Skip to main content
Premium Trial:

Request an Annual Quote

Hopkins Team Develops Method to Improve Rare Mutation Detection for Early Cancer Dx


By Julia Karow

Rare variants are still tricky to track down using next-generation sequencing because it is difficult to distinguish them from sequencing errors. To improve on this, researchers at the Johns Hopkins Kimmel Cancer Center have developed a template tagging method that allows them to distinguish errors from real variants.

The scientists recently published their approach, called Safe-Sequencing System, or Safe-SeqS, in the Proceedings of the National Academy of Sciences. They plan to use it in early diagnostic tests for cancer.

"If we can search for specific DNA mutations in patients, perhaps we can detect tumors at an early, more curable stage," said Isaac Kinde, a graduate student in the Ludwig Center for Cancer Genetics and Therapeutics at Johns Hopkins and the first author of the paper.

The idea is to search for common cancer mutations — primarily point mutations — in DNA that has been shed by the tumor into bodily fluids such as blood, stool, or saliva. However, the tumor DNA only makes up a tiny fraction of the total DNA content of the sample, and is present in minute quantities. Analyzing it requires a process that does not lose too much of the DNA, and a method that can distinguish between bona fide mutations and technical artifacts, Kinde explained. Safe-SeqS might be able to do just that.

First, the researchers assign a unique identifier, or barcode, to each DNA template molecule. This could be an endogenous barcode that is already part of the template or an exogenous barcode that is added to its end. Next, they amplify each tagged template, resulting in many daughter molecules. A mutation that was present from the start should be contained in all the daughter molecules, whereas a variant that was introduced through a technical error should only appear in a fraction of the daughters.

The number of different template molecules that can be analyzed in this way is determined by the number of barcodes — in their paper, the researchers used primers with a stretch of 12 to 14 random nucleotides. But according to Kinde, what usually limits the technique is not the number of barcodes but the sequencing capacity of the instrument, because for each template molecule, several daughter molecules need to be sequenced in order to determine that a mutation is genuine. For most experiments, though, the current capacity of the Illumina Genome Analyzer, which he and his colleagues have been using, is more than sufficient, "and it's increasing all the time."

Overall, Safe-SeqS improves mutation detection at least tenfold compared to conventional massively parallel sequencing, he said. And the advantage over other techniques that detect point mutations at specific locations is that Safe-SeqS can query several bases at a time.

The technique is also compatible with any massively parallel sequencer, both current platforms and those still in development.

So far, the researchers have only shown that they can analyze a single amplicon using exogenous barcodes, but there is a "relatively straightforward way" to amplify and analyze several amplicons at a time, Kinde said.

The approach bears some resemblance to another tagging method to improve the accuracy of variant calling in next-gen sequencing that was recently published by Population Genetics Technologies, a UK-based startup (IS 5/10/2011).

The main difference, according to James Casbon, PopGenTech's bioinformatics officer, is that the Hopkins method sequences each template molecule several times to verify a variant, whereas PopGenTech sequences a variant in multiple template molecules to establish it.

[ pagebreak ]

He also pointed out that PopGenTech, in its paper, predicted the number of "collisions," or templates receiving the same barcode by chance, which reduces the number of different mutations that can be detected.

The next step in developing Safe-SeqS, Kinde said, is to validate the method in a large number of samples from patients with a variety of cancer types as well as normal controls.

The focus will be on cancers for which common point mutations have already been discovered through sequencing studies. For colorectal cancer, for example, researchers already have "a good idea of the mutations that occur," he explained, so there are "many opportunities to detect point mutations in a feasible way with our technique."

Depending on how easily detectable the cancers are, the researchers might then focus on particular cancer types to develop early diagnostic tests. Kinde said that he and his colleagues plan to pursue patent protection for the technology and hope that a company might eventually commercialize it for early cancer detection, though he did not mention any concrete plans for that.

The method might also have other applications, for example to monitor rare drug resistance mutations in HIV or other viral diseases, he said, though his lab is not pursuing those areas. "We hope that this will encourage interest and that other people will also be using this technology," he said.

Have topics you'd like to see covered by Clinical Sequencing News? Contact the editor at jkarow [at] genomeweb [.] com.