Scientists from around the world are reported to have drooled as they witnessed the unveiling of a new DNA sequencing technology at a conference last week.
At the Advances in Genome Biology and Technology (AGBT) event, UK startup Oxford Nanopore claimed that a lab using a scaled-up version of its GridION hardware could sequence an entire genome within 15 minutes in 2013. Most sequencing companies are still racing to produce results in a day.
The drooling may have been metaphorical, but the abilities of Oxford Nanopore’s technologies are, the company says, definitely real. Nevertheless, the outfit came under fire over the weekend, as competitors and sceptical geneticists lined up to tear it apart - especially for not publishing data produced by its inventions.
“With no data release, how do you know this is not cold fusion?” quipped Jonathan Rothberg, who started Ion Torrent, a division of lab equipment maker Life Technologies. That company’s stock fell by 4.3 percent fell as the news from Oxford broke.
Hard data required
It was no surprise that Rothberg, as a competitor in a fierce market, laid into Oxford Nanopore. But he wasn’t the only one. One scientist keen to believe in the hardware said he needs hard data before he’ll buy into Oxford Nanopore’s claims.
And he’s backed up by Fiona MacDonald, head of the molecular genetics section at the West Midlands Regional Genetics Laboratory, who remembers receiving a pitch from Life Technologies when it was touting its Ion Torrent system around labs. At that time, the company had used it to sequence only bacterial genomes. “We said we wouldn’t buy it before we’d seen some data,” MacDonald remembers. “It’s important to see the data and see what it looks like.”
But Oxford Nanopore is keeping its cards close to its chest. An insider told Elements that the strategy is to continue developing the units with a select group of customers, because the customers will help them to establish a market for what they bill as a truly disruptive technology.
“The exquisite science behind nanopore sensing has taken nearly two decades to reach this point; a truly disruptive single molecule analysis technique, designed alongside new electronics to be a universal sequencing system,” said Dr Gordon Sanghera, the firm’s chief executive, in a statement.
Method in the madness
Here’s how the technology works. While conventional DNA sequencing techniques include dying different sections of the DNA distinct colours so that machines can ‘read’ the DNA, nanopore sequencing relies on electronic sensing.
An electrical current is fed through very small holes (‘nanopores’) in a membrane and then the DNA sample is passed through this hole. Since each pearl of information along the string of a DNA molecule has a certain shape and size, it disrupts the electrical signal in a characteristic way, allowing the GridION machine to infer its type.
You can read more on the technology in this previous Elements article.
Erring on the side of caution
This whizz-pop system may get scientists drooling, but they are not throwing themselves at it yet. “The error rate they’re talking about at the moment is 4 percent, which is higher than the error rate of the current system [we use],” says Dr Katie Snape, a clinical research fellow in the division of genetics and epidemiology at the Institute of Cancer Research and The Royal Marsden Hospital.
In her work, Snape uses next-generation sequencing technology to identify genes that make their carriers more predisposed to certain diseases. “Obviously, while it’s promising, they need to try to make that a bit more competitive.”
On the worrisome error rate, Oxford Nanopore’s chief technology officer told bioinformatician and blogger Nick Loman that he knows what the problem is “and we can fix it”.
In any case, Oxford Nanopore may need to compete on other fronts - especially if GridION is to be of use in initiatives like Macdonald’s Cancer Research UK-funded project. “We’re focusing on very specific genes,” she says. “We don’t want to sequence everything and [the GridION system, which focuses on whole genomes] would probably give us more information than we need.”
MacDonald raises a key issue faced by all scientists when choosing which sequencing technology to purchase: is it right for my research? As Snape says, “It may very well be that [Oxford Nanopore] comes up and overtakes or becomes parallel [to existing technologies], but you want to see the end points of what you want to achieve your research aims.”
But research is not the only market Oxford Nanopore is gunning for. One of the more headline-happy messages in the company’s announcement was that it can squeeze the GridION sequencing system onto a USB flash drive - perfect for consumers.
“The USB stick is an absolute game-changer,” Sanghera told Bloomberg. “It’s plug-and-play, on-the-go DNA sequencing.”
One researcher told Elements that MinION “seems a little bit gimmicky” at the moment, especially because there is a huge gulf between obtaining genome data and being able to interpret it.
No doubt the brains in Oxford are working on that as part of their strategy for this fierce race. But the company is running a marathon, not a sprint, and it is already seen to have “won” at AGBT.
Image courtesy of Oxford Nanopore.
Certainly every technology must be evaluated in the context of a specific application; this is a message that is too often lost.
But, it's strange to me to worry about “We don’t want to sequence everything and [the GridION system, which focuses on whole genomes] would probably give us more information than we need.” There is nothing to prevent you from filtering the DNA upstream of any platform the way scientists do currently. So no matter the throughput, you could still use PCR, hybridization selection, cloning and whatnot to winnow down the input material.
But in the end, it will likely soon be more cost-efficient for most biological questions to simply collect data across the entire genome/metagenome and then computationally filter down to what you are interested in; the fully-loaded (labor, capital, disposables, failures) cost of the computational filtering will beat the fully-loaded cost of doing the upstream experimental filtering.
For MacDonald, the matter is simply that she can currently sequence only the genes she's interested in, which Oxford's hardwire won't let her do at the moment, so why should she switch? But I agree with you that ultimately it'll make more sense financially to have WGS techniques and then use computational filtering later on.
There's so many possibilities with these various technologies at the moment and no one idea is sticking - yet. But we'll probably reach some kind of stability soon - probably once the time and cost of WGS is next to nothing.
As soon as you say “Oxford's hardware won't let her do at the moment” you've lost me.
Like any other sequencing technology after the original (Maxam & Gilbert or Sanger), the Oxford instrument takes a mixed population of DNA in and spits out base calls organized by input molecules (perhaps in segments). You can still do all the laborious sample prep you want; it simply doesn't require so much of it. Nothing prevents you from taking samples prepared for Illumina or 454 or Ion Torrent and simply sticking them in the Oxford instrument. Not understanding this is a bad reason not to switch; understanding the accuracy requirements of your experiment is where you'll find good reasons to stand put.