org/10.1371/journal.pone.0053115
Editor: Christophe Herman, Baylor College of Medicine, United States of America
Received: April 21, 2010; Accepted: November 27, 2012; Published: January 31, 2013
Copyright: © 2013 Hand et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was co-funded by the UK Biotechnology and Biological Sciences Research Council and the WALTHAM Centre for Pet Nutrition. The study was undertaken to support the development of a fundamental understanding of the canine gastrointestinal microbiota using 454-pyrosequencing. Scientists at the WALTHAM Centre for Pet Nutrition including co-authors Corrin Wallis and Alison Colyer contributed to the study design, data collection and analysis, the decision to publish and preparation of the manuscript. The work has no direct commercial value and is submitted in good faith within the conventions of scientific integrity and peer review. http://www.bbsrc.ac.uk/ http://www.waltham.com/.
Competing interests: The authors declare a contribution to funding of this project by WALTHAM Centre for Pet Nutrition, to whom CWP is a consultant and by whom DH was supported in part during the study through partial funding of his studentship. This competing interest does not alter authors‚ adherence to all the PLOS ONE policies on sharing data and materials.
The intestinal microbiota can be defined as the total population of microbial species that inhabit the digestive tract. This community of organisms is increasingly recognised as a major contributor to the digestion and utilisation of foods in the gastrointestinal tract, and a key factor in nutrition, development, immune function and other aspects of host physiology that contribute to health and wellbeing , , , , , , .
The microbiota of the mammalian digestive tract are very numerous, diverse and complex in their composition, comprising at least hundreds, perhaps thousands of interdependent and/or competing species , ,  that are generally incompletely characterised. The most diverse and abundant component of the digestive tract microbiota in monogastric mammals is the community associated with the contents of the large intestine. This community is usually dominated numerically by strictly anaerobic species, typically including members of the Bacteroidetes and Clostridia , . These groups include numerous diverse and often uncultured genera and species. Other members of the Firmicutes (low G+C Gram-positives) including anaerobic cocci are also often abundant  whilst the Actinobacteria and Proteobacteria, the latter including many of the medically important pathogenic genera, are often less abundant but nevertheless significant members of the population.
Many species found in the gastrointestinal tract are largely uncharacterised, due to the difficulty of culturing them routinely in the laboratory , , although efforts are continuing to culture new species, with considerable success , . Nevertheless the huge challenge posed by growth and characterisation of hundreds of often fastidious and highly diverse species has in practice made comprehensive and systematic culture-based analysis of mammalian gut microbiota an unattainable goal; furthermore it is a largely qualitative approach. In contrast, high throughput sequencing, for example by Roche-454 deep pyrosequencing of 16S rDNA amplicon pools now enables quantification of different microbial groups directly, based on the numbers of copies obtained of their signature sequences . Typically, a million or more sequence reads can be obtained per instrument run, and sequence tags can be used to enable reads from numerous different samples to be binned and analysed thus enabling simultaneous processing of large numbers of samples.
A number of recently published studies now describe the exploitation of these technologies to analyse the gut microbiota of humans ,  and other animal species such as non human primates , , mice , pigs , chickens , cats  and dogs , . In humans, where interest in the microbiota is intense as its importance in dietary processing and health is increasingly recognised , the complexity of rigorous studies is compounded by human genetic diversity and the difficulty of defining and controlling dietary, behavioural, environmental and other variables in a study population. In contrast, studies in defined animal populations offer opportunities to control these variables. Dogs for example include inbred lines represented by different breeds. Animals housed and fed under similar and controlled regimes are not subject to many of the confounding variables of human study populations.
Knowledge of canine intestinal microbiota is much less complete than for humans. However, several studies investigating canine microbiota using culture-independent methods have now been published , , , , . High throughput sequencing has been used to investigate the diversity of bacterial and fungal microbiota in canine faeces , to determine the effects of antibiotics on microbial diversity  and to study the effects of diet formulation  and dietary fibre on canine faecal bacterial phylogeny  and functional capacity . However canine gut microbiota are yet to be systematically characterised at the species level, and knowledge based on culture methods , , ,  is difficult to relate to the newer high-throughput culture-independent data. Thus there is no consensus view of the composition of ‘normal’ canine faecal microbiota, or the extent of its variation between individuals. We now describe the analysis of faecal microbiota in a group of miniature Schnauzer dogs housed on the same site and with known dietary intakes and genetic relatedness, and illustrate the potential for such an approach to generate fundamental new insights into canine gut microbiology.
Results and Discussion
Data acquisition and analysis
Pyrosequencing by Roche-454 GS FLX of faecal rDNA amplicons from 11 miniature Schnauzer dogs, including a replicate analysis of the MSs1 sample, yielded 247,501 reads representing 56.7 Mb of sequence. The average read length was 229 bp. After the raw sequence data were filtered for quality (see methods), which removed approximately 20% of reads (51,602 in total), sequences were binned by barcode to yield an average of 17,899± SD 1,434 reads per barcode.
