Perlegen scientists are cataloging the millions of SNPs (single nucleotide polymorphisms) that distinguish human beings from one another. An individual's DNA is amplified to enable it to be washed over a five-inch-square Affymetrix microarray wafer. The DNA is labeled with a fluorescent tag before it is bathed over the wafer containing millions of short oligonucleotide strands of predefined sequence. At each point across the wafer, the DNA sample will bind where it finds a complete match, the location revealed by fluorescence. If there is a mismatch, such as when a SNP is present, the probe will wash off and the site will remain dark.

From this point, the image is cleaned up and examined to determine the intensity of the 60 million different areas available on the wafer. Special algorithms that measure the intensity of fluorescence across the feature help determine which base (A, C, G, or T) is in that location, and whether the test material matches.

By examining these SNPs, Perlegen can make predictions about the genome in other populations and discover where differences may lead to greater understanding of the relationship between disease and the pharmaceuticals that treat them. Chief Scientific Officer David Cox says that working with these variants is the most important part of Perlegen's work. "[We can work with the variants] in real time and for real money," he says, explaining that being able to work with variants across a population in a cost-effective manner unlocks the door to understanding the 20 to 80 regions of the genome that are causal for a given disease or condition.

The company has analyzed the genomes of 25 individuals to date, with plans to use this information to do association studies in thousands of people. On August 16, Perlegen completed the first comprehensive human haplotype map. A separate public effort (the "HapMap") is being coordinated by the National Institutes of Health.

—Jennifer Patterson Lorenzetti 

Back to Taking Data Storage to Infinity—and Beyond