This article was taken from the April 2015 issue of WIRED magazine. Be the first to read WIRED's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online.
In June 2014, Heartlands Hospital in Birmingham reported an outbreak of Salmonella infection. It affected 30 patients and staff in two wards and spread to long-term in-patients on two adjoining wards. Salmonella food poisoning is associated with eggs and undercooked poultry, and although outbreaks are common, they are rare in hospitals.
On June 12, the infection-control team sent 16 strains of the bacteria from patients' faeces to Nick Loman, an expert in infectious diseases at the University of Birmingham. It wanted results ready for a meeting the following morning. "The hospital wanted to understand quickly what was happening," Loman says. "But routine genome sequencing is quite slow. It usually takes weeks or even months to get information back."
However, Loman had a new type of DNA sequencer that was, in theory, suited to the task. Called MinION, it was made by a British company called Oxford Nanopore and was based on a new technology called nanopore sequencing. "It was announced two years ago as the first portable device that could do rapid DNA sequencing in real time," Loman says. "Everybody was excited. Then for a long period nothing happened. We started to worry. Some people were even saying it was never going to happen because the technology they used defied the laws of physics."
Loman had received the sequencer in May 2014, as part of an early-access programme for researchers. It was the size of a USB stick, weighing about 100 grams. Loman paid a refundable deposit of £650 per device and used three of them for his research into infectious diseases. He didn't sign any confidentiality agreements and is free to publish any data, so long as he doesn't reveal trade secrets. A month later, he published the DNA sequence of a strain of Pseudomonas aeruginosa, a common hospital-acquired infection. It was the first publicly released data obtained from the MinION. "It was just to test the device," says Loman. "I really wanted to test it in an actual outbreak and see how useful it would be for quickly sequencing bacteria."
Loman used MinION to detect Salmonella in some of the samples sent from Heartlands, obtaining results in less than 15 minutes. Two hours later, he concluded that the majority of the Salmonella strains were part of the same outbreak, except for a second smaller outbreak linked to children who had been to Egypt. When the genome sequence of the Salmonella was compared with other strains in the public-health databases, the results suggested a link with cases not only in London, Bedford and Northampton but also in Germany, France and Austria. The source was a German egg supplier. "The MinION just blew me away," Loman says. "The idea that you could do sequencing on a sort of USB stick that you can chuck around does stretch credulity."
But tracing food poisoning is just the start. Today, more than 1,000 researchers from 63 countries are working to apply the MinION to a wide variety of problems: David Deamer, a biochemist at Santa Clara University, California, is using it in the search for extraterrestrial life; Brook Milligan, a biologist at New Mexico State University, relies on it to monitor illegally traded wood and wildlife; Nick Loman is in talks with the British military to take it to Sierra Leone to help doctors diagnose Ebola quickly. "These are great applications, but the one that gets me going isn't being done yet," says Clive Brown, CTO of Oxford Nanopore. "We're finding that everything that happens to your body is reflected in your blood. Cancer cells can grow and die quickly, shedding cancerous DNA into the bloodstream. Where I want to go is self-monitoring -- detecting things before they appear."
Oxford Nanopore was founded in 2005 by Hagan Bayley, a chemist at the University of Oxford, Gordon Sanghera, an expert in biosensor technology, and Spike Willcocks, who had worked at IP Group, an intellectual-property investment firm from which Oxford Nanopore received its first funding. Bayley's research involved using nanopores -- proteins that assemble as a pore on cellular membranes, perforating it -- to detect molecules that go through its inside channel. The initial goal was to use nanopores as sensors for various molecules, but soon the team decided to focus on DNA sequencing. Sanghera had started out at MediSense, an Oxford-based company, where he had been pivotal in launching the first commercial blood-glucose sensors. "When I joined MediSense, glucose measurements used to require ten minutes," says Sanghera. "Along came a digital device where you put in a drop of blood and you have a result in 30 seconds." To Sanghera, the parallels between glucose sensors and what Oxford Nanopore was attempting for DNA sequencing were obvious. "Measurement is everything, and this was a direct readout of our genetic source code," says Sanghera.
Oxford Nanopore made its first device on April 4, 2006. It was fairly primitive: a battery-powered, calculator-sized box into which a plastic chip containing one nanopore (designed to detect and measure cyclodextrin molecules) was inserted. It had a button to apply voltage. The only way to turn it off was to dismantle it. The next morning, Willcocks, Sanghera and Bayley went to London to meet investors. They had never taken a nanopore-based instrument out of the lab before. "Our engineer had finished making this at one in the morning," Willcocks says. "When we told Bayley we were doing a live demo, he said we were bonkers."
At the first meeting, as Sanghera was making his pitch, Willcocks attempted to operate the device. "Suddenly, I got a stable current, and then it started measuring the molecule," Willcocks recalls. He showed it to Bayley, who immediately interrupted Sanghera: "Guys, you have to look at this."
