Mobile DNA Sequencing in the Ebola Epidemic
By Aaron Krol
February 3, 2016 | Just over a year ago, Bio-IT World spoke to microbiologist Nick Loman about the recently released MinION DNA sequencer. The three-inch-long device, made by Oxford Nanopore Technologies of the UK, can read DNA in real time on a laptop, and Loman’s lab at the University of Birmingham was one of the first to receive one. Like many other early adopters we spoke to at the time, Loman was itching to try the MinION in real-world clinical contexts, following the genetic traces of an infection as it develops.
“You can imagine throwing one in a suitcase and taking it down to Sierra Leone to look at the Ebola outbreak,” he told us in that interview, which was conducted within weeks of the worst phase of the epidemic.
As a matter of fact, his team would go to Guinea instead.
Today, an article describing the months-long effort Loman helped organize, to sequence virus samples in the middle of the largest Ebola outbreak in history, has finally been published in Nature. The article features contributions from dozens of scientists and healthcare workers, hailing from organizations in twelve different countries in Africa, Europe, and North America. To those following the field, the publication will not come as a surprise. Loman, with colleagues including Joshua Quick and Lauren Cowley, have gladly discussed their team’s work in the press and at conferences. But the description of the project published today expands on what was already known, makes clear the capabilities and limitations of the MinION in the most extreme outbreak scenarios, and offers a detailed template for future efforts.
Quick was the first member of the Birmingham lab to fly to Guinea, departing in April 2015 with an entire mobile lab’s worth of equipment packed in a single luggage carrier. In the capital Conakry, he met up with local and international healthcare personnel at the European Mobile Laboratory in Donka Hospital, and set up his equipment on a pair of small tables they had set aside for the study. One table held four laptops with three MinIONs; the other had a heat block and a small thermocycler for performing PCR on viral RNA isolated from blood and urine samples.
Just two days later, Quick and his team on the ground were sequencing Ebolavirus. They used leftover samples from local laboratories, whose managers also provided information on when and where each patient had contracted the disease. The work, which would continue until October and eventually move to a treatment unit in nearby Coyah prefecture, ultimately sequenced virus from 142 patients, generating enough DNA data to construct a phylogenetic tree retracing different strains of Ebolavirus as they spread through the country.
The authors believe their work on Ebola represents the first attempt to use a MinION to study an ongoing viral outbreak. The technology’s trial by fire delivered several useful findings. For instance, DNA signatures from the various samples the researchers collected indicated that the Ebolaviruses circulating in Guinea could trace their origins to two main lineages―one that arose within Guinea early in the epidemic, and a second wave that crossed the border from Sierra Leone in late 2014. By making their data public, and comparing it with incoming data from a second sequencing project in Sierra Leone, the authors were also able to conclude that borders between the two countries remained porous during the outbreak, with virus regularly transmitted back and forth.
A Marvel, But Not a Panacea
Yet the Nature paper also makes clear that the MinION, as it exists now, is neither a frontline diagnostic nor a catchall tool for surveillance. Although the dream of a near-universal, on-the-go diagnostic for infections is closer than ever to reality, the team found that the MinION is not ready to sequence DNA agnostically and deliver complete pathogen genomes, at least in the severe constraints of an emergency clinic.
One problem is that the MinION’s analysis tends to break down when confronted with long sequences of the same DNA letter, or homopolymers, which occur in several parts of the Ebolavirus genome. That means genomes built from scratch will necessarily be incomplete and fragmented, something that also applies to many other species of pathogens. Another issue is that the MinION, compared to other sequencers, can only read a small amount of DNA per run. It’s more than enough to sequence the tiny genome of a virus, but not if the machine is also being fed the large amount of human DNA present in any clinical sample.
That meant the team’s workflow had to be carefully customized to the Ebola outbreak. The authors designed a series of PCR primer pairs to amplify Ebola DNA, eventually discovering that a set of 11 amplicons let them span the great majority of the virus’ genome. This let them efficiently pull the genetic material of the virus out of patient samples for sequencing―but also meant their approach was not general-purpose.
Other MinION users have done interesting work trying to identify pathogens from small snippets of DNA, but for now, work on actual patient samples is more likely to follow the example of the Ebola sequencing project.
The good news is that practical constraints for sequencing in an outbreak zone are crumbling fast. Working in overwhelmed, under-resourced environments, the team faced serious setbacks: the Internet connection on the ground would often stall or shut down, making it hard to send data back to Birmingham for analysis, and electricity was so unreliable that some of the equipment had to be run on uninterruptible power supply units brought in for the project. Yet, before the MinION, it would have been all but unimaginable to create a complete sequencing lab on the go, with all these constraints, and still be collecting DNA data within days of arrival.
Researchers may also be within sight of a solution to the connectivity troubles encountered in Guinea. Already, some genetic analysis can be done locally; Oxford Nanopore provided this project with a special version of its MinKNOW software, which performs initial interpretation of the MinION’s electrical readouts, to run on a laptop. More sophisticated analysis―comparing incoming Ebolavirus strains against a whole genome collected early in the outbreak, in search of markers that could track the virus’ evolutionary divergence―would have been much harder to do onsite. As the authors write, “There is a pressing need for a fully offline version of the analysis presented here.”
The work this large collaboration did in Guinea is a big step for sequencing in an outbreak, showing that large-scale genetics is ready to play a role in surveillance even in the most trying circumstances. Yet its success is not just a matter of the technological leaps that produced the MinION. The researchers’ commitment to open science was just as crucial. By releasing their data publicly and regularly, they were able to both share results as they worked, and merge their data with DNA sequence collected across borders and institutional boundaries, resulting in a fuller picture of the outbreak’s lines of transmission.
The European Mobile Lab was just one of many contributors to sites like ebola.nextstrain.org and virological.org, which sped the process of discovery at a time when the quick sharing of findings could make a real difference for public health. The entire analysis workflow the team used to find genetic markers in their Ebolavirus samples is also available at github.com/nickloman/ebov, for future programmers to scrutinize, reproduce, and modify.
Meanwhile, a very similar study, this one focusing on the MinION as a diagnostic tool for Ebola in Liberia, appears in this month’s issue of Emerging Infectious Diseases, a journal of the US Centers for Disease Control and Prevention. The swift adoption of this tool in the Ebola outbreak, by multiple groups working independently, suggests it may not be long before field sequencing is embraced as a standard part of the international response to health emergencies.
UPDATE 2/3/16: Nick Loman has written a fascinating post on the logistical efforts to get this study off the ground, which you can read on his lab blog.