NEW YORK (GenomeWeb) – An international research team has developed a next-generation sequencing method to better detect rare variants, which they said could have applications in oncology and beyond.
The method, dubbed SiMSen-seq for simple, multiplexed sensitive sequencing, relies on a three-cycle barcoding PCR step followed by adapter PCR for library construction. It is essentially a barcoding strategy, but differs from other techniques in its use of a stem loop structure to "protect" the barcode. The method enables variants to be detected at an allele frequency as low as .1 percent. The researchers also developed an analysis toolkit, Debarcer, for de-barcoding and error correction, which is freely available for research use.
Researchers from Boston University, the Ontario Institute for Cancer Research, and the University of Gothenburg in Sweden collaborated on the work and have filed a patent application on the method.
Anders Ståhlberg, an associate professor at the Sahlgrenska Cancer Center in Sweden, and Paul Krzyzanowski, program manager of OICR's genome technologies program, were both visiting the Boston University laboratory of Tony Godfrey, an associate professor of surgery.
The researchers had all encountered problems with multiplexed molecular barcoding techniques. For instance, Krzyzanowski explained in an interview, when attaching a barcode to a target, that barcode consists of a series of random bases that serve as a tag. Doing that in a multiplexed fashion results in lots of different stretches of random sequence, so that the barcodes themselves start acting as primers, creating off-target effects. So, the researchers began working together on a different type of barcoding method.
SiMSen-seq avoids the problem of sequencing off-target DNA by essentially keeping the barcode hidden until it is needed. The method uses primers with a stem loop structure. The barcode is contained within a hairpin during the first PCR cycle so that the different barcodes don't interact with each other and serve as primers themselves. However, the researchers "designed the hairpin structure to behave like a temperature controlled switch," Krzyzanowski said. After, the first PCR round, the temperature is raised, which dissociates the stem and enables the barcode to be read for sequencing.
In addition, the researchers optimized PCR conditions in such a way to enable the products from the first PCR step to be used for adaptor PCR library construction. Ståhlberg said in an interview that the team borrowed from techniques used in single-cell sequencing methods where researchers have to amplify limited amounts of sample. Specifically, he said, the team figured out ways to decrease the primer concentration and the time for the annealing step.
And finally, another important aspect is that it can work with limited amounts of DNA, 30 to 40 nanograms, Krzyzanowski said.
In the study, the researchers detailed the steps for designing an assay as well as for library preparation. They tested the protocol's multiplexing capability up to 40 amplicons., though Krzyzanowski said that the team is continuing to work on enabling it to multiplex higher numbers of amplicons. Forty amplicons essentially equates to around 4 kilobases of target, he said, which is not very large. For instance, one leukemia panel that the OICR runs is more than 1 megabase.
"We don't need all that for SiMSen-seq," he said, "but we need to get up to something that's useful across a wider range of applications."
The challenge with increasing the panel size, he said, is primarily the amount of labor and time that goes into designing and validating the panel itself and optimizing the performance of the different amplicons. In the current study, the researchers describe shifting primers two to three bases can help reduce the number of nonspecific PCR products. As the number of amplicons in an assay increases, more work is required to reduce off-target interactions and ensure uniform amplification.
Ståhlberg added that another restriction to increasing the level of multiplexing is the cost of sequencing. "It's not enough to see each original molecule once," he said. And once targets are amplified and sequenced multiple times, the costs quickly escalate.
He added that he viewed the technique as a "bridge" between high-throughput sequencing and digital PCR. It has greater sensitivity than standard NGS techniques, but greater throughput than digital PCR.
The researchers also developed a software tool called Debarcer to analyze data from SiMSen-seq assays. The tool uses standard bioinformatics approaches to first identify amplicons from the raw sequencing data. Reads are organized by barcode and the tool also filters out mutated barcodes. Krzyzanowski said that the current version of Debarcer is open source, but added that he would potentially look to refine it for specific commercial applications.
In addition, he said that he is working to develop it further so that it could be applied to other molecular barcoding techniques. "I would like to see Debarcer become the SAMtools of molecular barcoding," he said.
The researchers have filed a patent application on the SiMSen-seq method, and the Gothenberg and OICR groups are using the method for different applications.
Krzyzanowski's OICR team is developing a leukemia panel based on the technique that it plans to offer to researchers as a service or as a part of collaboration. The goal will be to use the panel for minimal residual disease monitoring where detecting low-frequency mutations is important, Krzyzanowski said. Currently, MRD monitoring is possible for patients who have BCR-ABL gene fusions, he said, because digital PCR techniques can be used to sensitively screen for that known fusion. But, not all patients have that gene fusion, so the idea would be to design an assay that could detect specific point mutations.
Ståhlberg's team is also working in the cancer field, specifically on developing the technique for liquid biopsy applications. However, he said, he is also collaborating with other researchers to apply it for use in immunology, forensics, and microbiology.