Willcocks had to keep the device on while they travelled around London meeting investors. "We actually amazed ourselves," he admits now, "that we could keep this device going outside the lab."
Geneticist and computer biologist Clive Brown joined Oxford Nanopore in June 2008 from the Wellcome Trust Sanger Institute, the biggest genomic laboratory in the UK. Brown had been a director of computer biology and IT at Solexa, a Cambridge-based company that in 2006 had made a popular DNA sequencer called Genome Analyzer, which worked by splitting apart and reassembling DNA in tiny pieces on a massive scale, then deciphering the results. In 2007, Solexa was acquired by Illumina, a San Diego-based genomics company, for $600 million, allowing the US company to enter the DNA-sequencing market. Since then, Illumina has dominated the market: its market share currently sits at around 80 per cent. "At Solexa, our VCs didn't trust us and sold it off as soon as they could," Brown says. "It crushed the people leading that project, me included. You have scientists who come in with a dream and build something, and then you get the rug pulled by some silver-spoon investor and they all go and become teachers or estate agents. It's hard as a British and European company, because very few people believe that you can do technology here."
That summer, Oxford Nanopore managed for the first time to accurately and continuously identify DNA bases, using a nanopore-sequencing technique called Base, an acronym for Bayley Sequencing. Base uses a protein nanopore coupled with an enzyme, whose role is to separate individual bases from a strand of DNA and sequentially introduce the bases into the nanopore. As the individual bases travel through the nanopore, they disrupt an electrical current flowing through it. Those disturbance signals are recorded electronically and interpreted to identify the DNA base. "At first, the idea of using a nanopore to measure DNA was a bit like Jules Verne writing about flying a rocket to the Moon," Brown says. "Very well, but how do you do that? People had been talking about nanopore sequencing at conferences for years and they were booed and laughed off. Many were saying that this idea would never work, that DNA will never go through that hole, that we would never decode the signal. The moment I saw what the signal-to-noise ratio was, I knew we could do it."
On January 12, 2009, Oxford Nanopore signed a commercial agreement with Illumina, by which Illumina would exclusively commercialise Oxford Nanopore's Base technology products and invest an additional $18 million. Illumina's CEO, Jay Flatley, became a board observer. "Oxford Nanopore's technology holds tremendous promise to achieve the sub-$1,000 human genome," said Flatley in the press release. "Making electrical measurements of unmodified DNA removed the need for complex sample preparation and the high-performance optics found in today's sequencing systems. We look forward to a long and productive partnership with Oxford Nanopore."
Despite the technology's initial promise, the next three years were plagued with difficulties. "We couldn't get the biology and electronics to work together," Willcocks says. "We were staring down the barrel of a shotgun." "It was soul-destroying. We had spent years trying to make that thing work," says Brown. "We performed miracles of protein engineering and experimental design, and we couldn't get a peep out of the data. Illumina started to get very tetchy and impatient. Those years were the worst of my life."
On September 27, 2010, a paper by Mark Akeson, a geneticist at the University of California, Santa Cruz, and a member of Oxford Nanopore's technology advisory board, was published in the journal Nature Nanotechnology. It demonstrated an innovative method of nanopore sequencing called strand sequencing, which involved a new enzyme that could move long strands of DNA into the nanopore, rather than individual bases.
Using the new technique, but a different chemistry, Oxford Nanopore sequenced a viral genome and a human genome in a matter of months. "We had a meeting and everyone shouted at each other," says Brown. "At the end, we decided to abandon Base and go with strand sequencing instead. That's when we became a proper company."
February 17, 2012, Clive Brown gave a talk at the Advances in Genome Biology and Technology (AGBT) meeting in Marco Island, Florida, an annual event where major breakthroughs in DNA sequencing are usually announced. The talk was titled "Single-molecule strand sequencing using protein nanopores and scalable electronic devices". Brown showed the first whole genome sequenced using nanopore strand sequencing. He also announced the GridION, a DNA sequencer with about 8,000 nanopores that could complete a whole human genome in around 15 minutes. Right at the end of his talk, Brown introduced the MinION, the first ever portable real-time DNA sequencer. "I mentioned it almost like a footnote," says Brown (pictured, opposite page). "There was a murmur in the audience, this rumbling noise. The reaction afterwards was off the scale. I had to hide in my hotel room."
Chad Nusbaum, codirector of genome sequencing and analysis at the Broad Institute of Harvard and MIT, and one of the conference's organisers, blogged: "People have been watching nanopore technology since the 90s and have accumulated a high level of scepticism over the years, so it is a game changer if someone can go to the podium and say, 'Yes, this works.'"
A few months later, however, a significant design flaw was found in the device's microchip. It took Oxford Nanopore nearly two years to fix the problem. "For those two years our name has been in the mud," says Brown. "People have accused us of peddling 'vapourware' and cold fusion and saying we're full of shit."
