16S rRNA gene sequencing on a benchtop sequencer: accuracy for identification of clinically important bacteria.

G.S Watts,D.M Wolk,M.J Slepian,M.M Oshiro,D Dhingra,L.D Cranmer,K Youens‐Clark,B.L Hurwitz,G.S Metzger

doi:10.1111/jam.13590

Abstract

AimsTest the choice of 16S rRNA gene amplicon and data analysis method on the accuracy of identification of clinically important bacteria utilizing a benchtop sequencer.Methods and ResultsNine 16S rRNA amplicons were tested on an Ion Torrent PGM to identify 41 strains of clinical importance. The V1–V2 region identified 40 of 41 isolates to the species level. Three data analysis methods were tested, finding that the Ribosomal Database Project's SequenceMatch outperformed BLAST and the Ion Reporter Metagenomics analysis pipeline. Lastly, 16S rRNA gene sequencing mixtures of four species through a six log range of dilution showed species were identifiable even when present as 0·1% of the mixture.ConclusionsSequencing the V1–V2 16S rRNA gene region, made possible by the increased read length Ion Torrent PGM sequencer's 400 base pair chemistry, may be a better choice over other commonly used regions for identifying clinically important bacteria. In addition, the SequenceMatch algorithm, freely available from the Ribosomal Database Project, is a good choice for matching filtered reads to organisms. Lastly, 16S rRNA gene sequencing's sensitivity to the presence of a bacterial species at 0·1% of a mixture suggests it has sufficient sensitivity for samples in which important bacteria may be rare.Significance and Impact of the StudyWe have validated 16S rRNA gene sequencing on a benchtop sequencer including simple mixtures of organisms; however, our results highlight deficits for clinical application in place of current identification methods.

Highlights

When infection is suspected, timely and accurate identification of pathogens can improve patient outcomes and reduce the excessive use of broad-spectrum antibiotic therapy
When 400 base pair sequencing chemistry became available for the Ion Torrent Personal Genome Machine (PGM), utilizing longer 16S rRNA amplicons containing more than one hypervariable region became feasible
The taxonomic level reached by the V1–V2 hypervariable region contained greater than 95% of all reads analysed except for three of the 41 strains: Enterococcus casseliflavus, Shigella sonnei and Proteus vulgaris (Table S1)

Summary

Introduction

Timely and accurate identification of pathogens can improve patient outcomes and reduce the excessive use of broad-spectrum antibiotic therapy. 16S rRNA gene sequencing on a benchtop sequencer widespread use and utility, culture-based diagnosis of infection relies on outgrowth of one, or a few, dominant organisms over fastidious organisms present in the sample despite that slow, or nongrowing, organisms may be clinically important. This problem is exacerbated in chronic infections that are commonly multi-organismal in nature (e.g. diabetic foot ulcers), when antibiotics are present, and when the causative organism is fastidious. While 16S rRNA gene sequencing has been successfully used clinically, drawbacks beyond time and cost have been noted, including lack of universality and the inability to differentiate between true pathogenic organisms, opportunists, and commensals (e.g. Shigella species versus Escherichia coli, respectively), sequencing errors, and intra-species sequence heterogeneity (Ashelford et al 2005; Boudewijns et al 2006; Janda and Abbott 2007; Woo et al 2007; Kunin et al 2010)

Methods

Results

Conclusion