By Julia Karow
A new tagging method developed by UK-startup Population Genetics Technologies promises to improve the accuracy of variant calling in next-generation sequencing, with possible applications in disease research and microbial diversity studies.
The method, published last month in Nucleic Acids Research, allows users to distinguish between true variants and errors introduced during amplification prior to sequencing. PopGenTech is now applying it in several population genetic studies with academic groups as well as pharmaceutical companies, including a study investigating genetic predisposition to Alzheimer's disease.
PopGenTech's approach "has the potential to allow us to study a large number of samples for candidate regions and get the relevant sequence data in a relatively short time," said Julie Williams, a professor at Cardiff University School of Medicine and the company's collaborator on the Alzheimer's study.
PopGenTech was founded in 2005 by Nobel prize winner Sydney Brenner, Applied Biosystems founder Sam Eletr, and investor Philip Goelet to commercialize a proprietary approach that "drastically reduces the amount of work involved in sequencing large numbers of samples," according to its website.
Launched with seed funding from the Wellcome Trust, the firm in 2008 raised £3.8 million ($6.2 million) in Series A venture capital funding from Auriga Partners, Noble Fund Managers (now Beringea), and Compass Genetics Investors. Last year, it won a $1.2 million award from the Wellcome Trust to study genetic predisposition to Alzheimer's disease in collaboration with Cardiff University.
The firm currently has 14 full-time employees and is housed in a bioincubator facility of the Babraham Institute.
As part of the new method, PopGenTech fragments genomic DNA and tags each fragment with an oligo that contains a unique DNA barcode and a run of degenerate bases, along with adaptor sequences required for next-gen sequencing. The degenerate base run allows the scientists to distinguish between individual template molecules. Several hundred samples are then pooled and specific regions of interest PCR-amplified. The PCR amplification process, by its nature, can change the representation of template molecules and may introduce errors. But PopGenTech's tags allow the researchers to distinguish between PCR errors and true variants, because the errors are associated with only one rather than several different template molecules.
According to James Casbon, PopGenTech's bioinformatics officer and the first author of the study, the firm has used the method effectively to analyze populations of samples, both to eliminate false positives during SNP calling and to improve allele calling. "It helps you, really, get the value out of your sequence data," he said.
In addition, he said, it could be applied in microbial diversity studies, for example to determine the number of viral quasispecies in a sample, but he and his colleagues have not yet tested it for that. It could also be used to study tumor samples, for example rare variants present in a subpopulation of cells.
The company has filed for patents on the method and its applications, according to Conrad Lichtenstein, PopGenTech's chief scientific officer, and is using it in all its ongoing studies that involve tagging and pooling genomes. "We want to be absolutely sure when we identify genetic variants that they are bona fide true variants and not just errors that are part of the workflow," he said.
So far, the company has used the method only in conjunction with the 454 platform, which it likes because of its long reads, but it could also be deployed on other sequencing platforms, Casbon said. PopGen Tech currently outsources all of its sequencing but is in the process of bringing a sequencer in house, though it has not yet decided which one.
[ pagebreak ]
As a proof of concept, the company has tested its method on 250 HapMap samples, for which genotyping data for common variants is available. It also identified new rare variants in these samples, which it validated by Sanger sequencing. A manuscript describing this validation study is currently under review, according to Lichtenstein. The company is also improving various aspects of the method, he said, but is not disclosing any details.
PopGenTech is now using the method in a number of population genetic studies, offering it as a service as well as through collaborations. These studies aim to discover associations between rare variants and clinical outcomes in order to allow development of clinical biomarkers to detect these variants, Lichtenstein explained, and to identify genetic pathways that are involved in drug toxicity and drug efficacy.
He noted that PopGenTech will share intellectual property of such discoveries with its collaborators, which provide the company with DNA samples and clinical data.
In a collaboration with a group at Cardiff University, for example, the company is currently studying predisposition to Alzheimer's disease, following up on a genome-wide association study that identified about 15 candidate genes for predisposition to late-onset Alzheimer's (IS 10/12/2010). Eventually, it will sequence these genes in 1,500 individuals.
Williams said that preliminary results from the first few hundred samples "are looking fine," and they hope to complete the study within the next few months. "We are pushing new territory, both for the design of the experiments and the technology that they have developed," she said.
In addition, PopGen Tech is involved in several cancer studies funded by pharmaceutical companies and by the UK's National Health Service, including studies of acute childhood lymphoblastic leukemia, clear cell renal cell carcinoma, and breast cancer.
Have topics you'd like to see covered in In Sequence? Contact the editor at jkarow [at] genomeweb [.] com.