NEW YORK – Researchers at the Broad Institute have developed a new sequencing method called Concatenating Original Duplex for Error Correction (CODEC) that can accurately detect rare mutations in samples such as tumor tissues.
At the annual Molecular & Precision Med Tri-Con conference in San Diego last month, Viktor Adalsteinsson, associate director of the Broad's Gerstner Center for Cancer Diagnostics and the study's corresponding author, presented details of the approach, a variation of the duplex sequencing method.
First published in a BioRxiv preprint last year, CODEC promises to cut down on the amount of sequencing required to distinguish low-frequency mutations in cancer and other samples from DNA sequencing errors with high accuracy. The authors have filed a patent application on the method, although Adalsteinsson said that it was too early to comment on its potential licensing and commercialization.
Researchers at the University of Washington first pioneered duplex sequencing in 2012, which separately tags each strand of a DNA molecule to later match and compare those complementary fragments.
"By reading both strands of each DNA molecule, [duplex sequencing] enables a more accurate readout of the DNA sequence than if you only read one strand on its own as most sequencers are traditionally configured," Adalsteinsson said in an interview.
CODEC attempts to build on the strengths of duplex sequencing by linking the Watson and Crick strands of a DNA duplex into a single molecule, which can then be amplified and sequenced in a standard NGS workflow.
According to Adalsteinsson, the method can be "a simple add-on to most major NGS workflows."
Last year, his group published another method, called Minor Allele Enriched Sequencing Through Recognition Oligonucleotides, or MAESTRO, that combines duplex sequencing with mutation enrichment.
In a head-to-head comparison of CODEC and duplex sequencing, Adalsteinsson and his colleagues found similar error rates, on the order of one in 1 million, when applying both techniques to pan-cancer panels and to whole-exome sequencing libraries. CODEC, however, achieved this error rate with approximately 100 times fewer reads.
Less sequencing implies less money spent on sequencing, and while Adalsteinsson did not provide specifics on CODEC's costs, he mentioned that his lab is already experiencing cost savings in using the method.
He expects CODEC's accuracy to open some "exciting possibilities" for looking at various biological samples. These include research applications, such as benchmarking whole genomes, germline variant calling, and shotgun metagenomic sequencing for microbiome analysis, as well as diagnostic applications.
"Since posting the preprint, we've had a lot of interest from many collaborators and have a number of different studies underway," Adalsteinsson said.
Interest has come from both academic groups and the commercial sector, he said, with projects ranging from applying CODEC to cancer detection and monitoring with liquid biopsies to rare mutation detection and analysis of somatic mosaicism in blood cells and tissue biopsies. However, he declined to name collaborators.
While TwinStrand Biosciences, a company that has been developing duplex sequencing commercially, is not collaborating with Adalsteinsson's group, CEO Jesse Salk praised the advances being made in duplex sequencing through publications from "strong academic groups" like the Broad team.
"This says a lot about the increasing appreciation within the scientific and clinical genomics community of the value of high accuracy and low-frequency mutation detection," he said via email.
TwinStrand also works with many companies and academic groups seeking to adapt the company's own duplex sequencing technology to various applications, such as Foundation Medicine, which is applying duplex sequencing to liquid biopsy testing.
Salk noted that there may not be a one-size-fits-all duplex sequencing solution.
"Depending on the application, different may be desirable," he said. "There are important and often complex trade-offs between different duplex sequencing approaches that balance accuracy, conversion efficiency, cost, speed, reliability, ease of targeted sequencing, and other important factors."
Long-read sequencing platforms have also taken advantage of improved error rates from sequencing both strands of a DNA duplex, but their overall accuracy still lags behind what short-read sequencing can achieve with duplex sequencing.
Pacific Biosciences, for example, relies on circular consensus sequencing, or sequencing both strands that are connected into a circle several times over, to create highly accurate HiFi reads. Also, Oxford Nanopore last year released a new chemistry, called Kit 12, that can be used to combine base calls from both strands for a quality score of Q30, translating to an error rate of approximately one in 1,000.
"Being able to leverage the high-throughput nature of next-generation sequencing, but conferring substantially greater accuracy, we think [that] we're going to begin to reveal new and exciting biology [that] we've never seen before," Adalsteinsson said.