NEW YORK (GenomeWeb) − Researchers from the University of Birmingham in the UK last week publicly released data they generated with Oxford Nanopore Technologies' MinIon nanopore sequencer, the first group to do so since the company started its early access program this spring.
The data, a single 8.5-kilobase read of comparatively low accuracy from the Pseudomonas aeruginosa genome, demonstrates that the technology works in the hands of users and provides some indication of what the platform can currently do.
The results add to those presented earlier this year by researchers at the Broad Institute, though those data were generated by Oxford Nanopore rather than Broad researchers.
Some predict the MinIon, which is the size of a pencil case and appears to generate long, noisy reads at low overall cost, will soon present formidable competition to Pacific Biosciences' platform and will target some of the same applications.
"It's surpassed all my expectations," said Nick Loman, a research fellow in the Institute of Microbiology and Infection at the University of Birmingham, who posted the first read. "I am amazed that we are getting so much useful data within a week of getting started."
Other participants in the MinIon Access Program are about to complete the mandatory burn-in phase – test runs with phage lambda DNA – while several have received the MinIon hardware and software but are still waiting for reagents.
Over the coming weeks, more users are expected to share results from their first experiments. While they are not permitted to reveal their lambda test results, they can talk about runs with their own samples.
First user experience
The Birmingham team received its MinIon device a few months ago and obtained its first flow cell and library preparation reagents in early June.
After seeing "very long alignable reads" with lambda DNA, the researchers told Oxford Nanopore they were ready to embark on their own experiments. Rather than having to meet certain company-determined specs during the test phase, users self-certify that they have completed burn-in, he explained.
The read Loman's team published last week came from their first run with a sample of their own, DNA from P. aeruginosa strain 910, originally derived from hospital water. In their second run, they have sequenced a large bacterial genome to greater than 40x coverage, he said.
The reason they chose to post the P. aeruginosa read is that it enabled them to identify the bacterium's serotype, O6, indicating that the data could be useful for diagnostic microbiology applications.
Analyses of the read by two bioinformatics researchers, who used different alignment tools and posted their results here and here, showed that the read is about 68 percent identical to the P. aeruginosa genome and has many errors, particularly gaps.
According to Loman, it is too early to say how representative that read is of the overall data quality. He pointed out that it is a unidirectional read, and that bidirectional reads, which his team is working on obtaining, should have fewer errors.
The library prep process introduces a hairpin to a double-stranded DNA fragment, so both strands can be read when the fragment unfolds into a single strand.
Also, he said, nobody knows yet what the best aligner is for this type of sequence data, and improvements in base-calling from the raw signal could also make a difference.
According to a US-based MinIon Access Program participant, who requested anonymity because his employer has not authorized him to talk about his experience yet, "the data Nick is showing is kind of typical of the [MinIon] data in terms of the error rate and the fact that only a small number of the reads are called on both strands."
Reads can be longer than the one posted, though. "I really have seen a read that aligns to essentially all of lambda," which is about 40 kilobases in length, the US participant said. "You really can get those monster reads, they're just very rare."
Overall, the Birmingham team's experience with the MinIon has been "incredibly smooth considering how early we are in the process," Loman said.
The device is "very robust" – he said he dropped it a few times with no ill effect – and library preparation is "straightforward."
According to the US user, library prep currently requires micrograms of DNA. It involves DNA shearing, end-repair, ligation of adaptors, and MinIon-specific steps, and takes on the order of three hours.
While the software has had "a few hiccups," Loman said, Oxford Nanopore's team is constantly upgrading and fixing it.
In terms of applications, he said the long reads should be useful for scaffolding genomes, and, in conjunction with real-time analysis, for diagnostic microbiology, his team's main interest. In the future, they could also be useful for de novo genome assembly.
Little software is currently available "to exploit the platform characteristics fully," he said, "but the community [is] already rising to the challenge."
The Exeter Sequencing Service at the University of Exeter in the UK is nearing completion of its burn-in phase and plans to run a variety of bacterial genomes in order to characterize the error profile of the technology.
In addition, they plan to run some "medically relevant samples" and DNA from a small eukaryote, Konrad Paszkiewicz, director of the Wellcome Trust Biomedical Informatics Hub and head of the sequencing service at Exeter, told In Sequence.
