Findings Hint at Genetic Links in Autism

by Tom Valeo

March, 2008

Several studies published almost simultaneously in three journals have identified genes linked to autism, underscoring the power of new techniques for scanning the human genome.

These techniques enable researchers to scan the common points of variation in all 3 billion bases in the human genome and discover potentially troublesome deletions, replications or changes of sequence in individual base pairs. As a result, genotyping that until recently would have taken two or three years now takes two or three hours, helping researchers identify variations long suspected to be at the root of disorders such as autism.

“This is utterly revolutionary,” says Tom Insel, Director of the National Institute of Mental Health. “Not only can we detect common variants faster, cheaper and better than in the past, but with new sequencing technologies we can quickly read each letter in the text of a DNA sequence. Even a rare event that occurs in only a few children with autism can now be detected by sequencing regions of interest.”

In addition, researchers have access to enormous DNA repositories collected from people with autism. A study in the Feb. 14 New England Journal of Medicine, for example, used DNA from 751 families in which at least one person has autism. All of the families were enrolled in the Autism Genetic Resource Exchange, the world’s first gene bank for autism. Their DNA was compared with DNA from thousands of people who do not have autism.

Without such a large sample, this type of research would not have enough power to reveal these very subtle differences in the DNA of people with autism, Insel says: “It’s hard to distinguish signal from noise with a million differences, but if you have several thousand subjects you can do that.”

These new methods of scanning the human genome not only find trouble spots that cause autism; they also are helping to break autism down into subcategories based on genetic variation. Some researchers say genetic testing may even provide a definitive and detailed diagnosis of the disorder.

“I think there will be some regions of the genome that, when deleted, will be highly predictive of outcomes of autism spectrum disorder,” says Stephen Scherer, director of the Center for Applied Genomics at the Hospital for Sick Children, a professor of molecular and medical genetics at the University of Toronto and the lead author of an article in the Feb. 8 American Journal of Human Genetics on chromosome variation in autism. “From our current data it is already apparent that for a proportion of individuals, it will be possible to describe their autism spectrum disorder based on the underlying structural characteristics of their genome.”

This new knowledge also may lead to effective treatments for autism, Insel says. “We don’t have medical treatments for the core symptoms because we don’t know the pathophysiology of autism,” he says.

“We don’t know what has gone wrong, but genetics may help us to get there.” Tom Valeo is a medical writer based in St. Petersburg, Fla. He can be reached at