Public Release: 

Important but limited role of de novo variants in autism spectrum disorders

Baylor College of Medicine

HOUSTON -- The search for gene mutations associated with autism and autism spectrum disorder is as complicated as the disorder itself.

In a study that appears online today in the journal Nature (, an international cohort of scientists that includes those at Baylor College of Medicine (, sequenced the exomes of 175 trios (a child with autism spectrum disorder and his/her two parents) and found many new (de novo) gene mutations but few that could be called a definitive cause of the disorder.

Autism and autism spectrum disorders encompass a wide definition of autism, incorporating developmental difficulties in verbal and non-verbal communication, problems in social interactions as well as differing responses to stimulation. Many people with autism have developmental delay while others appear to have superior abilities in specific fields such as music and math. The search for genes associated with this disorder has proven difficult with few single genes associated with the disorder except in the case of special syndromes such as Rett and fragile X.

After the researchers sequenced the exomes of the trios, they analyzed the sequences, looking for de-novo mutations and genome variants that occurred in the child but were not found in the parents. The sequencing took place at the Baylor College of Medicine Human Genome Sequencing Center (, the University of Pennsylvania, the Broad Institute of MIT and Harvard, Mount Sinai School of Medicine and Vanderbilt University.

"The rates of de novo mutations were only slightly elevated over what we would have expected to see, said Dr. Aniko Sabo, assistant professor in the Baylor College of Medicine Human Genome Sequencing Center ( ) and one of the authors of the study. "Almost all de novo mutations were found in different genes. Only three genes had two de novo variants. Furthermore, the breakdown of the types of mutation (those that potentially affect the function of the gene/protein and those that are silent) also followed expectations closely."

The group next focused their attention only on those mutations that were likely to disrupt the function of gene and the protein associated with it.

"In this subset of variants, we observed an increased protein-protein connectivity among the genes affected as well as greater connectivity to genes previously implicated as candidate genes in autism spectrum disorder," she said.

In the Nature report, the researchers said that their models "indicate that de novo SNV (single nucleotide variation) events will probably explain less than 5 percent of the overall variance in autism risk."

Two companion papers appear in the same Nature issue, and based on the combined data from all three reports, the researchers noted three genes with two de novo loss-of-function variants. Because of very low expected rate of this type of de- novo mutations, these are unlikely to occur by chance. Furthermore looking at the data from an ongoing exome study, two of the genes (KATNAL2 and CHD8) had additional loss-of-function variants in people with autism and none in the controls (people who do not have autism),implicating them as autism risk factor genes.

"Overall, these data underscore the challenge of establishing individual genes as conclusive risk factors for ASD (autism spectrum disorder), a challenge that will require larger sample sizes and probably, deeper analytical integration with inherited variation," the researchers wrote.

"Many studies are pointing to extreme genetic heterogeneity of autism disorders, estimating hundreds of autism risk genes to be involved, and multiple genes contributing to the disease in each person. But the hope and expectation is that these genes will converge on a smaller number of molecular pathways." said Sabo. "The next step is to increase the numbers of people in our studies, develop better understanding of effects of particular variants and integrate inherited variants, and de novo mutations in our analyses."


Institutions taking part in this work include BCM, The University of Texas Health Science Center at Houston, the Broad Institute, Mount Sinai, University of Illinois at Chicago, University of Pittsburgh School of Medicine, Carnegie Mellon University, Perelman School of Medicine at the University of Pennsylvania and Vanderbilt University.

Researchers from BCM who took part include Uma Nagaswamy, Donna Muzny, Jeffrey G. Reid, Irene Newsham, Yuanqing Wu, Lora Lewis and Yi Han. Drs. Richard Gibbs, director of the BCM Human Genome Sequencing Center, and Eric Boerwinkle, co-director of the HGSC and professor at the UT School of Public Health were principal investigators on the project.

Funding for this work came from the American Recovery & Reinvestment Act (ARRA), the National Institutes of Health, the Seaver Foundation, the NIH National Center for Research Resources, and the Kennedy Center for Research on Human Development.

For more information on basic science research at Baylor College of Medicine, please go

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.