|Your choice of sequencing approach matters. Think|
about your goals and the methodological caveats
before starting your experiments.
The field of microbiome research has been hugely popular in the past few years. It has forced us to rethink our approaches to various medical practices, and has captured the imaginations of both amateur and professional scientists. With this popularity has come an influx of scientists trying to incorporate the microbiome into their own research. It is of course great that people want to get into the field, but unfortunately it is deceptively difficult for newcomers who are not always aware of how best to get started. This has led to the execution of poorly designed studies that could have been improved by more methodological resources in the literature. To this end, my colleague (and lab mate) led a research project to evaluate the differences between sequencing methods of the skin microbiome, a consideration that is often overlooked by newcomers to the field. This week I want to briefly hit the highlights of the paper and suggest that you read it if you are interested in starting any skin microbiome work.
The study was led by Jacquelyn Meisel in Elizabeth Grice's laboratory, and was published in the Journal of Investigative Dermatology (the premier dermatology research journal). In their study, Meisel et al evaluated the effects of three different sequencing methods for studying the skin microbiome.
- Whole metagenome shotgun (WMS) sequencing, which means the entire genomes (or genome fragments called contigs) of the skin bacteria were sequenced instead of a specific region (e.g. 16S rRNA). This method is costly and more difficult to analyze, but can provide answers to many questions regarding the genomic structure of the communities that cannot be answered using techniques involving marker genes.
- 16S rRNA V4 region gene sequencing, which means the fourth variable region (V4) of all bacteria within the bacterial community is sequenced and used to provide taxonomic/phylogenetic information. Variable regions are used because the high throughput sequencing technologies cannot span the entire length of the gene, and the variable regions allow for the greatest differentiation between different bacteria (if we used a conserved region, they would all look the same). This method is great because it is cheaper, provides strong taxonomic/phylogenetic information about the community, and is sufficient to answer many research questions. It does not provide sequences for the entire genomes however.
- 16S rRNA V1-3 region gene sequencing, which is the same approach as V4, although it is covering variable regions 1-3 instead of four. Different variable regions provide different resolution between members of the community because they are differentially variable between groups of bacteria. This region in particular is longer than V4, which means it can provide more information at the expense of being more difficult to sequence.
|Illustration of the variable regions within the 16S rRNA|
gene. The valleys are regions of low conservation, and
are labeled as variable regions 1-9. <Source>
So what did the group find? The highlight was that the V4 region poorly characterized the skin community, while the V1-3 and metagenomic approaches were much more accurate (accuracy was determined by sequencing a known community and comparing the results to the known composition). The most striking limitation to sequencing the V4 region was its inability to capture Propionibacteria.
The reason for using metagenomic approaches over 16S sequencing is thought to be that the metagenomic data allows for an understanding of the functional potential of the community. Meisel et al found that the functional predictions made using 16S data was similar to that found in the metagenome samples, meaning you are getting comparable results but paying considerably more for the metagenomic data.
The group also evaluated the effects of these methods on the resulting diversity calculated for the communities. They found that the resulting diversity was in fact impacted by the sequencing approach, highlighting a danger in comparing results from different studies that used different sequencing methods.
Now I know I have an obvious bias since I was a part of this research, but Jackie (Jacquelyn) led an excellent study that provides an important resource to the field. If you are curious about the importance of sequencing methods, or if you yourself want to incorporate this type of study into your research, I suggest checking this paper out. It can help you to interpret other skin microbiome studies, and could prevent you from making costly mistakes in your own research.
As always, I would love to hear your questions, comments, and concerns in the comment section below, or through email/Twitter. You can find my information to the right.