One of the long-held goals for molecular testing of any sort (from testing the food supply for food-borne illness, to livestock health, to forensic testing, to testing for infectious disease, to screening for cancer) is to obtain inexpensive results without a lot of expensive equipment. The latest effort at ‘point-of-care’ testing has been to take existing diagnostic equipment and technology (namely real-time PCR) and miniaturize it as much as possible, for example the Cepheid GeneXpert Omni (profiled here from AACC 2015) or the Belgium-based BioCartis’ Idylla (likewise profiled here). The detection technology has stayed constant, but the use of clever optical-quality plastic and cartridges that purify as well as run the PCR as it is monitored in real-time for optical fluorescence means a system in a few-tens-of-thousands of dollars to acquire, and running costs of a few-hundred dollars.
For the sake of completeness, I neglected in the past to write up the Roche cobas LIAT system. LIAT is an acronym for ‘Laboratory In A Tube’, and like the Cepheid and BioCartis systems, also is real-time qPCR-based.
What is the problem to be solved here? There are multiple dimensions, but mainly it is two-fold: to lower the cost of instrumentation and consumables,and to miniaturize and simplify the operation of the system to get the test closer to the where the sample is collected. Many times a central laboratory with specialized equipment and trained operators will delay results and increase costs. These samples can be from a bunch of spinach suspected of harboring a nasty strain of O157:E7 E. coli or a throat swab of a sniffling 6-year old.
The cost of instrumentation is moving down-market from the high five-figures to the low five-figures, which is an improvement; the assays are becoming easier to run as well, with the Cepheid cartridges being the first widely adopted diagnostic platform (if memory serves) to combine sample preparation with a real-time assay in a single easy-to-use cartridge.
It is in this market backdrop that there has been a steady increase in interest in single-molecule detection systems. First with Oxford Nanopore (in case you were wondering, I wrote up about ONT way back in February 2014), a system powered by a simple USB port and about the size of a large candy bar, has seen use for field use in Africa to characterize Ebola (here’s a nice Guardian piece from 2016 about strain identification within 24 hours, complete with some great photos of Guinea scientists running the system and linking to the related Nature publication). Subsequently, work on Zika detection using Oxford Nanopore has been performed, and the MinION has recently been deployed on the International Space Station.
However this single-molecule biological nanopore system has some limitations. Yes it is very compact, and the price to acquire is about $1,000 and as little $600 to run (note: a friend tells me that it is $600 per run in volume pricing, including the cost of a 1D library preparation). This $600 for a single run will limit its effectiveness, as there are not many situations where the value can justify a $600 test. I can imagine life-or-death situations like Sepsis or perhaps a real-time fatal epidemic such as Ebola testing in Africa.
Another limitation is their sample preparation steps, but Oxford’s upcoming electrowetting device called the Voltrax may help solve it (no expectations on when that will be available). There is still a requirement for the specialized equipment (centrifuges, -20C freezers for enzymes, thermal cyclers, etc) that place practical limitations on sequencing as a field-deployable system for diagnostics. There are still formidable infrastructure issues in an inhospitable place like the remote countryside of Guinea – Nick Loman gave a great talk at ESHG in Barcelona last year about what that all involved; here is his description of the work for those interested.
Two Pore Guys has developed a new approach, not for sequencing but for probe detection, using single-molecule solid-state nanopores. By ‘solid-state’, this means the pore is made of semiconductor materials, rather than biological nanopores (like used at Oxford Nanopore as well as the upcoming Genia sequencing platform at Roche). Two Pore Guys uses two nanopores linearly aligned, with a space in the middle for the detector to measure capacitance. The photo can explain it better.
DNA is drawn through the two aligned nanopores via an electrical current, and the use of two nanopores instead of one allows better control of the speed of the DNA as it passes through the pore. For Oxford Nanopore (as well as other biological nanopores), a lot of protein engineering has gone into the ‘ratchet’ mechanism to slow down the speed of the DNA strand as it passes through. And their ability to generate sequence has been 25 years in the making.
