One of the frustrations of scientists researching autism and families struggling with it is that genetic studies in the past have found only small pieces of the puzzle affecting handfuls of relatives.

But now, a new study that appeared last month in the online version of the journal Nature says that researchers have found the first genetic variation that is common in a wide swath of people with autism -- one that may account for 15 percent of total cases.

That study and another one that appeared alongside it were led by Hakon Hakonarson, director of the Center for Applied Genomics at Children's Hospital of Philadelphia, using research done at several institutions across the nation, including the University of Pittsburgh.

Autism is estimated to occur in as many as 1 out of every 150 children. The disorder affects four times as many boys as girls, and is characterized by poor social skills, repetitive behavior and narrow interests.

Dr. Hakonarson said in an interview last week that one of the new study's most exciting results is that the genetic variation that stood out in people with autism lies between two genes that control the connections that brain cells make with each other.

That finding bolsters brain imaging work that was pioneered at Carnegie Mellon University and the University of Pittsburgh, which showed that adults with autism had poorer connections between different regions of their brains than people without the disorder did.

"The work that [Carnegie Mellon's] Marcel Just and [Pitt's] Nancy Minshew have done showing this lack of connectivity has really held up," said Geraldine Dawson, the chief science officer of Autism Speaks, the nation's leading autism advocacy group.

"It looks like autism is indeed a syndrome where there is a lack of coordinated functioning between different regions of the brain," said Dr. Dawson, who is a co-author of the Nature papers.

The lead study in Nature looked at "snips," the moniker given to single nucleotide polymorphisms, which are the variations we all have in the chemical letters of our genetic code.

Researchers looked at nearly 600,000 snips in people with autism and control groups that didn't have the disorder, and found that one region on Chromosome 5 stood out.

The segment lies between genes that make cell adhesion molecules, which are crucial in forming the synapses that connect one brain cell to another.

Interestingly, the variant they found exists in about 55 percent of the general population, but was present in 65 percent of the people with autism, which was enough to be statistically significant, Dr. Hakonarson said.

As opposed to many previous autism genetic studies, this one involved a large sample of people -- 5,500 people with autism, 1,500 unaffected relatives and 6,500 control-group people who did not have autism in their families.

The samples came from the Autism Genetic Resource Exchange run by Autism Speaks, the Autism Genome Project headed by Dr. Hakonarson, as well as samples assembled at the University of Washington and UCLA, Dr. Dawson said.

Because more and more families affected by autism are volunteering genetic material for research, studies like these that include enough people to display strong statistical power will become increasingly common, she predicted.

"The sheer number and the way the statistics were done make the finding very reliable," she said. "When you have these very large samples, it makes you very confident."

The second study done by the Hakonarson team looked at "copy number variants," another common genetic pattern.

We typically inherit one copy of each gene from our mothers and fathers, but it's not unusual for us to have some extra copies, some missing copies and some smaller additions and deletions involving parts of genes.

In that analysis, Dr. Hakonarson said his team found at least nine regions that were unusual in people with autism. One of those areas involved genes that make a product called ubiquitin.

Ubiquitin also seems to be important in how brain cells connect, but its key role may involve pruning unneeded or extra connections, he said.

That could provide an intriguing link to other autism research, which has shown that children with the disorder tend to go through this pruning process at a later age than children without autism.

Bernard Devlin, a computational genetics professor at Pitt who studies autism, said the new research is promising, but the results need to be replicated in an even larger group of people with the disorder.

"The way I think of the genome-wide significance of these studies is that we're in the first quarter of the football game, and what the final score will be we can't say. Whether this 'snip' [in the Hakonarson study] is really important depends upon other data sets that are still coming in."

On the other hand, Dr. Devlin said the results are promising, and there will soon be another study published that will identify a different genetic variation in people with autism, but one that is also near genes that affect brain cell development.

The next step in his research, Dr. Hakonarson said, will be to develop laboratory animal models of these genetic variations. Scientists will then be able to see how the variations affect the animals' behavior and figure out how the proteins made by the genes differ.

And that ultimately should allow them to develop medications that could improve the brain connections in people with autism and lessen or reverse the disorder.

"This is something that is a very doable thing," he said, "but it will take time. I would say give us another six to 10 years, and we will have something that can target this mutation and fix the consequences of it."