Skip to main content
Many of the most severe mutations identified in patients with autism affected proteins that work together in one large interconnected network.
Investigator, University of Washington
Many of the most severe mutations identified in patients with autism affected proteins that work together in one large interconnected network.


Over the past decade, scientists have added many gene mutations to the list of potential risk factors for autism spectrum disorders—but researchers still lack a definitive explanation of autism’s cause. Now, a chance finding from the University of Washington’s Jay Shendure and Howard Hughes Medical Institute (HHMI) investigator Evan Eichler gives an important clue to the puzzle.

In their study of more than 200 families, 39 percent of the 125 most severe mutations identified in patients with autism affected proteins that work together in one large interconnected network. Members of this network control a variety of fundamental developmental processes, such as whether a gene is turned on or off by changing the large-scale packaging of DNA, a process called chromatin remodeling.

Eichler’s team published its results on April 4, 2012, in the journal Nature. Two other research teams, one at Yale Medical School that included Matthew State and HHMI investigator Richard Lifton, and a second led by Mark Daly at Harvard Medical School, also published papers in the same issue of Nature that provide new insights into the genetic underpinnings of autism spectrum disorders.

“I think we’ve stumbled on a network that could be really important,” says Evan Eichler. “Other studies have hit on some of these genes one at a time, but I was surprised to see so many of them related as part of a highly interconnected set of proteins.”

Autism spectrum disorders, a set neurodevelopmental disorders characterized by impaired communication and social skills, affects more than one percent of children in the U.S., according to the latest numbers released by the Centers for Disease Control and Prevention. Previous studies on autism have concluded that genetic mutations present at birth are responsible for the disorder, but few cases have been explained in full by specific genetic mutations. “There is good evidence that mutations in as many as two dozen different genes can contribute to autism, but hundreds of other genes have also been suggested to be involved,” Eichler noted.

In their latest search for gene mutations linked to autism, Eichler and his colleagues relied on the Simons Simplex Collection, a collection of genetic samples from families in which only one member has the disorder. The presumption is that in these cases there will be enrichment for sporadic mutations in one of the parents’ egg or sperm cells that causes autism, rather than a mutation that a parent carries in all the cells of his or her body. “These sporadic mutations might explain about 25 percent of autism cases,” Eichler says, “although that’s just a guess.” By comparing the genetics of the parents, who don’t have autism, and the child who does, researchers can spot genetic differences and narrow down the list of mutations that could be responsible for the disorder in that individual.

Eichler’s team looked at the exomes—the full set of all protein-coding genes—from 209 such families and found more than 250 mutations carried by autistic children but not by their parents. Many of these were likely benign, but some of the mutations had been linked to autism or related genetic disorders in the past and others were strong new candidates. Only a handful were seen in more than one patient, but when the researchers started sorting through the list, Brian O’Roak, a postdoctoral fellow in Eichler’s lab, noticed that many were mutations affecting proteins in a related network.

“At first, I was skeptical that this was important,” says Eichler. “You can find connections between all sorts of proteins if you look hard enough. But we went back and sequenced the exomes of 50 unaffected siblings.” None of the unaffected siblings had mutations in the network.

Thirty-nine percent of the mutations on Eichler’s original list were mutations in genes that coded for proteins in the network, which he calls the βcatenin/chromatin-remodeling protein network. “Understanding how the mutations affect cells will require further experiments,” says Eichler, “but converging on a network is a step forward in the field of autism genetics.”

“If a list of genes just keeps growing and growing, it becomes very daunting and becomes harder to make any sense of,” he says. “But here we have converged on a pathway that helps to make some sense of that list.”

In addition to highlighting the network’s potential role in autism, Eichler’s team also studied whether the sporadic mutations they found in children came from mutations in a father’s DNA or a mother’s DNA. Eighty percent of the new mutations, they found—both those linked to autism and those seen in unaffected siblings—came from a father’s DNA. The high level of sporadic mutations inherited from fathers could help explain why autism has been associated with older dads.

“There are many mutations outside the network that are still linked to autism, and many mutations still to be discovered. Future genetic screens—that include regulatory DNA in addition to protein-coding sequences, for example—will likely reveal these additional mutations,” Eichler says.

“Despite finding this link between many mutations, I think autism is really an umbrella term for many disorders,” he says. “There are probably going to turn out to be many molecular flavors of autism which might be distinguished only at the level of genotype.”