Knowing your genome may not do much to improve predictions about disease

Since scientists first mapped the human genome more than a decade ago, excitement has built about a new era of personalized medicine, in which people will ultimately be able to use the information embedded in their DNA to better prevent and treat disease.

But even as the technologies that decode genomes have gotten faster and cheaper, questions have remained about how useful that information will be to a doctor and a patient trying to understand risk for a complicated disease caused by a combination of genetic and environmental factors, such as diabetes or cancer.

A new study led by researchers at the Harvard School of Public Health adds to the emerging picture of how difficult it will be to use genetic information in the clinic to make meaningful predictions about risk for common diseases. The researchers studied common gene variations and environmental risk factors that have been linked to breast cancer, type 2 diabetes, and rheumatoid arthritis. They found that knowing those genetic variations and including possible interplay between them and the environment -- two genes that interact to increase risk more, for example -- a patient would get only a modest increase in risk prediction, of 1 to 3 percent at most

“Although there are many genetic mutations that have been found associated with risk of some diseases, their clinical utility is still hotly debated,” said Hugues Aschard, a research fellow in epidemiology at Harvard and lead author of the study, published in the American Journal of Human Genetics.


Some scientists argue that risk models underperform because they do not factor in the interplay between genes and environment, Aschard said, but his modelling indicated that even if such effects are common and scientists can figure them out, “the improvement in terms of prediction will not change much.”

Jeffrey Barrett, a group leader at the Wellcome Trust Sanger Institute, said that the study highlights another difficulty in turning ever-increasing amounts of genomic information into useful prognostic tools for common, complex diseases.


He added that the study does not undermine the utility of studying the interplay of genes and environment in order to better understand the biology of disease. Instead, the study illustrates that the idea of DNA as destiny -- a cheek swab at birth lays out a baby’s future struggles with diabetes or heart disease -- is extremely oversimplified.

“I think we have in the past overpromised a little bit: Personalized medicine is coming and it’s going to revolutionize everything,” Barrett said. “It may mean that it’s unlikely we’re ever going to use it in that ‘Gattaca’ future way, you have a DNA sequence and it will tell you everything about your future.”

Carolyn Y. Johnson can be reached at cjohnson@globe.com. Follow her on Twitter @carolynyjohnson.