Are all faecal bacteria detected with equal efficiency? A study using next-generation sequencing and quantitative culture of infants' faecal samples.

Background: Many species of intestinal bacteria are present in moderate numbers in the faecal microbiota, which is dominated by obligate anaerobes. Little is known regarding the detection sensitivity of next-generation sequencing for these microbes in samples of complex microbiota.

Methods: Twenty stool samples from six healthy infants, who were followed from 1 week to 1 year of age, were previously cultured quantitatively for total population counts, as well as for counts of relevant facultative bacteria and a limited selection of obligate anaerobes that are prevalent in the neonatal microbiota. The same samples were analysed by Next-generation sequencing (NGS, pyrosequencing) of the 16S rRNA gene (V1-V3 regions; average read length, 500 nucleotides; average number of reads per sample, 30,000). We used the bacterial culture data to determine the lowest bacterial populations that could be detected by NGS. Different DNA extraction kits (QIAamp DNA Stool Mini, ZR Faecal DNA MiniPrep, and PowerSoil DNA Isolation) were compared for efficacy in extracting DNA from Gram-negative and Gram-positive Type strains.

Results: NGS yielded one read per 10 CFU/g faeces of the Gram-negative commensal gut bacteria Bacteroides and Enterobacteriaceae, but only one read per 10 CFU/g faeces of Gram-positive bifidobacteria. The Gram-positive facultative bacteria Enterococcus was often undetectable by DNA-based methods despite being present at >10 CFU/g faeces. The DNA extraction kits tested varied considerably in their ability to extract DNA from bacterial samples, and showed considerably lower efficacies in extracting DNA from Gram-positive than from Gram-negative bacteria.

Conclusions: NGS has lower sensitivity for detecting Gram-positive bacteria than Gram-negative bacteria, due at least in part to inefficient extraction of DNA from Gram-positive bacteria. Therefore, enzymatic lysis may enhance the yield of DNA and increase the sensitivity of NGS methods for Gram-positive bacteria, and the inclusion of positive and negative controls during DNA extraction is indicated for validation purposes. The differential extraction of DNA from bacterial samples by different DNA extraction kits may limit comparability between studies on the gut microbiota. Finally, quantitative culture methods detect certain bacteria with greater sensitivity than NGS techniques, and thus culture- and DNA-based methods can be used in tandem to define the complex composition of the gut microbiota with greater accuracy.