To make an initial assessment of the microbiota in each dog, the reads were analysed using the Bayesian classifier algorithm from the Ribosomal Database Project (RDP, http://rdp.cme.msu.edu/) . This method was chosen to take advantage of the straightforward RDP pipeline, fast analysis and widely used, high quality sequence database. Identification of reads to the genus level with increasingly strict RDP classifier bootstrap values resulted in a large and uneven drop in assigned read numbers between the dogs (Fig. 1). Dogs MSs1 and MSs7 in particular were subject to disproportionate reductions in the proportion of reads that could be classified with increasing filtering stringency based on the bootstrap score. A possible explanation for this observation is given below in relation to the taxonomy of sequence reads from these dogs.
The graph shows RDP bootstrap score (x axis) and number of reads that would be classified to the genus level (y axis) for each dog.
A bootstrap score of 50% has been recommended for genus level identification . However, we decided to reduce this to 30% to maintain a more even representation of reads across the group of dogs, accepting the trade-off that a minority (estimated to be ∼15%, see methods) of classifications will be incorrect.
Composition of microbial communities
First we combined RDP classified data from all eleven dogs to generate an overview of the microbiota composition, as represented by the read sequences. In the following discussion we assume that the presence of sequences reflects the presence of the corresponding organisms in the faecal biomass, with the caveat that DNA from dead bacteria or even potentially from ingested foods or other sources may persist and be detectable in the total faecal DNA fraction. There is very little literature available on which to estimate the stability of dietary DNA during passage through the digestive tract, but there is some evidence of persistence in the GI tract of recombinant (GM) DNA from dietary sources , .
We discovered that at the phylum level (see Table 1), the combined community was dominated by the Fusobacteria (39.17% of reads) followed by the Bacteroidetes (33.36%), Firmicutes (15.81%), Proteobacteria (11.31%), Actinobacteria (0.33%) and several other phyla at lower abundances. The clear abundance of sequences attributable to phylum Fusobacteria is somewhat unusual for a mammalian gut community , , , , and our observations echo findings of others in a study of dogs , although contrasting reports demonstrated greater dominance of Firmicutes, Actinobacteria and Bacteroidetes phyla , . Thus there is little clear concensus on the predominant phyla present in dogs and the significance of an enriched Fusobacteria phylum in the dog microbiota is obscure. Notable in these data is the virtual absence of phylum Bacteroidetes in sample MSs1 and very low numbers in sample MSs7, although this phylum is well represented in all the other dogs; replication of this analysis confirmed the result so it does not appear to be due to any technical error. Furthermore, classification to phylum level of reads that could not be assigned to genera, as shown in Table 2, did not assign any of those reads to the Bacteroidetes phylum.
In addition to the phylum level inter-animal variation evident from Table 1, analyses at the genus level of the faecal microbiota from individual animals revealed considerable divergence in abundances of the major genus level taxonomic groups (Fig. 2). The percentage coefficient of variation (CV) for each genus was calculated on the log10 data, given as 100× SD/Mean. CVs ranged from 6 to 193% (with mean counts of 1347 and 1 respectively); the median CV was 44%. In most of the animals the five most abundant genera represented approximately 60–80% of the bacteria present. However even among the most abundant genera the data were highly variable between animals; only the prominent members of the Fusobacteria phylum (Fusobacterium, Cetobacterium and Ilyobacterium; CVs of 8.6% +/−0.27, 6.2% +/−0.2, and 16.5% +/−13.7 and mean counts of 1406, 1347 and 164 respectively) were among the most abundant groups in all animals. Fusobacteria have been shown to ferment carbohydrates and certain amino acids to produce butyrate, acetate and other volatile fatty acids . Cetobacteria have been little-investigated but have been isolated from human faeces as well as fish and shown to ferment peptides and carbohydrates  and to produce vitamin B12 . Ilyobacter has been characterised as a 3-hydroxybutyrate fermenting anaerobe , . None of the other most abundant genera, Bacteroides, Prevotella and Sutterella, were highly abundant in every one of the dogs tested. Notably, members of the little-investigated β-Proteobacterial genus Sutterella accounted for a large majority of Proteobacteria phylum members. The observed divergence between animals does not support the concept of a ‘core microbiota’ at the genus level in which the major constituents of the community will be more or less universally present at comparable levels in different individual hosts. This finding is in agreement with studies in humans , . A study of microbiota in monozygotic and dizygotic twins  indicated that no species-level phylotype at an abundance ≥∼0.5% of the total number of phylotypes in all of the samples from a total of 154 human individuals was universally present. Faeces collected from Koreans was found to contain even lower levels of shared microbiota, with only 0.005% of species-level phylotypes represented in at least 75% of individuals . Other studies of the human intestinal microbiota have reported the levels of shared microbiota to be slightly higher with 2.1% of species-level phylotypes being present in more than 50% of the samples .
