In the past five years, scientists have identified more than 3,000 common genetic mutations associated with diseases including cancer, Alzheimer's and diabetes, thanks to insights gleaned from genome-wide association studies (GWASs). But the inherent value of these studies has come under scrutiny, in part because they largely ignore rare mutations. Given this flaw, researchers have called for renewed focus on the rare mutations that might be more likely than common ones to cause illness.

The GWAS approach involves comparing the genomes of healthy people with those suffering from illness to pinpoint disease-associated single nucleotide polymorphisms (SNPs) typically present in at least 5% of the population. “A GWAS study is meant to capture most of the common variation in the genome, and that's something it does very well,” says Jonathan Sebat, a geneticist at the University of California–San Diego.

So far, though, the common variants that have been identified by GWASs confer relatively small increases in risk and explain only a fraction of the heredity that clearly exists in many common diseases.

In an analysis published last month in the journal Clinical Genetics (doi:10.1111/j.1399-0004.2010.01535.x, 2010), medical geneticist Ivan Gorlov and his colleagues from the MD Anderson Cancer Center in Houston argued that scientists should speed the search for rare mutations that slip under the radar. “Until now, we've studied two tales of the distributions of polymorphisms: extremely rare mutations with strong effects, or monogenic diseases, and common polymorphisms with small effects on disease risk,” explains Gorlov. “Common sense tells us that the most cases should be in the middle.”

According to the new study, more than half of all SNPs are probably rare in terms of their prevalence in the population, and these are the ones that are most likely to cause disease. “They are likely to be slightly deleterious,” Gorlov says, “and because of this, they will be under the pressure of negative selection that will drive their frequencies down.”

As a result, Gorlov contends, scientists conducting GWASs need to maximize their sample sizes so that rarer mutations can be assessed. Others agree. “The bigger your sample size is, the more things you find,” says John Witte, a genetic epidemiologist at the University of California–San Francisco.

Others, such as Sebat, believe that more drastic measures will ultimately be necessary. “If you really want to understand how much rare variation contributes to disease, what you need are complete genome sequences,” Sebat says.