Improving Sequencing Errors and Detection of Rare Mutations

There is little doubt within the scientific community of the immense impact that next-generation sequencing (NGS) has had on translational research. Techniques like liquid biopsy have allowed investigators to harness the power of NGS to identify genetic mutations from blood samples noninvasively. Identifying such mutations enables earlier diagnosis of cancer and can inform treatment decisions for many patients. However, NGS does not come without its own caveats, as high error rates are common among NGS techniques—often limiting their clinical application. Now, researchers at the Johns Hopkins Kimmel Cancer Center have developed a new system they dubbed SaferSeqS (Safer Sequencing System) to improve upon previous technology and correct sequencing inefficiencies and errors.

The Johns Hopkins team published findings of their new technology recently in Nature Biotechnology through a new study titled, “Detection of low-frequency DNA variants by targeted sequencing of the Watson and Crick strands.” The new SaferSeqS technology detects rare mutations in blood in a highly efficient manner and reduces the error rate of commonly used technologies for evaluating mutations in the blood more than 100-fold.

The presence of a mutation in a clinical sample could be an early indicator that a person has developed cancer, said study lead author and MD/PhD candidate Joshua Cohen. Cancer is a genetic disease driven by oncogenes and tumor suppressor genes. A small portion of cancer cells shed their DNA into the bloodstream, allowing their mutations to be detected via blood samples. Detecting such mutations in the blood rather through surgical biopsy of cancerous tissue is called “a liquid biopsy.” Such blood-based tests have the potential to detect cancer at an earlier stage when it can be put into remission by surgery and chemotherapy. The challenge, Cohen explained, is that the vast majority of DNA present in the blood sample is shed by noncancer cells, and only a tiny fraction of the DNA is derived from the tumor. In patients with relatively early-stage cancers, a 10 mL blood sample will only contain a handful of molecules with a mutation.

“To detect cancers when they have the best chance of being cured requires a detection method that will pick up cancer signals that are present at extremely low frequencies,” Cohen noted. “The technical challenge in detecting these mutations is akin to finding a needle in a haystack.”

The researchers addressed this challenge, with SaferSeqS, by efficiently tagging both strands of each original molecule present in an individual’s blood with a unique barcode. It required new biochemical approaches to do this in an efficient manner with the small number of degraded DNA molecules that are usually present in the blood. The investigators use the structural redundancy of the double-stranded DNA molecule to distinguish real mutations from errors, an approach called duplex sequencing. If both strands of a DNA molecule contain the identical mutation, it is far more likely that it is a confirmed mutation and not an error.

“We describe a method, termed SaferSeqS, that addresses these challenges by (1) efficiently introducing identical molecular barcodes in the Watson and Crick strands of template molecules and (2) enriching target sequences with strand-specific PCR,” the authors wrote. “The method achieves high sensitivity and specificity and detects variants at frequencies below 1 in 100,000 DNA template molecules with a background mutation rate of <5 × 10^–7 mutants per base pair (bp). We demonstrate that it can evaluate mutations in a single amplicon or simultaneously in multiple amplicons, assess limited quantities of cell-free DNA with high recovery of both strands, and reduce the error rate of existing PCR-based molecular barcoding approaches by >100-fold.”

“What makes SaferSeqS unique is the efficient tagging of both strands of the majority of DNA molecules circulating in the blood, the low error rate achieved through analysis of both strands of these DNA molecules, and the manner in which the molecules of interest are enriched prior to sequencing,” Cohen added. “Altogether, these advancements underlie the power of the new technology. Every molecule is sacred because it has the potential to be the one with the mutation we’re looking for—because the absolute number of molecules is low, the technology has to be highly efficient at capturing each molecule to sensitively identify mutations.”

The researchers wanted to test the specificity and sensitivity of SaferSeqS in a clinically relevant setting, so they compared the samples to previous results from the CancerSEEK test, a single blood test that screens for eight common cancer types developed and reported by the same research team.

In the study, the Hopkins investigators revisited 74 blood samples from patients with cancer that had false-negative results—undetectable mutations—in the 2018 CancerSEEK study using SafeSeqS. In their newest study describing SaferSeqS, the researchers reassessed these blood samples. Using SaferSeqS, they observed a marked improvement in sensitivity, finding previously undetectable mutations in 68% of the samples tested.

Taking these results together, the researchers concluded that SaferSeqS is highly sensitive and specific for detecting extremely rare cancer-related mutations, is potentially efficient and cost-effective for clinical use, and reduces the error rate of existing mutation-detection approaches more than 100-fold.

The next step, the researchers said, is to validate the results and demonstrate the clinical usefulness of the technology in prospective clinical trials.

“The SaferSeqS strategy affords highly reliable technical specificity, which translates to a better way to provide clinically meaningful results for patients with relatively early-stage and small tumors,” Cohen concluded.