Sample number is shown on the x axis and percentage reads classified on the y axis. Genera with fewer than 100 reads in all samples were pooled and are shown as ‘rare genera’. MSs1#2 denotes replicated read data on the same DNA sample for dog MSs1. The complete data set is tabulated in table S1.
Among prominent genera identified in this study, the combined Prevotella and Bacteroides (phylum Bacteroidetes) abundances tended to be inversely related to phylum Fusobacteria in abundance. We hypothesise that this distribution may relate to ‘competition’ for the same niche by these groups of bacteria. This type of relationship has also been observed in human studies where the combined contributions of Firmicutes and Bacteroidetes have been shown to remain largely constant over time and between individuals . A hypothesis regarding this type of relationship is that the gastrointestinal gene pool remains largely constant throughout life but the microbes themselves are continuously replaced in response to environmental change.
As mentioned above, samples from dogs MSs1 and MSs7 were subject to a disproportionate reduction in the number of reads that could be classified to genus level, as the classifier bootstrap score was increased above 30%. In assignment to phyla of the sequences that could not be classified to genus level using a 30% RDP bootstrap score, faeces from dogs MSs1 and MSs7 yielded more Fusobacteria reads (15–20%) than the other dogs (3–8% of reads) (see Table 2). The literature on isolation or detection of Fusobacteria (with the exception of Fusobacterium prausnitzii, reassigned to the Firmicutes, family Ruminococcaceae as Faecalibacterium prausnitzii in 2002 ) from intestinal contents or faeces is sparse , , , and indicates that these organisms are not often prominent in the human gut microbiota. It has been suggested that their association with the mucosa may make isolation difficult . The low numbers isolated and characterised in the past may also explain the paucity of representative sequences in the RDP database, hence limiting ability to classify these sequences to genus level.
Extent of coverage of microbiota diversity
The overall extent of diversity of the digestive tract microbiota is an important question in terms of understanding its complexity and biological role. The identification of sequence reads by matching them to known sequences in the RDP database clearly has limitations in determining overall diversity, since sequences that do not match well to known organisms have been discarded from the analysis above. This is particularly relevant for identification of microbes from less well studied hosts. We therefore determined the diversity of sequences based on the allocation of reads to ‘operational taxonomic units’ or OTUs, independently from any known sequence homologies, using RDP infernal aligner and complete linkage clustering tools . The OTUs represent notional taxa, and at a 97% sequence identity level (i.e. 3% sequence difference from others), reads within an OTU are likely sampled from the same species and individual OTUs approximate to different species. Data showing diversity in an ecological context, where a habitat may not have been exhaustively sampled, can conveniently be analysed by mathematically modelling the occurrence of ‘new’ and repeat sequences in the sample to predict the total diversity in the system. Such analyses can be based on rarefaction curves, shown in Fig. 3, where it can be seen that by these analyses there appear to be significant additional numbers of OTUs yet to be discovered.
Number of reads is shown on the x axis and number of OTUs at 97% sequence identity on the y axis. Also shown are the technical replicates of the analysis of DNA from MSs1.
It is also possible to predict total numbers of OTUs present in such samples by Chao1 richness estimate as shown in Fig. 4. As might be expected from the smaller proportion of the total microbiota taken up by the five most abundant genera in this animal, MSs9 shows a larger estimate of total microbiota diversity than, for example, MSs1. From the fact that the confidence intervals from these two ChaoI estimates do not overlap we can infer that these dogs differ significantly in their estimated species diversity. Overall it appears that in terms of OTUs, faecal microbiota diversity is likely to range from approximately 500 to 1500 OTUs per animal. However these estimates should be viewed with caution, as errors in Roche-454 sequencing at homopolymeric runs may contribute to an overestimate of diversity in the determination of OTU numbers . Furthermore erroneous sequences arising from, for example, chimeric sequence formation resulting from DNA-DNA hybridisation between unrelated sequences during PCR amplification were not excluded in this study. Since completing this analysis we have assessed chimera occurrence in a very similar data set using the Perseus algorithm  and numbers of sequences discarded as chimeras were small, not exceeding 1% of the total, hence there may be small numbers of chimeric sequences present in our data set. Despite these reservations it is clear that there is substantial variation between animals in the diversity of the faecal microbiota.
Genetic relatedness between animals and variation in microbiota
There were several closely related animals in this study cohort (see Table 3). Dogs MSs3, MSs4, MSs5, MSs6 and MSs10 were siblings, the first four being littermates. MSs5 was the mother of MSs7 and MSs8, while MSs8 was the mother of MSs9 and MSs11. MSE was the father of six of the study dogs. Furthermore there were some dietary differences (see Table 3), but 8 of the 11 dogs were fed Pedigree Adult Small Bite (SB). Although this study was not designed to be statistically powered for assessment of genetic or dietary influence on the microbiota, there were some indications of their effects.
