By Julia Karow
Addressing a major challenge in DNA nanopore sequencing, a team of researchers led by Columbia University has developed an integrated nanopore sensing platform that allows them to take faster measurements of weak current signals.
One of the big issues with nanopore sequencing is that DNA moves through the pores at a speed between 10 nanoseconds and 1 microsecond per base — too fast to measure the ionic current associated with each base accurately. Many research labs and companies, including Oxford Nanopore Technologies, have therefore put a lot of effort into slowing the DNA down, for example by coupling the molecules to ratcheting enzymes, such as polymerases (see story, this issue).
While those are useful strategies, "it would be nice to come closer to the natural rate at which DNA goes through the pore," said Jacob Rosenstein, a graduate student in Kenneth Shepard's lab at the department of electrical engineering at Columbia University.
He and his colleagues have developed a new complementary metal-oxide semiconductor, or CMOS, current preamplifier that improves the signal recorded from solid-state nanopores, thus allowing them to take faster measurements. Earlier this month, they published a description of their nanopore sensing platform online in Nature Methods.
The speed at which weak currents through nanopores can be measured is not limited by the speed of the electronics, Rosenstein explained, but by the signal-to-noise ratio. "Whenever you try to measure something faster, you inevitably have more noise in the measurement." One way to enable faster measurements is, thus, to lower the noise.
One source of noise is ions or electrons accumulating in certain areas for short periods of time, a phenomenon called "parasitic capacitance." For example, ions can gather on either side of the membrane that carries the nanopore, and electrons can accumulate in wires between the device and the amplifier or in the input to the amplifier itself.
That problem is well known in electrophysiology, which has addressed it by using a headstage, or preamplifier, that is placed next to the nanopore. Rosenstein and his colleagues have developed this concept further by designing a circuit of very small CMOS preamplifiers that are exposed to the fluid chamber on the trans side of the nanopore. To avoid parasitic capacitance, they used very low-capacitance nanopores, designed their preamplifier for low-input capacitance, and placed it in the fluid chamber to minimize any wiring.
Usually, the types of preamplifiers they used require a large value resistor, he said, but those are not available for an integrated CMOS process, so instead, they designed a system with an active control loop that behaves similarly to a resistor. They had prototypes of their design manufactured by a standard commercial semiconductor foundry line.
The researchers then used their nanopore platform, which consists of eight low-noise current preamplifiers mounted underneath silicon nitride nanopores, to measure the current trace of short DNA oligonucleotides of up to 50 base pairs.
While they were unable to determine individual bases from the current trace, they could record the current signal with a bandwidth of 1 megahertz, about ten times faster than commercially available amplifiers, which support bandwidths of up to 100 kilohertz.
Besides allowing DNA to be measured faster, the signal improvements from the preamplifier could also help increase the accuracy of nanopore sequencing, Rosenstein said. "It's never going to hurt you to have lower noise."
In addition, because the preamplifiers are small — about 500 would occupy a square centimeter — the number of nanopores on an array could probably be scaled to thousands.
Increasing the number of nanopores to millions, however, might be difficult. "At least with this particular type of design, when you make an integrated circuit, you can't make it arbitrarily large because of yield problems in the manufacture," he said.
In principle, the nanopore sensing platform could also be used with biological nanopores, but the signal level from those is usually lower than that from solid-state nanopores, which limits improvements in measurement speed. "If you need to resolve a 1 picoampere signal, you will not be able to do it at 1 megahertz, even with our system," Rosenstein said. "To make use of the higher bandwidth, you need to be starting with a little bit stronger signal."
The researchers have patented their technology through Columbia University and are currently in discussions with several parties who are interested in licensing it for DNA sequencing and other applications, he said.
More Nanopore Sensing Platforms
Several companies, including Oxford Nanopore and Genia, have developed their own nanopore sensing platforms. While few details about their chip designs are currently available, they are "probably thinking along similar lines," Rosenstein said. Both Oxford Nanopore and Genia currently use biological nanopores.
Oxford Nanopore, for example, has developed an instrument called the GridIon that contains "custom, high-performance electronics" to measure the ion current through individual nanopores at high frequency, according to its website. Data is transferred from a sensor array chip in a cartridge, which initially carries 2,000 nanopores, to a custom application-specific integrated circuit, or ASIC (IS 2/21/2012).
In addition, Oxford Nanopore has developed the MinIon, a disposable instrument that combines the nanopores, sensor chip, and ASIC in a small device.
Startup Genia has also developed a nanopore sensor platform, operating an alpha chip version with several hundred nanopores, and plans to scale the number of sensors to a million for a commercial product (IS 1/17/2012).
A company spokesperson told In Sequence that Genia uses a different architecture than what the Columbia researchers developed. Genia CEO Stefan Roever said in an e-mail that the paper shows that "by moving the CMOS technology away from discrete electronics, it is possible to dramatically improve signal-to-noise at higher bandwidths."
"Challenges still remain with using solid-state pores to do single-base discrimination," he added. "If these can be solved, solid-state pores may become a viable alternative to biological nanopores."
Rosenstein and his team are now improving their platform further to enable even faster measurements, both by making nanopores with increased signal levels and by further decreasing their noise. "With all this together, there is probably room to go to several megahertz, approaching 10 megahertz" in bandwidth, he said.
One way to increase the signal is to use graphene nanopores. The researchers are collaborating in this area with Marija Drndic's group at the University of Pennsylvania. Drndic is also an author on the Nature Methods paper.
In collaboration with Meni Wanunu at Northeastern University, another author, they are studying the translocation of "various small molecules" through the nanopores.
In addition, the Shepard group is now exploring the use of other types of transducers, such as carbon nanotubes or field effect transistors. Those provide stronger signals, potentially allowing for measurements as fast as tens of megahertz, but "they have other challenges with signal to noise, and their own properties," Rosenstein said.
Have topics you'd like to see covered in In Sequence? Contact the editor at jkarow [at] genomeweb [.] com.