After developing the chemistry for its sequencing-by-expansion nanopore technology for several years, Stratos Genomics is ready to grow its development and commercialization efforts.
The Seattle-based company is in the midst of a Series B funding round in which it hopes to raise at least $15 million by the end of March to develop a low-cost sequencing platform that combines the speed and throughput of nanopores with improved resolution and signal to noise, both for targeted and whole-genome sequencing.
Stratos Genomics was founded in 2007 as a spinoff of the Stratos Group, an engineering firm that acts as a holding company to incubate new technologies in various industries, including life sciences, medical devices, and consumer electronics. Stratos Genomics, which has almost 20 employees, shares space with its parent near the Seattle downtown waterfront. Allan Stephan serves as CEO of both firms.
In 2010, Stratos Genomics raised $4 million in a Series A financing round led by Fisk Ventures that also included the Stratos Group. The same shareholders provided an additional $2.1 million in 2011, another $2 million in 2012 after the firm demonstrated that it could sequence a 36-base DNA template, and $2 million more in the fall of 2013 after Stratos showed it could generate a 210-base read.
In addition, Stratos Genomics has garnered grants totaling about $1.3 million from the National Human Genome Research Institute and from the Defense Advanced Research Projects Agency.
The firm's technology, called sequencing by expansion or SBX, first converts DNA into a larger surrogate molecule, called an Xpandomer, which contains reporter molecules that represent the DNA bases and generate a strong signal. The Xpandomer, which is about 10 to 100 times longer than the original DNA, then passes through a nanopore with a detector to read out the signal, which is translated back into the DNA sequence.
Stratos hopes to use this approach, which is protected by several patents, to overcome the resolution problems many nanopore sequencing technologies have been grappling with because the DNA passes too fast through the pores, and "to fundamentally address the signal-to-noise problem of single-molecule detection," CSO Mark Kokoris told In Sequence.
A polymerase synthesizes the surrogate molecule off a DNA template by incorporating so-called X-NTPs, nucleotides that contain a hairpin with a reporter, a tether, and a cleavable bond. The hairpins are then opened through a rapid chemical cleavage reaction, which expands the surrogate molecule.
It is unclear how long the DNA molecules to be converted can be, but the company is targeting 200 bases for its first commercial applications. Because the biochemical conversion is separated from the detection, it also has the ability enrich for long surrogate molecules.
Initially, Stratos pursued a different, ligase-based approach in which it used hexamer, tetramer, or dimer probes with built-in reporters as building blocks to generate the surrogate molecules in a ligase-mediated reaction. While this requires fewer building blocks per DNA molecule than the polymerase-based approach – 34 hexamer probes for 200 bases of DNA, for example, instead of 200 X-NTPs – it needs a larger library of building blocks, for example 4,096 different hexamer probes instead of only four types of X-NTPs.
Using five different hexamer probes, the company demonstrated last year that it can read a stretch of 210 bases, its longest template converted so far. The hexamer probes also allowed it to encode homopolymers, which some sequencing technologies find difficult to score accurately.
But to keep the library size small and the cost of development low, Stratos decided a little over a year ago to develop the X-NTPs and switch from the ligase to the polymerase approach, according to Bob McRuer, the company's CTO. Last year, Stratos "hit several milestones in terms of making intermediate structures" for the X-NTPs, and by the end of 2013, it was able to make a full library of all four expandable nucleotides for the first time, "which is a major achievement for us," he said.
For now, the company is focusing on the polymerase-based approach, although it might pick up the ligase-based approach again and develop it in parallel, if funding permits.
Stratos does not reveal the exact nature of its reporter molecules, which are "mixed composition polymers" that differ according to the detection technology used. "We can put in reporters that are very specific and tuned to the detection technology, so we can have extremely high signal to noise and be very discriminative between two different bases," Kokoris said.
Although the firm uses wild-type alpha hemolysin as its nanopore detector at the moment, which Kokoris said is "performing extremely well," it could use other types of detectors, such as solid-state nanopores, electron beam scanning, or nanowires.
According to McRuer, it is premature to speculate on the technology's accuracy, though the firm has seen "extremely good specificity" with the small ligation probe set it has used so far.
Stratos is now embarking on optimizing its biochemistry, including engineering polymerases to better incorporate the X-NTPs and ways to expand the substrate more efficiently.
Ultimately, the company wants to commercialize platforms for both targeted and whole-genome sequencing. "We really think that we have some economic advantages and engineering advantages in both spaces," CEO Stephan said.
One advantage over other nanopore sequencing platforms will be the ability to detect homopolymers accurately, Kokoris said, noting that "the market has spoken very clearly that that is a requirement" of any sequencing system.
Further, because the signal to noise ratio is higher than in other nanopore systems, throughput can be faster. "We believe that high signal to noise is necessary to go faster in order to get the throughput to compete at a whole-genome level," he said.
Finally, separating the biochemistry and the detection steps means the detection instrument can be lower in cost.
Stratos hopes to achieve commercialization in less than three years, Stephan said. One version of the technology, optimized for targeted sequencing, could be produced for less than $5,000. Another, targeted at whole-genome sequencing, would have a production cost of about $20,000 and be able to sequence a human genome in about 2.5 hours. The biochemical conversion would be simple, easy to automate, take about 30 minutes, and require less than $40 per genome in consumables.