Principal component analysis (PCA) on the log10 (counts +1) of the most abundant genera in the microbiota of individual dogs revealed clustering of samples obtained from genetically related animals (see figure 5). There was an apparent grouping of littermates MSs3, MSs4, MSs5 and MSs6, despite MSs5 receiving a different diet from the other three littermates (see Table 3). The PCA loadings plot (not shown) indicated that this clustering was predominantly due to these dogs having higher levels of Lactobacillus, Streptococcus, Lachnospiraceae, Faecalibacterium, Bacteroides, Lachnospira, Coriobacterineae and Erysipelotrichaceae. In addition, dogs MSs9, MSs10 and MSs11 were more correlated with the group of four littermates than the other dogs. This may be due to the fact that they were also closely related; MSs10 was a sibling of the littermates, MSs9 shared the same father as MSs10 and the littermates and MSs9 and MSs11 shared the same mother. Dogs MSs7 and MSs8 were also littermates and formed another cluster predominantly due to higher levels of Roseburia, Lachnobacterium, Propionigenium, Anaerofilum, Peptostreptococcaceae and Cetobacterium. MSs1 and MSs2 shared the same mother but also received different diets to the majority of other dogs which might also explain their lower level of correlation with other animals. MSs1 received a standard Pedigree wet diet and MSs2 was on a veterinary diet whereas all the other dogs received Pedigree adult small bite (SB) with the exception of MSs5 which was receiving a trial diet at the time of faeces collection.
Each point is labelled with the dog ID followed by the mother ID and is coloured according to the father ID; red (MSC), green (MSD), blue (MSE), black (MSF) and purple (MSG). Dotted circles identify littermates.
The findings are in contrast to other canine studies where no discernable effect of littermates was noted in gut microbial composition in a cohort of six dogs comprising three pairs of littermates . Our data are only indicative and a larger study including more littermates is required to confirm these preliminary findings. However, studies in mice and humans have shown that genetically related individuals have a more similar gut microbial composition than unrelated individuals , . In contrast, a study of gut microbiota in pairs of twins and their mothers found no significant difference in degrees of similarity in gut populations . These contradictory findings suggest that other factors as well as host genotype are likely to contribute to the composition of the intestinal microbial community.
A control was included in the analysis whereby DNA from the same sample (MSs1) was analysed (amplified and sequenced) in duplicate. The distribution of genera within these replicate data sets was more similar than between any two individual animals (figure 5), indicating that the variations seen between animals are not solely due to technical error in the analysis. Clustering of log10 (count +1) of genera showed that the repetitions for MSs1 are correlated to 0.99.
Summary and biological significance
The data described comprise one of the first studies to date of the faecal microbiota in a number of closely related dogs, in terms of both breed and family relationships. Considerable differences between individuals were observed, especially in quantitative terms, between the major groups of bacteria detected, albeit a broad similarity is also perceptible in the dominance of anaerobic organisms in the Fusobacteria and Bacteroidetes phyla. This study also shows that genetically related dogs have a more similar faecal microbial composition than unrelated dogs. Taken together, these finding suggest that genetics along with other factors such as diet and age are likely to contribute to shaping the composition of the canine intestinal microbial community. We emphasise however that due to the lack of prior information about biological variation in this system, we have not achieved the necessary statistical power in this study to claim statistical significance in the majority of our observations.
Materials and Methods
Animals and faeces collection
The eleven study animals were healthy adult Miniature Schnauzer dogs housed at the WALTHAM® Centre for Pet Nutrition, and many of the group were closely related genetically. Details are given in Table 1. The dogs were pair-housed in high quality kennel accommodation which exceeded both Home Office and European regulatory requirements. All dogs had constant inside and outside access, including access to paddocks throughout the day, and received similar levels of on- and off-lead exercise. Dogs had varying degrees of cross-contact from sharing pens and during exercise.
Dogs received a variety of diets as detailed in Table 3. One faecal sample was collected per dog during daily exercise and a cross-section of stool to include both surface and internal content (approx 2 g) was frozen on dry ice no more than 15 minutes following defaecation. The frozen samples were stored at −80°C for between one and two weeks before DNA extraction.
Faecal DNA extraction
Faecal DNA was extracted from faecal samples using a QIAamp DNA Stool Mini kit (Qiagen). A subsample of 190–220 mg of frozen faeces was processed following the ‘Protocol for Isolation of DNA from Stool for Pathogen Detection’ detailed in the manufacturer's instructions, using a lysis temperature of 95°C.
Extracted DNA was eluted from the spin columns in 200 µl of Qiagen AE buffer (10 mM Tris-Cl and 0.5 mM pH 9.0 EDTA). Extracted DNA was then quantified and checked for purity (based on UV absorption spectrum and 260∶280 nm and 260∶230 nm absorption ratios) on a ND1000 spectrometer (Nanodrop Technologies Inc) and samples with poor yields (<∼15 ng/μL) and/or highly aberrant absorption ratios were were re-extracted.
