Skip to main content
Premium Trial:

Request an Annual Quote

Method to Improve Nanopore Sequencing Accuracy Developed by University of Washington Team


SAN FRANCISCO (GenomeWeb) – Researchers from the University of Washington have developed a method to increase the accuracy of nanopore sequencing that relies on varying the voltage applied to DNA as it moves through the nanopore. They  have applied for a patent on the method, which was described in a study published Nature Biotechnology last week.

Jens Gundlach, a professor of physics at the University of Washington and the senior author on the study, demonstrated the method using his lab's nanopore set-up, which is based on a protein nanopore developed from Mycobacterium smegmatis porin A. However, he said the method could be applicable to any nanopore sequencing method. In the study, his team was able to increase the single-pass base accuracy from around 63 percent to 79 percent.

The researchers and UW have filed for a patent on the method but Gundlach declined to disclose potential commercialization plans. Illumina previously licensed the MspA nanopore technology from Gundlach's lab but he did not comment on whether the company had also licensed, or planned to license, this new method.

Oxford Nanopore Technologies, the only company with a commercial nanopore sequencing technology on the market, was previously involved in a patent dispute with Gundlach and Illumina. Illumina sued Oxford Nanopore in 2016, alleging that the firm's nanopore sequencing instruments involved the use of the MspA pore and thus infringed on the patent Illumina had licensed from UW. Oxford Nanopore did not publicly disclose what type of pore it was using, but the firms settled later that year and Oxford Nanopore ultimately released a version of its technology that used a CsgG pore from Escherichia coli, which it licensed from VIB in Belgium.

In the new study, Gundlach developed the variable voltage technique using his lab's MspA pore setup. He explained that typically, when DNA translocates through a nanopore with the use of a processive enzyme that ratchets the molecule through at a speed slow enough for a signal to be detected, that signal is based on the presence of a nucleotide within a pore. Ultimately, basecalling is done based on interpreting current signatures that are derived from a set of four nucleotides, since surrounding bases impact the current trace from a given nucleotide within the nanopore. That works out to "256 combinations of ion current values that you need to discern," Gundlach said. The challenge is that different combinations are often indistinguishable from each other, he explained

Instead, varying the applied voltage changes the force on the DNA, with higher voltages pulling more strongly on the DNA, and so shifting slightly the DNA that is within the nanopore. The team varied the applied voltage between 100 mV and 200 mV at a frequency of 200 Hz in a pattern that's known as a triangle waveform, Gundlach said. For each step of the enzyme, there were around 10 waveform cycles, although the exact number varied. This resulted in a continuous signal curve, which allows for the 256 four-nucleotide current signatures to be more precisely parsed. "The signature becomes a curve segment that has more information, and we can use that to distinguish between the 256 four-nucleotide combinations that are involved in generating the ion current," Gundlach said.

The variable voltage also helps with another problem of nanopore sequencing, which is identifying when the enzyme has moved too fast for the DNA to be detected. Initially, when the researchers implemented the variable voltage technique, the team noticed that every now and then, the ion current curve that was generated would have "jumps" in it, where the curve was no longer a smooth, continuous line. Gundlach said that the researchers figured out that those jumps corresponded to the enzyme errors — when it would simply race past a nucleotide, for example. Being able to identify these jumps allowed the researchers to correct for those enzymatic errors,  he added.

The main challenge with the method, Gundlach said, was interpreting the new type of signal. Varying the voltage meant the readout no longer directly corresponded to a sequence. Any given nucleotide would elicit a different signal, depending on the actual voltage that is being applied when it is in the nanopore. "We had to correct for the effect of the ion current changing all the time," Gundlach said. "We just had to subtract and correct for the fact that more current flows through when the voltage is increased."

Mike Schatz, an associate professor of computer science and biology at the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University, who was not involved with the study but has broad experience in the sequencing field, including with nanopore sequencing, said in an email that it is a "pretty interesting paper as accuracy is one of the biggest remaining challenges before we will see total adoption of nanopore sequencing." However, he noted that it was important to "keep in mind that they performed these tests using a very low throughput academic nanopore technology that consists of a single pore."

In addition, he noted that while the improvement in accuracy that the team demonstrated — from around 63 percent to 79 percent — was impressive, his lab routinely achieves a single-read accuracy of nearly 90 percent using Oxford Nanopore's MinIon and PromethIon with 1D reads. He said it would be interesting to see whether the technique described in the Nature Biotechnology study could be applied within one of Oxford Nanopore's devices.

Gundlach said that his group plans to keep working to improve the method and to further understand the helicase enzyme itself. In addition, his group has been collaborating with Floyd Romesberg's lab at Scripps Research to use the variable voltage to measure "unnatural bases" — synthetic nucleotides that can form a third type of base pair and would have applications in DNA storage. Conventional next-generation sequencing methods would not be able to detect these unnatural bases, Gundlach said, but in previous work, the UW and Scripps labs have shown that nanopore sequencing can read them. Gundlach said the two are now collaborating to use the variable voltage method.