Technical variation in metagenomic analysis must be minimized to confidently assess the contributions of microbiota to human health. Here we tested 21 representative DNA extraction protocols on the same fecal samples and quantified differences in observed microbial community composition. We compared them with differences due to library preparation and sample storage, which we contrasted with observed biological variation within the same specimen or within an individual over time. We found that DNA extraction had the largest effect on the outcome of metagenomic analysis. To rank DNA extraction protocols, we considered resulting DNA quantity and quality, and we ascertained biases in estimates of community diversity and the ratio between Gram-positive and Gram-negative bacteria. We recommend a standardized DNA extraction method for human fecal samples, for which transferability across labs was established and which was further benchmarked using a mock community of known composition. Its adoption will improve comparability of human gut microbiome studies and facilitate meta-analyses.

EEN impacts on gut microbiota composition and changes fecal metabolic activity. It is difficult to infer a causative association between such changes and disease improvement, but the results do challenge the current perception of a protective role for F. prausnitzii in CD.

ABSTRACTQuantitative real-time PCR assays targeting thegroELgene for the specific enumeration of 12 human fecalBifidobacteriumspecies were developed. The housekeeping genegroEL(HSP60in eukaryotes) was used as a discriminative marker for the differentiation ofBifidobacterium adolescentis,B. angulatum,B. animalis,B. bifidum,B. breve,B. catenulatum,B. dentium,B. gallicum,B. longum,B. pseudocatenulatum,B. pseudolongum, andB. thermophilum. The bifidobacterial chromosome contains a single copy of thegroELgene, allowing the determination of the cell number by quantification of thegroELcopy number. Real-time PCR assays were validated by comparing fecal samples spiked with known numbers of a givenBifidobacteriumspecies. Independent of theBifidobacteriumspecies tested, the proportion ofgroELcopies recovered from fecal samples spiked with 5 to 9 log10cells/g feces was approximately 50%. The quantification limit was 5 to 6 log10groELcopies/g feces. The interassay variability was less than 10%, and variability between different DNA extractions was less than 23%. The method developed was applied to fecal samples from healthy adults and full-term breast-fed infants. Bifidobacterial diversity in both adults and infants was low, with mostly ≤3Bifidobacteriumspecies andB. longumfrequently detected. The predominant species in infant and adult fecal samples wereB. breveandB. adolescentis, respectively. It was possible to distinguishB. catenulatumandB. pseudocatenulatum. We conclude that thegroELgene is a suitable molecular marker for the specific and accurate quantification of human fecalBifidobacteriumspecies by real-time PCR.

SummaryNon-digestible milk oligosaccharides were proposed as receptor decoys for pathogens and as nutrients for beneficial gut commensals like bifidobacteria. Bovine milk contains oligosaccharides, some of which are structurally identical or similar to those found in human milk. In a controlled, randomized doubleblinded clinical trial we tested the effect of feeding a formula supplemented with a mixture of bovine milkderived oligosaccharides (BMOS) generated from whey permeate, containing galacto-oligosaccharides and 3'-and 6'-sialyllactose, and the probiotic Bifidobacterium animalis subsp. lactis (B. lactis) strain CNCM I-3446. Breastfed infants served as reference group. Compared with a non-supplemented control formula, the test formula showed a similar tolerability and supported a similar growth in healthy newborns followed for 12 weeks. The control, but not the test group, differed from the breast-fed reference group by a higher faecal pH and a significantly higher diversity of the faecal microbiota. In the test group the probiotic B. lactis increased by 100-fold in the stool and was detected in all supplemented infants. BMOS stimulated a marked shift to a bifidobacteriumdominated faecal microbiota via increases in endogenous bifidobacteria (B. longum, B. breve, B. bifidum, B. pseudocatenulatum).

The gut microbiota is implicated in the pathogenesis of Crohn's Disease (CD) (1) . Exclusive enteral nutrition (EEN) is a successful treatment but its mode of action remains unknown (2) . We assessed changes in the fecal microbiota milieu during EEN. Five faecal samples were collected from CD children; 4 during EEN (start, 15, 30, end EEN∼60 days) and a fifth on free habitual diet. Two samples were collected from healthy controls. Faecal pH, bacterial metabolites ( Figure) and quantitative changes (q-PCR) of total and 7 major bacterial groups implicated in CD were measured. 68 samples were from 15 CD children and 40 from 21 controls. Faecal pH and total sulphide increased and butyric acid decreased during EEN (Figure). F. prausnitzii spp. concentration significantly decreased after 30 d on EEN (Figure). In patients who responded to EEN, the magnitude of the observed changes was greater and the concentration of Bacteroides/Prevotella group also decreased. All these changes reverted to pre-treatment levels when the children returned to their free habitual diet.EEN impacts on gut microbiota composition and changes faecal metabolic activity. It is difficult to infer a causative association between such changes and disease improvement but the results do challenge the current perception of a protective role for F. prausnitzii in CD.