Needles emerge from the genomic haystack

double-helix-genomegov.jpgScientists yesterday reported finding seven new gene variants linked to diabetes, a sign that the disease-gene hunt may have finally attained a sort of critical momentum.

Researchers have been looking for genes related to particular disorders for more than 30 years, and have been frustrated in that hunt for almost as long. With the exception of a few conditions caused by a single malfunctioning gene — most notably, perhaps, cystic fibrosis and Huntingdon’s disease — most disease appears to result from a complex interplay of multiple genetic effects and environmental factors. Even worse, until recently most studies that tried to untangle the genetic side of the equation operated by laborious trial-and-error — researchers would study families or other groups of people with a particular disease, then apply various biochemical and mathematical tools to try to identify which portions of the genome they shared, in the hopes that they’d eventually narrow the search down to a particular gene. The process produced so many unproven — and largely inaccurate — “candidate” genes that biologists began to joke about “gene of the week” discoveries.

One of the major changes in recent years has been the advent of a new way of sifting through genetic data. Instead of sequencing entire genomes — or even chunks of genomes — researchers have started restricting their searches to the specific ways in which one individual’s genes differ from those of other people. Think of it this way: Sequencing a human genome involves reading all three billion nucleotide “letters” of it, the overwhelming majority of which are the same from person to person. But if you set two — or eight, or 500 — genomes next to one another, the places where individual “letters” vary will jump out. Researchers have now mapped 10 million or so of these genetic variations, which are technically known as single-nucleotide polymorphisms, or SNPs (pronounced “snips”), and compiled them into a database known as the HapMap. (One private biotech, Perlegen Sciences, donated at least 2.1 million SNPs it had discovered to the project.)

In other words, sorting through SNPs is a lot faster than trolling through entire genomes, which in turn makes it possible to scan genetic information from much larger groups of people with a particular disease. In practice, these techniques may finally be turning disease-association studies into more of a science than an art.

Unfortunately, that doesn’t mean that medical breakthroughs are likely to flow immediately from the resulting gene-disease links. The latest diabetes findings raise the number of genes that contribute to the disease to ten — but those ten genes only account for two to 20 percent of overall diabetes risk. What’s more, it’s not entirely clear what these gene variants do differently that might contribute to the disease, much less how modern medicine might go about preventing or repairing the damage. So there’s plenty of slogging yet to be done.

The NYT’s Nicolas Wade has more.