The device had been Brown's idea from the outset, so he was naturally protective. "Before, we did what everybody else does: you build a big ugly box, put some stupid name on it, like BigSeq 3000, and then charge £200,000 for it," he says. "Inside, it's all off-the-shelf components and lots of space. It's what the industry has been doing for 30 years." Brown conceived a much simpler and smaller instrument. He says that, in early 2011, when he first presented his idea at a board meeting, Flatley said he couldn't see the use for it. "He was dismissive and chauvinistic," says Brown. Oxford Nanopore decided to make the MinION anyway. "People believed that Illumina pretty much owned us," says Brown. "So we felt pressured to put the record straight and to make a public statement that we wanted to commercialise it ourselves."
The MinION is based on strand sequencing, a technology for which, unlike Base, Illumina didn't own the commercial rights. On November 15, 2013, Oxford Nanopore and Illumina severed their ties, announcing the divestiture of Illumina's shares for $56.4 million. Ten days later, Oxford Nanopore announced the Minion Access Programme, designed to give access to MinIONs for a refundable deposit of £650. About 3,000 researchers applied and, by June 2014, about 500 were invited to join. "We found that when we started to talk to customers, they were so excited about it. Every single person would have ten interesting things to do with a decentralised device like that. We decided to allow people to try it out and come up with interesting applications," says Sanghera. "The conventional path to release technology is via the endorsement of opinion leaders," says Brown. "We bypassed that group and put the product straight into the hands of the mob. In this case, the mob happens to be a group of very well-intentioned people who want to cure diseases."
Oxford Nanopore occupies two four-storey buildings at the Science Park on the outskirts of Oxford, from where it manufactures and distributes its products. "It's all we need for world domination in DNA sequencing," says Sanghera one morning last November. "Because we're only using single molecules, the concentrations we need are very small to meet the global production forever. Even if we sequenced every baby in the world at birth, we could do it from these buildings." So far, they have raised £180 million from investors in the UK, US and Europe.
In a meeting room, Sanghera and Willcocks set up an experiment with the MinION to decode the DNA of a phage, a virus that infects bacteria. Sanghera bangs the device against the office table and then turns it upside down. "The device is very robust," he says. This is the only sequencing machine that can work upside down."
A technician pipettes the liquid sample containing the phage's DNA on to the MinION, which is connected to a laptop via a USB. Preparing the sample still requires some degree of chemical preparation, but the company is planning on launching an upgrade that will enable users to sequence DNA directly from a biological sample, such as a drop of blood.
On the screen is a graph showing the electrical current in one nanopore, which Willcocks calls a "squiggle". Each MinION has about 2,000 of these. Each DNA base is about half a nanometre, corresponding to an atom or two, and each nanopore reads about 30 bases per second, as a single strand of DNA goes through it. The initial current flowing through the pore is disturbed by the DNA and that disturbance is used to identify the individual bases. Each piece of DNA is timestamped and geotagged and sent to the cloud for data processing. "You look at that and it looks like noise," says Willcocks. "But we have enough variations in the signal to determine the sequence. We use algorithms similar to voice-recognition to decode it. Each datapoint correlates to DNA moving one position, one base. We're sequencing in real time." "Being able to do things in real time changes the dynamic of the decisions that you make," says Brown. "If you're out in the field, it has to be real time. If it takes two days to get an answer, it's useless."
Sanghera suggests a useful test: "If you've bought a beefburger and are worried about its contents, just take the MinION, stick it in and it will tell you if there's horse in there." "We did just that recently," Willcocks says. "What was it we found in that beefburger? There was some human DNA in there, right?"
Willcocks shows a plot that displays the length of the DNA strands. The average is about 7,000 bases; some are as long as 18,000. "Standard sequencing methods need to break DNA strands into portions of a few hundred bases," he says. "DNA is made of three billion bases. When you need to put the jigsaw back together, it's very hard to do. We can measure whatever length of DNA you load on to the chip. It makes it easier putting the jigsaw back together, and also helps you discover new biology that otherwise you wouldn't."
With the MinION, Oxford Nanopore sees itself targeting an unmet need for access to DNA sequencing: as the first portable DNA sequencer, it was designed for life-science researchers such as Nick Loman, but is equally as useful for farmers who want to sequence their crops as it is for doctors in West Africa who want to diagnose Ebola as quickly as possible. "Only about 4,000 labs in the world have an Illumina machine," Brown says. "Now, how many labs are there that would like to do DNA sequencing? A hundred thousand? What about universities? Schools? What about the little guy? Our idea is that you don't need an enormous grant, you don't need a massive reputation, you don't need big teams of technicians. It's very anti-establishment. We haven't played the game."
João Medeiros is WIRED's science editor. He wrote about Stephen Hawking in 01.15
This article was originally published by WIRED UK