Amplification by PCR of 16S rDNA
A barcoded 16 S rDNA tag approach was used to amplify a ∼500 bp region (bases 28–514, excluding primer annealing regions in the E. coli sequence) which includes the V1, V2 and V3 regions of the 16 S rDNA sequence, although only the 5′∼230 bases including barcode and primer were sequenced which included variable regions V1 and V2. The primer sequences were as follows: forward primer 5′-GCCTCCCTCGCGCCATCAG[N8]AGAGTTTGATYMTGGCTCAG-3′ and reverse primer, 5′-GCCTTGCCAGCCCGCTCAGTIACCGIIICTICTGGCAC-3′. The forward primer comprised, from the 5′ end, the 454 sequencing adapter A, a sample barcode octamer (denoted by N8) and rDNA-specific sequence 27f-YM . The reverse primer comprised from the 5′ end, 454 sequencing adapter B and rDNA-specific sequence I533r . Individual 50 µl PCRs were set up as follows; 25 µl Extensor ready mix (Thermo Scientific), 3 µl of each primer (10 pmol/µl), 1.5 µl Nuclease-Free Water (Promega) and 17.5 µl (35 ng) faecal DNA. Amplification was for 30 cycles with the following conditions: 94°C for 3 min:00s (3:00) followed by 10 cycles of 94°C for 0:45, 55°C for 0:30, 72°C for 1:00, then 19 cycles of 94°C for 0:45, 55°C for 0:30, 72°C for 1:30, and a final cycle of 94°C for 0:45, 55°C for 0:30 and 72°C for 7:00. Amplicon abundance and size were checked using agarose gel electrophoresis and were then purified using a QIAquick PCR Purification Kit (Qiagen).
Purified PCR amplicons from different dogs were pooled on an equimolar basis based on ND1000 spectrometer readings. This pool was then sequenced from a primer annealing to adapter A on a 454 GS FLX sequencer (Roche) using FLX chemistry and picotitre plates following the manufacturer's protocols.
Using local databases and code written in Python, the raw sequence read data were initially filtered to remove sequences below 150 nt, those containing one or more ambiguous bases and those with a mismatch against the 27f-YM primer sequence. Sequences were then binned by barcode sequence and each bin was randomly resampled using the Random module in Python, based on the Mersenne Twister algorithm . This allowed standardisation of sequence read number to 10,000 reads per dog, the highest ‘round number’ that would not exclude any of the dogs, to reduce bias in the subsequent comparative data analysis between dogs. This was deemed necessary because it is well recognised that estimates of species richness are dependent on sample size , , and a recent report indicates that equalisation of sample sizes is crucial in comparisons between samples .
Taxonomic assignment of reads was done using a downloaded copy of the 2.0 version of the Bayesian classifier algorithm from the Ribosomal Database Project (RDP, http://rdp.cme.msu.edu/) . Classification at the phylum and genus levels was done using bootstrap scores ranging from 0% to 100%. A value of 30% was selected because although a significant number of genus identifications are likely to be incorrect, a substantially greater number of correctly identified genera are then included in the output data; the value of a high level of sampling has been emphasised in a recent review . The number of incorrectly classified reads at a bootstrap value of 30% was estimated be approximately 15% based on in silico modelling of the classifications of a set of ∼250 sequences of known taxa trimmed to 250 bp spanning regions V1 and V2 (data not shown). We believe that on balance the use of this bootstrap value increases the validity of the determination of proportions of different genera present.
Operational taxonomic units (OTUs) were determined for the total of 120,000 reads combined from the 10,000 resampled reads per dog, using the RDP infernal aligner and Complete Linkage Clustering tools . Rarefaction curves and Chao1 richness estimates ,  were calculated for each dog using Mothur version 1.8.1 .
Principal component analysis (PCA) was performed on the log10 (counts +1 to allow for zeros) to determine if variability of the most abundant genera (>0.1% of total reads) was associated with gender, mother, father, sibling group or diet. Genera present at <0.1% in all 12 samples were classified in a single group of “rare” taxa, and reads that were unclassified, using the predefined criteria, formed another group. All analyses were performed in SIMCA-P version 10 (Umetrics).
We acknowledge the contribution to data analysis of Michigan State University and the Ribosomal Database Project. We thank the staff of the Centre for Genomic Research, University of Liverpool for sequencing work.
Conceived and designed the experiments: DH CW CWP. Performed the experiments: DH. Analyzed the data: DH CW AC. Contributed reagents/materials/analysis tools: DH AC CW. Wrote the paper: DH AC CW CWP.
- 1. Stecher B, Hardt WD (2008) The role of microbiota in infectious disease. Trends in Microbiology 16: 107–114.
- 2. Flint HJ, Duncan SH, Scott KP, Louis P (2007) Interactions and competition within the microbial community of the human colon: links between diet and health. Environmental microbiology 9: 1101–1111.
- 3. Garrett WS, Gordon JI, Glimcher LH (2010) Homeostasis and inflammation in the intestine. Cell 140: 859–870.
