Speed-Reading the Genome

Reading genomes is a messy business. Even the terminology—like "shotgun"—evokes images of inelegant science. But Woburn, Mass.-based US Genomics plans to change that. Inventor Eugene Chan based the GeneEngine™ on the same mechanisms cells use to read DNA. He designed a system in which DNA is first linearized and then threaded through a nanofluidic chip at high speeds. Before the analysis, the DNA sample is treated with a set of fluorescently labeled tetramers that cover the thr

May 13, 2002
Jim Kling
Reading genomes is a messy business. Even the terminology—like "shotgun"—evokes images of inelegant science. But Woburn, Mass.-based US Genomics plans to change that. Inventor Eugene Chan based the GeneEngine™ on the same mechanisms cells use to read DNA. He designed a system in which DNA is first linearized and then threaded through a nanofluidic chip at high speeds. Before the analysis, the DNA sample is treated with a set of fluorescently labeled tetramers that cover the three million or so known human single nucleotide polymorphisms (SNPs). "What we recognize is sets of sequences as opposed to the base pair," says Chan, the company's chairman and CEO.

The result is a device with the potential to analyze a single, long strand of DNA. That means far less handling of DNA and therefore fewer problems with contamination and loss. Chan claims that it is also faster than existing systems; it analyzes 10 to 30 million base pairs per second. And best of all, he says, the system may not require polymerase chain reaction (PCR).

The US Patent and Trademark Office recently awarded US Genomics patent No. 6,355,420, "Methods and products for analyzing polymers," to cover its GeneEngine technology. Though the system isn't commercially available yet, the company has entered into collaborations with the Wellcome Trust Sanger Institute and Pui-Yan Kwok, dermatology professor at the University of California, San Francisco, to validate its claims and enhance further development.

The GeneEngine is a potentially powerful tool for haplotype analysis, says Kwok. In haplotype analysis, scientists correlate inheritance of a disease gene with various markers. "It's really difficult to know which alleles of individual markers are on the same chromosome," he says, because somatic human cells—like those of all eukaryotes—are diploid. Yet researchers cannot identify haplotypes without this information.

Currently, the only way to determine which alleles are grouped together on the same chromosome is to rely on calculations that depend on inheritance patterns, or to create somatic cell lines, separating the paired chromosomes and placing them in separate cell lines for analysis. Both methods are laborious and not precise, says Kwok. The process would be much easier, however, if all the alleles could be sequenced in a single run of DNA sample, as US Genomics claims. "We're hoping that if the claim of being able to look at single molecules at really fast speed [is true], we can label the alleles and see which ones go together."

The advantages wouldn't stop there. Single molecule detection would allow researchers to group patients' samples together. "That's one of the attractive things about this technology, because now you can do two experiments and get all of the information you need for a lot of these disease genes; ... you pool all of the affected individuals in one and all of the controls in the other."

Over the next several months, Kwok and his team will focus on improving the DNA labeling, and testing GeneEngine to determine the maximum length of DNA it can handle. Although he is excited about GeneEngine's potential, Kwok is not convinced it will meet all of the company's claims. "I don't know if [it will work] without PCR," he says, "but even if we have to do PCR it is still a lot faster than any other method that I know of. Actually, there's no competitor in terms of looking at haplotypes."

US Genomics has plans to license GeneEngine to pharmaceutical and biotech companies in the near future, but the technology is probably still several years away from being more generally available, according to Chan. In time, the company hopes to see the GeneEngine applied to sequencing as well as analysis. Many areas of the genome are currently difficult to sequence because they can't be cloned and therefore can't be expanded by PCR. Other areas are so riddled with repeats that they are difficult or impossible to patch back together after sequencing via traditional methods. The GeneEngine could open these regions to sequence detection.

Other researchers are developing competing approaches to linear DNA analysis, according to Kwok. US Genomics is getting a lot of attention, in part because "they do a good job explaining and educating people about their vision," he says. But like other technologies that have shown promise, GeneEngine is not likely to be the final answer to genome sequencing and analysis. "Mass spectroscopy was touted as a way to sequence DNA ... and now it has found a really good niche in analyzing SNPs," says Kwok. "I would say [US Genomics'] vision over the next few years will probably lead to good haplotyping and genotyping applications, but in terms of sequencing whole genomes, we're still three or four breakthroughs away."

Jim Kling (jkling67@cs.com) is a freelance writer in Washington, DC.