The finding opens new research targets for better understanding the disorder and for treating it.
Research into the causes of autism has focused on genetic causes because so many families have multiple children with the disorder. Thus far, only about 10 percent of autism cases have a known genetic cause. Boston-area researchers estimate the gene glitch they've identified accounts for another 1 percent of cases.
They found a segment of a chromosome which has genes linked to brain development and various developmental disorders was either missing or duplicated far more often in autistic people. The defect was inherited in some cases, but more often the result of a random genetic accident.
The results from the Autism Consortium study, released online Wednesday by the New England Journal of Medicine, confirm those of smaller studies by U.S. and Canadian research groups in the past year. The consortium verified its findings by checking two other DNA databases.
"They really did nail it," said Dr. Andrew Zimmerman, director of the Kennedy Krieger Institute's Center for Autism & Related Disorders in Baltimore, who was not involved in the research.
He predicted children newly diagnosed with autism or other developmental disorders now will be tested for this defect on chromosome 16 and that studies of many more DNA samples may reveal other autism-related gene variations.
Already, the findings are starting to be used to give some parents long-sought answers to burning questions: What caused autism in their child and how likely is it that any future children also would have autism, long known to run in families?
"We've provided very compelling evidence that this particular small stretch of the genome provides an important clue to the biological roots of autism," said lead researcher Mark J. Daly, an assistant genetics professor at Harvard Medical School and an investigator for the consortium, which includes researchers from 14 Boston-area universities and medical centers.
When the biological pathways involved are figured out, scientists can try to design drugs to target chemicals in the brain to treat autism, said Geraldine Dawson, chief science officer of the advocacy group Autism Speaks.
"I think chromosome 16 is now going to be a hotbed for autism research," said Thomas Lehner, head of the genomic research branch at the National Institute of Mental Health. "It gives us a very important lead."
Another study researcher, Dr. David Miller of Children's Hospital Boston, said the chromosome 16 variations increased the risk of autism a hundredfold. But he said the disorder must be due to a combination of genetic variations since there were cases of people who had the defect but didn't have autism.
Autism, a complex, poorly understood disorder, is characterized by repetitive behaviors and poor social interaction and communication skills. Research has mainly centered on genetic causes, and on whether it could be caused by the mercury-based preservative once used in childhood vaccines, which has been repeatedly discounted.
The number of children diagnosed with autism has risen in recent years to as many as one in 150 American children, but experts are unsure whether its prevalence really is increasing or the trend is due to a broader definition of autism.
For their study, consortium researchers scanned all 46 chromosomes from DNA samples from 1,441 children with autism or related disorders. They also scanned DNA from most of their parents and 2,800 other people, none known to have autism.
The researchers found a 25-gene segment of chromosome 16 was missing in five children with autism; none of their parents had the deletion. That shows that in some cases the genetic glitch is not inherited from the parents, but instead due to a random accident while an egg or sperm is being formed.
Another seven autistic children had a chromosome 16 duplication, but all but one had parents with the same duplication.
The researchers confirmed their findings by looking at DNA databases from Children's Hospital Boston and Iceland. The same defect was found in 1 percent of those with autism or related disorders. It was found in just seven of about 19,000 Iceland samples from people without the disorder.