- 4. Clemente JC, Ursell LK, Parfrey LW, Knight R (2012) The impact of the gut microbiota on human health: an integrative view. Cell 148: 1258–1270.
- 5. Honda K, Littman DR (2012) The microbiome in infectious disease and inflammation. Annual review of immunology 30: 759–795.
- 6. Sekirov I, Russell SL, Antunes LC, Finlay BB (2010) Gut microbiota in health and disease. Physiol Rev 90: 859–904.
- 7. Lepage P, Leclerc MC, Joossens M, Mondot S, Blottiere HM, et al.. (2012) A metagenomic insight into our gut's microbiome. Gut. doi:10.1136/gutjnl-2011–301805.
- 8. Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen L, et al. (2005) Diversity of the Human Intestinal Microbial Flora. Science 308: 1635–1638.
- 9. Ley RE, Hamady M, Lozupone C, Turnbaugh PJ, Ramey RR, et al. (2008) Evolution of mammals and their gut microbes. Science 320: 1647–1651.
- 10. Spor A, Koren O, Ley R (2011) Unravelling the effects of the environment and host genotype on the gut microbiome. Nature reviews Microbiology 9: 279–290.
- 11. Gill SR, Pop M, Deboy RT, Eckburg PB, Turnbaugh PJ, et al. (2006) Metagenomic analysis of the human distal gut microbiome. Science 312: 1355–1359.
- 12. Savage DC (1977) Microbial ecology of the gastrointestinal tract. Annu Rev Microbiol 31: 107–133.
- 13. Green BD, Keller M (2006) Capturing the uncultivated majority. Current Opinion in Biotechnology 17: 236–240.
- 14. Greetham HL, Giffard C, Hutson RA, Collins MD, Gibson GR (2002) Bacteriology of the Labrador dog gut: a cultural and genotypic approach. Journal of Applied Microbiology 93: 640–646.
- 15. Duncan SH, Louis P, Flint HJ (2007) Cultivable bacterial diversity from the human colon. Lett Appl Microbiol 44: 343–350.
- 16. Lagier JC, Armougom F, Million M, Hugon P, Pagnier I, et al. (2012) Microbial culturomics: paradigm shift in the human gut microbiome study. Clin Microbiol Infect 18(12): 1185–1193.
- 17. Sogin ML, Morrison HG, Huber JA, Mark Welch D, Huse SM, et al. (2006) Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proceedings of the National Academy of Sciences of the United States of America 103: 12115–12120.
- 18. Dethlefsen L, Huse S, Sogin ML, Relman DA (2008) The pervasive effects of an antibiotic on the human gut microbiota, as revealed by deep 16S rRNA sequencing. PLOS Biol 6: e280.
- 19. Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL, Duncan A, et al. (2009) A core gut microbiome in obese and lean twins. Nature 457: 480–484.
- 20. McKenna P, Hoffmann C, Minkah N, Aye PP, Lackner A, et al. (2008) The macaque gut microbiome in health, lentiviral infection, and chronic enterocolitis. PLOS pathogens 4: e20.
- 21. Yildirim S, Yeoman CJ, Sipos M, Torralba M, Wilson BA, et al. (2010) Characterization of the fecal microbiome from non-human wild primates reveals species specific microbial communities. PLOS One 5: e13963.
- 22. Ravussin Y, Koren O, Spor A, Leduc C, Gutman R, et al. (2012) Responses of Gut Microbiota to Diet Composition and Weight Loss in Lean and Obese Mice. Obesity 20(4): 738–747.
- 23. Kim HB, Borewicz K, White BA, Singer RS, Sreevatsan S, et al. (2011) Longitudinal investigation of the age-related bacterial diversity in the feces of commercial pigs. Veterinary microbiology 153: 124–133.
- 24. Callaway TR, Dowd SE, Edrington TS, Anderson RC, Krueger N, et al.. (2010) Evaluation of the bacterial diversity in the rumen and feces of cattle fed diets containing levels of dried distiller's grains plus solubles using bacterial tag-encoded FLX amplicon pyrosequencing (bTEFAP). J Anim Sci.
- 25. Handl S, Dowd SE, Garcia-Mazcorro JF, Steiner JM, Suchodolski JS (2011) Massive parallel 16S rRNA gene pyrosequencing reveals highly diverse fecal bacterial and fungal communities in healthy dogs and cats. FEMS microbiology ecology 76: 301–310.
- 26. Suchodolski JS, Dowd SE, Westermarck E, Steiner JM, Wolcott RD, et al. (2009) The effect of the macrolide antibiotic tylosin on microbial diversity in the canine small intestine as demonstrated by massive parallel 16S rRNA gene sequencing. BMC Microbiol 9: 210.
- 27. Garcia-Mazcorro JF, Lanerie DJ, Dowd SE, Paddock CG, Grutzner N, et al. (2011) Effect of a multi-species synbiotic formulation on fecal bacterial microbiota of healthy cats and dogs as evaluated by pyrosequencing. FEMS microbiology ecology 78: 542–554.
