NEW YORK (GenomeWeb) – Oxford Nanopore Technology is working on a droplet-based nanopore sequencing technology that draws on elements of its VolTrax automated sample preparation device and its existing nanopore technology.
Clive Brown, the company's chief technology officer, provided an overview of the new approach during a presentation at the firm's user meeting last week. While the meeting was closed to media, a video of the talk, which included an update on Oxford Nanopore's business and product development, appeared on the firm's website this week.
Oxford Nanopore has shipped between 6,000 and 7,000 of its MinIon devices so far, which Brown said remains "by far the most mature, most widely used" of the firm's sequencing platforms. A new version of the MinIon flow cell called D Chip with an improved electronic chip, or ASIC, is in development that will enable longer run times of up to 100 hours, resulting in yields of up to 30 gigabases internally, he said.
For its medium-throughput GridIon sequencing platform, the company now has more than 100 customers in 24 countries. Of those, 13 have been certified by the firm as service providers, and several more are in the process of becoming certified, he said. Current GridIon service providers are based in the US, China, Australia, Japan, France, the UK, and the Netherlands, and include both companies and academic institutions.
The PromethIon, the company's high-throughput sequencing platform, is now fully commercialized, and Oxford Nanopore has shipped 40 systems so far, with another 20 to be delivered by the end of June. The platforms currently being shipped are beta systems, which can run up to 24 flow cells, and customers "buy into an upgrade path," Brown said, with the next version allowing users to run 48 flow cells in parallel.
He acknowledged that the delayed roll-out of the PromethIon "was a bit of an embarrassment" for the company, but added that "we now got our act together" and the platform has high data yields in the hands of customers.
Early-access PromethIon customers talked about their initial experience with the platform earlier this year, reporting yields of up to 50 to 100 gigabases per flow cell. Brown said that many customers now achieve more than 50 gigabases, "quite a few" obtain 60 to 70 gigabase, and "a handful" get more than 100 gigabases per flow cell. Internally, the company has generated more than 160 gigabases on a PromethIon flow cell and has a road map to get to 250 gigabases and later 500 gigabases per flow cell, he said. Crop breeding company KeyGene of the Netherlands is the first PromethIon customer that has been certified as a service provider.
Oxford Nanopore is working on a single-channel flow cell for the PromethIon that will be easier to load and run, which it expects to release broadly in June. That flow cell will have a higher pore yield and is less prone to air bubbles, Brown said. The current multi-channel flow cells will be redesigned, based on feedback from early customers, and will come back later this year.
The company will also make all kits for the MinIon — in particular the direct RNA sequencing kit — available for the PromethIon in June.
Longer term, Oxford Nanopore is working on a new, automated droplet-based sequencing technology that would enable sequencing of small amounts of DNA, including single cells. It will combine elements of the VolTrax sample prep platform and the MinIon into a single, yet-unnamed device. "That's our next likely major product launch," Brown said.
Instead of using nanopores inserted in a membrane on top of a well on a chip, the new technology will measure current through a nanopore that sits at the interface between two microfluidic droplets. To do that, two droplets, one containing nanopores in solution, are brought into close proximity on the VolTrax device and next to recording electrodes. When the droplets come into contact, a bilayer forms at their interface. Then a voltage is applied, causing a pore to insert into the shared membrane. DNA moving through the pore from one droplet to the other can then be sequenced by measuring the current through the pore, similar to how it is done on the MinIon.
This setup would allow for long sequencing runs, and DNA could be further manipulated — for example, PCR-amplified — after sequencing. Overall, it would enable "a very large number of very complicated workflows," Brown said.
The throughput of the new device would be low — the target is 128 channels, which Brown said is "ambitious" — but it could potentially allow single cells to be inserted, lysed, and sequenced.
Brown said the firm has already built a breadboard array with a DNA extraction zone, a library preparation zone, and a sequencing zone. He also showed data from a proof-of-concept experiment where single HeLa cells were moved on the array and lysed with a voltage. "If you have a tiny amount of material, this is a device that can do that," he said.
While he did not provide a timeline for commercialization, he said that "we are not talking years."
Brown also mentioned a number of improvements to its existing sequencing technology, including its software and sample prep devices, among them one that came from a customer.
