It’s been a really long time since I wrote anything here. Interesting mix of good and bad stuff in the meantime. Anyway…
We recently uploaded a bioRxiv preprint based on a small project of ours:
While I have a few minutes between flights in Copenhagen as I’m headed out to the field again, I thought I’d share some thoughts about it.
Extreme metagenomics? Really? In the sense we tried shotgun metagenomics to study an ecosystem people would consider “extreme”, maybe it is. But as this was a preprint, I felt I could indulge a tongue in cheek moment by alluding to this as a science equivalent of Extreme Ironing: Doing a normal task in a new and unusual environment.
So is there a value to sequencing metagenomes in the field? Hopefully. In the preprint we set out the trifecta of problems Arctic microbiologists face in my opinion: Firstly, the Arctic is warming fast, and that has unpredictable impacts on its ecosystems, fuelling microbial feedbacks and disruption to field plans. Secondly, to study these environments there are often challenges around sample integrity and timeliness incurred by transfer and process of samples back to a lab. This means that the insights gained are retrospective. Thirdly, while we can make many measurements of microbial community activities and abundance in field labs, being able to validate the capture of certain taxonomic groups on site could be very useful. For example: imagine wishing to test a hypothesis that genus X does Y using a field experiment – it would be useful to know that your source community contains genus X before starting the experiment, rather than finding out much later that your experiment measured genus Z instead because genus X was on holiday. Similarly, imagine drilling into a subglacial lake and wishing to confirm that your sample(s) are clean/distinct from drilling fluid or surficial contaminants before wrapping up your drilling season. Both applications are helped by on-site sequencing.
Why metagenomics? One could use 16S rRNA gene sequencing, but this necessitates an additional PCR step of course. This introduces primer biases, and also slows the process. In field conditions rather than a pre-PCR lab it also makes your analyses more vulnerable to contaminants. We preferred to do a quick 10-15 minute transposase based library prep from ca. 200 ng community genomic DNA. Conceptually I’d rather think of the output generated as a non-16S based 16S like taxonomic profile (!) than the kind of depth or coverage one would wish to have to reassemble genomes or delve into potential functionality with great detail.
Did it work? On balance, yes. There are things we’d do differently, principally around DNA extraction and bioinformatics, but we were able to arrive at community profiles comparable to published datasets from glacier community.
What next: I’d like to start some tests with mock communities and optimize the process further. Another area is considering how off-grid we can take this. On our way home from Svalbard I had five hours sitting in this same transfer centre and I decided to hack my way through the extraction, prep and sequencing process. I think we can do the same kind of experiment with a sub-five kilogram package without access to mains electricity or the internet.
Will you publish this in a peer reviewed journal? Hopefully. I’m new to both preprints and nanopore so I have no feel for where & how, or the worth of our observations. Our intention in uploading to bioRxiv now was to illustrate the possibility of going from extraction & sequencing in the field to sharing insights on a faster timescale than before. Given the trends for accelerated warming in the Arctic, speed may be of the essence for future work to be relevant.
I’ll update this as I get time & wifi over the next week.