Science Editor

In medical school, one of the things that students are taught is "when you hear hoofbeats, think horses, not zebras" – that is, common symptoms are most likely to be caused by common diseases. Or in practical terms, when you see a barfing child, think "stomach virus," not "brain tumor."

That link from the common to the common may not hold as far as the genetic causes of diseases are concerned, though. After sequencing nearly 30 individuals and looking at their genetic makeup, a team of researchers concluded that common diseases are most likely underpinned by many different rare genetic variants, not a few common ones.

The bottom line, Qianqian Zhu told BioWorld Today, is that "rare variants are more likely to occur in functional regions of the human genome and cause disease." Zhu is a researcher at Duke University's Center for Human Genome Variation, and the lead author of the paper on the findings, which was published in the March 31, 2011, issue of the American Journal of Human Genetics.

The work has implications for how natural selection works. But it may also provide part of the explanation for why the identification of risk variants has not done as much to further drug discovery as was once hoped.

If common variants are less likely to be in a functional region of the genome than rare ones, then that also means that they are less likely to cause disease – meaning that from the point of view of drug discovery, there may not be a lot of low-hanging fruit in genomics.

Indeed, senior author David Goldstein, director of the Center for Human Genome Variation, said in a prepared statement that "I am entirely convinced that sequencing, which is becoming less expensive every month, will unlock a lot of the causes of genetic disease. What we can do clinically with that information will become the primary challenge."

In their study, the authors sequenced the complete genomes of 29 individuals, a number that allows them to see most genetic variations that occur in at least 1 percent to 2 percent of the population. They then looked at all 3.5 million single-nucleotide polymorphisms (SNP) in those genomes.

The team, Zhu said, used "three major criteria" to separate functional from nonfunctional regions of the genome: First, how conserved it was, with more conserved genes more likely to be functional; second, whether they were in a gene itself, including an exon, intron or untranslated region; and third, whether it was in a region that was likely to have regulatory function.

They found that the more frequent a SNP was, the less likely it was to be in a functional region of the genome.

Goldstein said that "the magnitude of the effect is dramatic and is consistent across all frequencies of variants we looked at. . . . It's not just that the most rare variants are different from the most common, it's that at every increase in frequency, a variant is less and less likely to be found in a functional region of the DNA."

Usually, the more common variant of a SNP is also the older one, precisely because it is the older one. But in some cases, new variants can spread rapidly and become more frequent within a relatively short period of time, at least evolutionarily speaking.

To see whether such relatively new but common variants spread rapidly because they are useful, the authors also compared human genome sequences to those of chimpanzees and rhesus macaques to see whether recent but fairly frequent variants were under positive selection pressure.

Strikingly, when they focused on the variants that are new and common in humans, they found no relationship between how common a variant was, and whether it was in a functional or a nonfunctional region. Because nonfunctional variants, by definition, cannot confer an advantage on their owner, the results suggested that whatever the reason for the rapid spread of some variants, it was not that they did their owners any good.

The authors concluded from their results that that natural selection works more often via so-called "purifying" selection – that is, selection acting against changes that have a negative effect – than via positive selection, where a trait that is beneficial to its owner is more likely to make it into the next generation.

"Positive selection is not a general pattern," Zhu said. "Negative selection is more dominant."