Researchers led by Matt Loose's group at the University of Nottingham discovered that Oxford Nanopore's own MinKnow software incorrectly split long reads into shorter reads. In a BioRxiv preprint posted earlier this month, they described how they were able to correct for that, which led them to discover a single contiguous read of 2.3 megabases that was previously incorrectly reported as 11 consecutive reads.
Brown speculated that this could open the door to even longer reads, up to whole-chromosome length, provided the DNA is not fragmented and doesn't get stuck in the pore. To address the latter problem, the company included a new feature called "progressive unblock" in its recent update of the MinKnow software to version 2. Previously, unblocking a pore by reversing the potential tended to destroy the channels, Brown explained, but the new way of doing it is more gentle, resulting in larger overall data yields.
Flongle soon, SmidgIon on the backburner
The flow cell adapter, called Flongle, that Oxford Nanopore has been working on for a while is going to be available to early-access user this quarter and is slated for full commercialization in the third quarter, Brown said.
The device contains the expensive electronic parts that are currently part of the MinIon and GridIon flow cells, so Flongle flow cells will be cheaper and one-time-use. However, they will have only 128 channels instead of the 512 of the current MinIon flow cells. Initially, a Flongle flow cell is expected to generate 1 gigabase of data per 16-hour run, which might increase to 3 gigabases over time with optimization. This throughput will be particularly useful for gene panels, as well as for bacterial and viral genomes, he said. Pricing per flow cell, which will be sold in bundles, will be on the order of $100.
Commercialization of the SmidgIon, a sequencing device with even smaller flow cells that is powered by a smartphone, has been put on the backburner because sample preparation is not easy enough yet, Brown said, but might come back once large-scale equipment is no longer needed to get samples ready for sequencing.
To that end, the company has been working on Zumbador, a one-tube DNA extraction device it has now renamed Ubik. After applying a saliva sample or swab to the tube, the cells get lysed and the DNA attaches to a solid-phase support, which sinks to the bottom under gravity and drops into the sequencing flow cell, where the DNA is further prepared for sequencing. The Ubik prep takes about 10 minutes but requires more development work to get better yields, Brown said. In an on-stage demonstration during his talk, a company researcher used the Ubik to extract DNA from Brown's saliva and sequence it on the MinIon.
The firm has also been working on version 2 of the VolTrax library prep device, after rolling out version 1, which Brown called "a prototype", to early-access users. The new version, which will start shipping in August and be able to process 10 samples at a time, will have 15 inlets, can do PCR, will have improved magnetic control, and offer fluorescence detection for DNA and RNA quality control. Pricing for a starter pack, which includes the device, various kits, and MinIon flow cells, will be $8,000.
Another device, called MinIon Mk1C, which combines a MinIon or Flongle device with the MinIT compute module that the company developed to replace the required laptop for the MinIon, is targeted to become available by the end of this year.
For targeted sequencing, the firm has been developing a method that uses deactivated Cas9 enzyme and guide RNA to pull out regions of interest and transport them close to the pore. The plan is to make this available as a kit "later in the year," Brown said.
Direct RNA sequencing continues to be popular, and the firm will soon release a new version that will sequence RNA at 110 bases per second, up from the current 70 bases per second. It is also looking into new motor proteins to increase speed even further, he said.
Oxford Nanopore has also continued to work on improving its sequencing accuracy. One way to do so is through better software, and training the software on a wider range of DNA contexts, Brown said. As previously mentioned, homopolymers remain the largest source of errors. Another idea is to use two or more nanopore types, each with different error modes that are not correlated. The company has explored this concept by testing two pores, R8 (or lysenin) and R10, which are now "under very active development as potential products," Brown said. Even a mutant of the existing R9.4 pore was able to call some homopolymers correctly that the original pore was not, he added. Finally, accuracy could be improved by converting a DNA sample into a synthetic molecule, for example one that has Us instead of Ts, and sequencing both molecules.
On the software side, the company plans to enable the "read until" feature in its MinKnow software in the second half of this year, which allows users to do basecalling in real time and stop an experiment when sufficient data of a certain type has been collected.
Finally, Oxford Nanopore's Metrichor spinout has been developing a bioinformatics platform called EPI2ME that offers analysis workflows for researchers who do not have bioinformatics support. The platform can run either in the cloud or locally and is scheduled to be released on the MinIT in the third quarter. Oxford will charge for running a workflow with so-called "Metricoins" that customers can purchase.