- 28. Greetham HL (2003) Diversity studies of the canine gastrointestinal microbiota. Theses Reading University School of Food Biosciences.
- 29. Simpson JM, Martineau B, Jones WE, Ballam JM, Mackie RI (2002) Characterization of fecal bacterial populations in canines: effects of age, breed and dietary fiber. Microbial ecology 44: 186–197.
- 30. Suchodolski JS, Camacho J, Steiner JM (2008) Analysis of bacterial diversity in the canine duodenum, jejunum, ileum, and colon by comparative 16S rRNA gene analysis. FEMS Microbiology Ecology 66: 567–578.
- 31. Suchodolski JS, Xenoulis PG, Paddock CG, Steiner JM, Jergens AE (2009) Molecular analysis of the bacterial microbiota in duodenal biopsies from dogs with idiopathic inflammatory bowel disease. Vet Microbiol Nov 10 (E-pub ahead of print).
- 32. Middelbos IS, Vester Boler BM, Qu A, White BA, Swanson KS, et al. (2010) Phylogenetic characterization of fecal microbial communities of dogs fed diets with or without supplemental dietary fiber using 454 pyrosequencing. PLOS One 5: e9768.
- 33. Swanson KS, Dowd SE, Suchodolski JS, Middelbos IS, Vester BM, et al. (2011) Phylogenetic and gene-centric metagenomics of the canine intestinal microbiome reveals similarities with humans and mice. The ISME journal 5: 639–649.
- 34. Balish E, Cleven D, Brown J, Yale CE (1977) Nose, throat, and fecal flora of beagle dogs housed in “locked” or “open” environments. Appl Environ Microbiol 34: 207–221.
- 35. Beasley SS, Manninen TJK, Saris PEJ (2006) Lactic acid bacteria isolated from canine faeces. Journal of Applied Microbiology 101: 131–138.
- 36. Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naive Bayesian Classifier for Rapid Assignment of rRNA Sequences into the New Bacterial Taxonomy. Appl Environ Microbiol 73: 5261–5267.
- 37. Claesson MJ, O'Sullivan O, Wang Q, Nikkila J, Marchesi JR, et al. (2009) Comparative Analysis of Pyrosequencing and a Phylogenetic Microarray for Exploring Microbial Community Structures in the Human Distal Intestine. PLOS One 4: e6669.
- 38. Martin-Orue SM, O'Donnell AG, Arino J, Netherwood T, Gilbert HJ, et al. (2002) Degradation of transgenic DNA from genetically modified soya and maize in human intestinal simulations. Br J Nutr 87: 533–542.
- 39. Wiedemann S, Lutz B, Kurtz H, Schwarz FJ, Albrecht C (2006) In situ studies on the time-dependent degradation of recombinant corn DNA and protein in the bovine rumen. J Anim Sci 84: 135–144.
- 40. Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER, et al. (2006) An obesity-associated gut microbiome with increased capacity for energy harvest. Nature 444: 1027–1031.
- 41. Dowd SE, Sun Y, Wolcott RD, Domingo A, Carroll JA (2008) Bacterial tag-encoded FLX amplicon pyrosequencing (bTEFAP) for microbiome studies: bacterial diversity in the ileum of newly weaned Salmonella-infected pigs. Foodborne Pathog Dis 5: 459–472.
- 42. Ritchie LE, Burke KF, Garcia-Mazcorro JF, Steiner JM, Suchodolski JS (2010) Characterization of fecal microbiota in cats using universal 16S rRNA gene and group-specific primers for Lactobacillus and Bifidobacterium spp. Vet Microbiol 144: 140–146.
- 43. Shah HN, Olsen I, Bernard K, Finegold SM, Gharbia S, et al. (2009) Approaches to the study of the systematics of anaerobic, gram-negative, non-sporeforming rods: current status and perspectives. Anaerobe 15: 179–194.
- 44. Finegold SM, Vaisanen ML, Molitoris DR, Tomzynski TJ, Song Y, et al. (2003) Cetobacterium somerae sp. nov. from human feces and emended description of the genus Cetobacterium. Systematic and applied microbiology 26: 177–181.
- 45. Tsuchiya C, Sakata T, Sugita H (2008) Novel ecological niche of Cetobacterium somerae, an anaerobic bacterium in the intestinal tracts of freshwater fish. Letters in applied microbiology 46: 43–48.
- 46. Stieb M, Schink B (1984) A New 3-Hydroxybutyrate Fermenting Anaerobe, Ilyobacter-Polytropus, Gen-Nov, Sp-Nov, Possessing Various Fermentation Pathways. Archives of Microbiology 140: 139–146.
- 47. Sikorski J, Chertkov O, Lapidus A, Nolan M, Lucas S, et al. (2010) Complete genome sequence of Ilyobacter polytropus type strain (CuHbu1). Standards in genomic sciences 3: 304–314.