At last week’s Molecular Tri-Conference in San Francisco, Two Pore Guys’ CSO Trevor Morin presented the technical details: 30nm solid-state pores, and as current draws single DNA molecules through, the changes in impedance (measured perpendicular to the direction of the DNA traveling through the pores) can be measured in both strength of signal and duration. This is not sequencing of individual base changes, this is recognition of a particular probe or molecule bound to a DNA strand.
By showing a brief video during his presentation (they showed this one, although likely it was an abbreviated version) they showed the ability to take a simple cheek swab, put it in some kind of lysis buffer, put a few drops onto a test strip (housing the micro-chip with the nanopores on it) and place it into a handheld reader.
Is this an oversimplified workflow? Can it be that simple? And such a small device do what they claim?
The elimination of an optical analysis of signal for genetic analysis is not new; Combimatrix is a company that started in the late 1990’s with repurposed 8K RAM chips, and Autogenomics uses a BioFilmChip from the same timeframe. Of course a decade later Ion Torrent would do the same for sequencing. But the real ‘secret sauce’ in Two Pore Guy’s case is single molecule analysis with solid-state electronic detection.
Instrumentation for single molecule analysis with optical detection is large. The original PacBio RS II was famous for its size; electrical detection via Oxford Nanopore is famous also for its size.
Admittedly the Two Pore Guys’ data was limited; they showed data that allowed up to five discrete kinds of signals to be detected (i.e. five analytes per sample). Each analyte could be a specific DNA sequence, or a protein bound to an antibody, or even a receptor-ligand conjugate. The beauty of the system is that the DNA doesn’t have to be biological in nature for the readout; one of their videos on Vimeo shows data from a synthetic molecule going through these pores. Imagine the DNA is just a rail, and the cargo on that rail is what you are interested in detecting (and picked up by the system).
An interesting feature is the that the pattern of the cargo can be varied with the same probe, and the amplitude of the signal as well as the time it takes to pass through (i.e. integration of the signal) can provide a unique signature. (For more information about this, here’s a video entitled ‘2PG Data Molecules’.)
The sample preparation was low-cycle PCR for semi-quantitative data; amplification-free is also claimed; a LOD of less than 50 discrete target molecules; 1% sensitivity and 2% error. (Note: these were all claims by Dr. Morin.) One early application and collaboration is with UCSF, looking at KRAS G12D and an 89-bp unbiased amplification target for ctDNA analysis.
Additional applications abound, and he referenced looking at ligand/antibody binding kinetics with the platform, with sensitivity in the low picomolar range, and a dynamic range of over three logs.
Their CEO Dan Heller then presented their business model, which struck me as unusual. They are offering an exclusive license to the assay developers on this platform on an application-specific basis, with narrowing by geography or other parameters as needed. And the royalty paid to Two Pore Guys was also flexible depending on the market; for agricultural or livestock applications, where there is high price sensitivity the royalty is set low, while for human applications where there is much less price sensitivity the royalty is four times as high.
He got into specifics in the Q&A afterwards, in both COGS and expected pricing. Suffice it to say, a given test may cost on the order of $10-$15 (or much higher depending on the market!), and the handheld, portable reader will be in the few hundreds of dollars.
In the strategic business consulting world, there is a powerful concept called the value chain. As a new Product Manager in the early 2000’s at QIAGEN I was involved with their Luminex partnership, and experienced terrible channel conflict at that time as we sold directly against Bio-Rad, offering the same platform but different assays and software to differentiate the platform. Luminex at that time had the platform but not the assays, and partnered with assay experts. Over time, they changed their leadership and their business model, and today they sell the platform and the assay. There is too much of the value chain tied up in the assay, less so on the platform.
Many diagnostic companies go that route; the market values the problem to be solved, not the technology that enables the problem to be solved. What good is a smartphone without the software apps to run on it? What good is assisted reality without useful objects to overlay my field of view? What good is a nanopore device without an assay to detect HIV directly from saliva?
I’m sanguine about Two Pore Guys – very much a technology to watch, across application markets from diagnostics to veterinary care to food safety to forensics.