His team received the MinIon in late April and sequencing consumables in early June, allowing it to start lambda test runs last week. He has been writing about the group's experience on the Exeter Sequencing Service's blog.
According to the blog, after plugging the sequencer into a USB 3.0 port of a computer, it installs the MinKnow software suite. A program called Metrichor uploads the raw data – ion current traces – to the Amazon Elastic Compute Cloud, where base-calling happens, either 1D base-calling for unidirectional reads or 2D base-calling for bidirectional reads.
"It works," Paszkiewicz said, adding that it is "incredible to witness reads being generated in real time" on such a small device. Also, he said, the MinIon is capable of producing "significantly more data than we anticipated," though he did not say how much.
"It is still rough around the edges and there's plenty of room for refinement, but there's no doubt that the underlying technology is sound," he said.
The error rate is high, consisting mainly of deletions, but the long reads, which he said reach up to tens of kilobases, can be aligned to reference genomes, or used as scaffolds for short-read data to finish genomes.
Library preparation is straightforward for anyone with experience in molecular biology, he said, and is very similar to sample prep for Illumina sequencers. Libraries currently cannot be stored for long periods of time but it's possible to stop the protocol at an earlier point for storage and to resume it later.
The software is GUI-based and straightforward, "provided the instructions are followed," he said, but bioinformatics expertise is required to extract and analyze the data effectively. Standard algorithms used to analyze Illumina or PacBio data "do not handle the long read lengths particularly well, but there are tools out there which do work," he said.
"It feels very much like the early days of Solexa sequencing," he said, referring to the sequencing-by-synthesis technology developed by Solexa, which Illumina acquired in 2007.
While his team has not fully evaluated the error profile of the MinIon data yet, "there is plenty of good quality data there with great scientific value," Paszkiewicz said. Initially, he and his colleagues plan to use the data for scaffolding and for full-length transcript sequencing.
Improvements could be made to the base-calling software, reliability of the flow cells, and library shelf-life, he said, and barcoded adaptors would be helpful. Also, new software needs to be developed by the community to take advantage of the MinIon reads.
Oxford Nanopore said a new chemistry will be available in the coming months, he added, which might include some of these improvements.
Overall, "this isn't a product a non-specialist could take on without considerable back-end support in terms of molecular biology and bioinformatics expertise," he said. "But the barriers to entry for users are much lower than for the other major platforms out there, especially as regards to the total cost of ownership."
Oxford Nanopore has not revealed pricing for the commercial MinIon sequencer yet but participants in the early access program are paying $1,000, plus shipment costs.
Competition for PacBio?
While early access users are just starting to produce their first MinIon data sets, some are already predicting Oxford Nanopore will soon compete with Pacific Biosciences.
"Even at this stage, this platform has the potential to steal large chunks out of the market from the likes of PacBio," Paszkiewicz said.
Over the last few years, PacBio has identified several applications for its long reads, which are lower in accuracy than other platforms, for example, in determining long-range cDNA structures and alternative splice patterns, "and now Oxford is going to go in and grab a bunch of those," according to the US early access user.
He cautioned, though, that the typical accuracy of MinIon reads currently appears to be lower than that of PacBio reads, based on the limited data available to the early access user community.
More data to come
In the meantime, many other early access users contacted by IS are awaiting the arrival of reagents, are in the midst of burn-in, or have run their own samples but are not ready to talk about their results yet. According to the US user, a number of program participants have reported their consumables were held up in customs.
Curiously, at least two large-scale sequencing centers – the Genome Center at Washington University and the Human Genome Sequencing Center at Baylor College of Medicine – were unable to participate in the MinIon Access Program. Oxford Nanopore declined to comment on why it did not include these centers, traditionally among the first to test new sequencing technology because of their experience.
Overall, "Oxford Nanpore's approach with the MAP has been a brave one," Paszkiewicz said, and they "should be proud of what they have accomplished so far."
"Researchers have different priorities to commercial companies, so it's a learning experience for MAP participants and Oxford Nanopore alike," he said, noting that there is "a lot of community goodwill" right now that the company needs to maintain.