In a recent paper entitled “Performance Comparison of Illumina® and Ion Torrent™ Next-Generation Sequencing Platforms for 16S rRNA-based Bacterial Community Profiling”, Stephen Salipante and colleagues at the University of Washington (USA) illustrate several of the challenges and limitations of any approach to 16S rRNA sequencing.
Amplifying the V1-V2 regions with a 320-350 base-pair amplicon, and sequencing on two comparable benchtop sequencers, the Ion PGM™ System produced 400bp single-end reads while the alternate platform produced bidirectional reads with an overlap of 150-180 base-pairs.
In a defined mock mixture of 20 bacterial species in equimolar concentrations and 18 human-derived microbiological samples, the results were ‘generally in good agreement’ (from the abstract).
A slightly higher overall error is documented on the PGM system, yet the authors note that ‘the absolute difference in error rates between the two platforms is not great’ (line 391, page 18 of the preprint). Furthermore the differences between the platforms did not change the final results of the species identification. Over the next few months we are introducing a new Hi-Q™ Enzyme to enable fewer indel errors during sequencing.
For all of this work, the authors performed a very careful analysis of full-length sequencing performance for 16S rRNA-based bacterial community profiling. In order to be called “present” in their sample, they argue that full length reads are stronger evidence for a call, and truncated reads (for unexplained reasons) on a small number of species represent a weakness in a platform’s ability to accurately identify the composition of a bacterial population. While more data is always better (and full length reads better than partial ones), the authors did not address some of the ‘larger-picture’ challenges that this study of platforms and sequencing V1 and V2 region amplicons would suggest.
Here are a few ‘larger picture’ questions that need to be addressed:
- Which 16S regions should be chosen in the first place?
- How will PCR inhibitors from environmental samples be suppressed?
- Is a commercial supplier for a 16S rRNA kit important? Alternatively, are resources available for optimizing individual assays and reagents?
- Is a pre-built, easy-to-use informatics pipeline important, along with the ability to generate visualizations? Does the bioinformatics staff have existing expertise or the inclination to learn tools such as QIIME?
- Are existing bioinformatics staff equipped to delve into mountains of uncurated datasources and screen through a multitude of uncharacterized matches?
For the last two questions, we’ve written here before about the 16S analysis capabilities of cloud-based Ion Reporter™ Software which you can try today with your own custom 16S primer sequences. The Ion Reporter software does not require full-length reads, which helps maximize the sensitivity of species detection, and accesses Applied Biosystems® MicroSeq® ID database. This is a proprietary, high-quality database with >15,000 manually curated organisms. In addition, the Greengenes database with >400,000 curated records can also be accessed via Ion Reporter™ software, and both databases can be used in sequential manner as an option.
To address the first three questions, we produced an Ion 16S™ Metagenomics Kit (ordering information is found here). It uses two primer pools covering seven hypervariable regions (V2, V3, V4, V6, V7, V8, and V9), contains a special PCR master-mix specifically to suppress PCR inhibitors from environmental samples, and is packaged ready-to-use. In addition, the AmpliTaq Gold® DNA Polymerase Ultra Pure is included for improved PCR performance.
“Performance comparison of Illumina and Ion Torrent Next-Generation Sequencing Platforms for 16S rRNA-based Bacterial Community Profiling”, Applied and Environmental Microbiology, Salipante, SJ, Kawashima T, et al. November 2014 Volume 80 Issue 21. doi:10.1128/AEM.02206-14