Main

Researchers have long suspected the importance of other variations such as insertions, deletions and inversions in our genome. These so-called structural variations, however, are more difficult to detect than SNPs. Convinced that crucial information is being overlooked, Evan Eichler and his group at the University of Washington in Seattle, set out to discover structural variations in the genome of one individual, using the human genome sequence as a reference.

They used a large fosmid library, built from the DNA of a test individual, and mapped where the two ends of each clone appear in the human genome reference. Because the size of the clones in a fosmid library is very consistent (about 40 kilobases; kb), the two paired ends most often map to sequences separated by the same distance—but the interest is in the exceptions. A pair of fosmid ends aligning to a pair of sequences separated by more than 40 kb in the reference reveals a deletion in the test individual. Conversely, a separation of less than 40 kb signals an insertion, and an inconsistency in their orientation indicates an inversion.

The study, published in Nature Genetics, revealed about 300 sites of structural variation. Once a structural variation has been identified, sequencing the corresponding clone is easy; this reveals the exact nature of the underlying structural change, and permits the design of genotyping tests to evaluate its frequency in a population.

The problem is the price tag. Analysis of each individual amounts to sequencing about one million paired ends, “a mini-genome project” as Eichler calls it. But this is not a one-time shot. His laboratory, in collaboration with Agencourt, University of Washington Genome Center and US National Human Genome Research Institute, is embarking on the analysis of another eight to ten individuals, selected from the HapMap collection to represent extremes of genetic diversity, with the goal of building a more global structural map of the human genome.

Undoubtedly, these data will broaden the human genome consensus and drive new genetic studies. “The key will be moving forward with the development of robust genotyping assays to do association studies with diseases,” says Eichler. Although his lab has established PCR-based genotyping tests for some of the variation sites, he is not convinced that this is the way to go. Instead, he insists, “we need new technology development to be able to detect inversions in a high-throughput and robust fashion.”

“Think about where SNPs were 10 years ago,” says Eichler, “in discovery phase; this is where we are now with structural variations”. Eichler is confident in the power of the method, in part because it relies on a skill that genome centers already have mastered: sequencing libraries of millions of clones. In his view, “combining lots of sequencing in a random fashion with structural information is the way of the future.”