One of the primary goals of genetics over the past decade has been to understand human health and disease in terms of differences in DNA from person to person. But even a relatively straightforward trait such as height has resisted attempts to reduce it to a particular combination of genes. In light of this shortcoming, some investigators see room for an increased focus on an alternative explanation for heritable traits: epigenetics, the molecular processes that control a gene’s potential to act. Evidence now suggests that epigenetics can lead to inherited forms of obesity and cancer.

The best-studied form of epigenetic regulation is methylation, the addition of clusters of atoms made of carbon and hydrogen (methyl groups) to DNA. Depending on where they are placed, methyl groups direct the cell to ignore any genes present in a stretch of DNA. During embryonic development, undifferentiated stem cells accumulate methyl groups and other epigenetic marks that funnel them into one of the three germ layers, each of which gives rise to a different set of adult tissues. In 2008 the National Institutes of Health launched the $190-million Roadmap Epigenomics Project with the goal of cataloguing the epigenetic marks in the major human cell types and tissues. The first results could come out later this year and confirm that different laboratories can get the same results from the same cells, says Arthur L. Beaudet of the Baylor College of Medicine, the project’s data hub. “One couldn’t automatically assume it would be so nice,” he says.

Up to this point, the best way to study epigenetic effects has been a strain of mice known as agouti viable yellow. In these mice, a retro­transposon—a bit of mobile DNA—has inserted itself in a gene that controls fur color. Mice bearing the identical gene can be yellow or brown depending on the number of methyl groups along the retrotransposon. Such methylation marks would normally be erased in the reproductive cells of an animal. But in 1999 a group led by geneticists at the University of Sydney in Australia discovered that methylation of the fur color genes persists in the female germ line, allowing it to be passed down to offspring like a change in the DNA.

Agouti viable yellow mice might have something to say about the human obesity epidemic. The animals have a tendency to overeat and become obese. In 2008 Robert A. Waterland, also at Baylor, discovered that this trait gets passed down and amplified from one generation of agouti to the next, so that “fatter mothers have fatter offspring,” he says. He is investigating whether the effect can be explained in terms of methylation patterns in the hypothalamus, the part of the brain that regulates appetite.

Retrotransposons could lead to other epigenetic effects. In the early 2000s geneticist David Martin of Children’s Hospital Oakland Research Institute in California reasoned that the silencing mechanism that keeps retrotransposons inactive might randomly shut down genes that are supposed to be left on. If the silencing occurred in a gene responsible for suppressing tumor formation, the result would appear the same as genetic mutations that predispose people to cancer.

Working with colleagues at St. Vincent’s Hospital in Sydney, Martin identified two individuals who had the characteristics of hereditary nonpolyposis colorectal cancer, which is usually caused by a mutation that inactivates one of a person’s two copies of the tumor suppressor gene MLH1, but who showed no signs of mutation. Instead the MLH1 of both was methylated in cells of the blood, hair follicles and inner cheek—all derived from different embryonic layers.

In Martin’s view, the result strongly suggested that the patients had inherited the silenced gene from one of their parents, like the case with agouti mice. Although some researchers have suggested that a genetic mutation in the fertilized egg cell could be responsible for the methylation pattern, Martin says the simplest explanation is an inherited epimutation. “Nobody has been able to explain why these things aren’t actually germ-line epimutations,” he says.

If epimutations can happen, the same effect should turn up in other genes. Martin’s colleague Catherine Suter of the Victor Chang Cardiac Research Institute in Sydney is studying whether melanoma patients have epimutations in genes associated with the cancer. It is also conceivable that epimutations could play a role in some cases of autism, Beaudet says.

Researchers agree they are just scratching the surface of understanding the role of epigenetics in health and disease. The NIH Roadmap Project should help by allowing them to compare models of disease with reference samples. In effect, “we’re trying to figure out how we work,” says epigenetics researcher Randy Jirtle of Duke University. “It’s an amazingly huge project, and it’ll never go away.”