- 48. Biagi E, Nylund L, Candela M, Ostan R, Bucci L, et al. (2010) Through ageing, and beyond: gut microbiota and inflammatory status in seniors and centenarians. PLOS One 5: e10667.
- 49. Claesson MJ, Cusack S, O'Sullivan O, Greene-Diniz R, de Weerd H, et al.. (2010) Composition, variability, and temporal stability of the intestinal microbiota of the elderly. Proc Natl Acad Sci U S A. doi:10.1073/pnas.1000097107.
- 50. Nam YD, Jung MJ, Roh SW, Kim MS, Bae JW (2011) Comparative analysis of Korean human gut microbiota by barcoded pyrosequencing. PLOS One 6: e22109.
- 51. Tap J, Mondot S, Levenez F, Pelletier E, Caron C, et al. (2009) Towards the human intestinal microbiota phylogenetic core. Environmental microbiology 11: 2574–2584.
- 52. Duncan SH, Hold GL, Harmsen HJ, Stewart CS, Flint HJ (2002) Growth requirements and fermentation products of Fusobacterium prausnitzii, and a proposal to reclassify it as Faecalibacterium prausnitzii gen. nov., comb. nov. International journal of systematic and evolutionary microbiology 52: 2141–2146.
- 53. Walter J, Margosch D, Hammes WP, Hertel C (2002) Detection of Fusobacterium species in human feces using genus-specific PCR primers and denaturing gradient gel Electrophoresis. Microbial Ecology in Health and Disease 14: 129–132.
- 54. Strauss J, White A, Ambrose C, McDonald J, Allen-Vercoe E (2008) Phenotypic and genotypic analyses of clinical Fusobacterium nucleatum and Fusobacterium periodonticum isolates from the human gut. Anaerobe 14: 301–309.
- 55. Swidsinski A, Dorffel Y, Loening-Baucke V, Theissig F, Ruckert JC, et al. (2011) Acute appendicitis is characterised by local invasion with Fusobacterium nucleatum/necrophorum. Gut 60: 34–40.
- 56. Probert HM, Gibson GR (2002) Bacterial biofilms in the human gastrointestinal tract. Current issues in intestinal microbiology 3: 23–27.
- 57. Cole JR, Wang Q, Cardenas E, Fish J, Chai B, et al. (2009) The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucl Acids Res 37: D141–145.
- 58. Quince C, Lanzen A, Curtis TP, Davenport RJ, Hall N, et al. (2009) Accurate determination of microbial diversity from 454 pyrosequencing data. Nat Methods 6: 639–641.
- 59. Quince C, Lanzen A, Davenport RJ, Turnbaugh PJ (2011) Removing noise from pyrosequenced amplicons. BMC Bioinformatics 12: 38.
- 60. Zoetendal EG, Akkermans ADL, Akkermans van-Vliet WM, de Visser JAGM, de Vos WM (2001) The host genotype affects the bacterial community in the human gastrointestinal tract. Microbial Ecol Health Dis 13: 129–134.
- 61. Ley RE, Backhed F, Turnbaugh P, Lozupone CA, Knight RD, et al. (2005) Obesity alters gut microbial ecology. Proceedings of the National Academy of Sciences of the United States of America 102: 11070–11075.
- 62. Frank JA, Reich CI, Sharma S, Weisbaum JS, Wilson BA, et al. (2008) Critical Evaluation of Two Primers Commonly Used for Amplification of Bacterial 16S rRNA Genes. Appl Environ Microbiol 74: 2461–2470.
- 63. Watanabe K, Kodama Y, Harayama S (2001) Design and evaluation of PCR primers to amplify bacterial 16S ribosomal DNA fragments used for community fingerprinting. Journal of Microbiological Methods 44: 253–262.
- 64. Makoto M, Takuji N (1998) Mersenne twister: a 623-dimensionally equidistributed uniform pseudo-random number generator. ACM Trans Model Comput Simul 8: 3–30.
- 65. Youssef NH, Elshahed MS (2008) Species richness in soil bacterial communities: a proposed approach to overcome sample size bias. Journal of microbiological methods 75: 86–91.
- 66. Hughes JB, Hellmann JJ (2005) The application of rarefaction techniques to molecular inventories of microbial diversity. Methods in enzymology 397: 292–308.
- 67. Lemos LN, Fulthorpe RR, Triplett EW, Roesch LF (2011) Rethinking microbial diversity analysis in the high throughput sequencing era. Journal of microbiological methods 86: 42–51.
- 68. Chao A (1984) Nonparametric estimation of the number of classes in a population. Scandinavian Journal of Statistics 11: 265–270.
- 69. Hill TCJ, Walsh KA, Harris JA, Moffett BF (2006) Using ecological diversity measures with bacterial communities. FEMS Microbiology Ecology 43: 1–11.
- 70. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, et al. (2009) Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol 75: 7537–7541.