NEW YORK – Oxford Nanopore Technologies has developed a new sequencing chemistry where a substantial fraction of reads has an error rate of less than 1 percent, or a Q20 quality score.
Released to a small group of development labs earlier this year, the Q20+ chemistry offers a modal raw read accuracy of 99.3 percent, according to the firm. The kit employs a refined motor enzyme and is used with R10 flow cells and the Bonito basecaller.
"It's a huge step forward," said Rasmus Kirkegaard, a researcher at the University of Vienna who has used the kits as part of the developer release to sequence a reference mock bacterial community sample from Zymo Research. "You can get basically reference-quality genomes from nanopore data only," he said.
The boost in accuracy not only improves the actual assemblies but also user trust in the results. "Even with the old data, we got high-quality consensus sequences, but people tended to distrust them," he said. "You would still need to polish them with Illumina. It will be interesting when people demonstrate you don't need to do that anymore."
In a blog post, Oxford Nanopore said it is planning to release the kits to early-access partners "in the coming weeks." Q20+ is one of several improvements the company has made to its long-read sequencing platform. The newest Bonito basecaller, updated in late 2020, includes a "super accuracy" mode that provides 98.3 percent raw read accuracy. Also, another new chemistry, called "Duplex," is an extension of Q20+ which enables sequencing of both the template and complement strand of DNA in succession.
A new PromethIon instrument, planned for release sometime in 2022, would be the size of a shoe box and run up to two flow cells at a time. A $60,000 "starter pack" would include 48 flow cells.
The firm also clarified that it is releasing new products in phases that differ with regard to where they are in the development cycle: "developer release," where new concepts are tried out; "early access," where products are subject to immediate change; "released," for fully available products that are still evolving; and "fully released." Products in different phases come with different warranties and levels of technical support.
The new chemistry will help as Oxford Nanopore plays catch-up in the market for highly accurate sequencing. "Five years ago, the best reads from a consensus of two strands had an error rate of 15 percent," Kirkegaard noted, and even two years ago, the highest basecalling accuracy available with nanopore sequencing was 96 percent. Illumina has long been the gold standard for accuracy in sequencing, but recently, Pacific Biosciences reached 99.9 percent accuracy for long reads with its HiFi sequencing protocol, which distills reads higher than Q20 from its Sequel IIe instrument. Reads can also be selected based on Q score on the Oxford Nanopore platform.
Like Oxford Nanopore's in-house team, Kirkegaard was able to get a modal accuracy greater than 99 percent. He posted data from his experiments using the PromethIon P24 instrument to GitHub, noting that the best chromosome contig he was able to obtain had an indel rate of 0.74/100 kbp and a mismatch rate of 0.37/100 kbp, which he called "pretty mind-blowing for a nanopore-only assembly."
According to Kirkegaard, data from Q20+ runs were good enough for an analysis to suggest that there was nothing to be gained from polishing an assembly made from such reads with Illumina reads. "To actually assemble genomes, we'd need longer reads than what was in the sample," he said, which were around 7 kb. "But that's a problem with the sample, not the sequencer."
An alpha version of Duplex will be part of the Q20+ early access program, an Oxford Nanopore spokesperson said. Some duplex reads are very close to reaching Q30 accuracy — 99.9 percent, or error rates of less than 1 in 1,000. In data posted to Twitter May 21, Kirkegaard showed three duplex reads greater than 15 kbp that had accuracies of 99.7, 99.8, and 99.8 percent, respectively. However, these duplex pairs accounted for fewer than 2 percent of reads, he said.
The Q20+ chemistry is named so because Oxford Nanopore expects it could be even more accurate in the future. "With further iterations on software, it will improve more," Clive Brown, chief technology officer at Oxford Nanopore, said in a presentation at the firm's virtual user group meeting last week.