New Genetic 'Fishing Net' Harvests Elusive Autism Gene
DURHAM, N.C. -- Duke University Medical Center researchers have developed a new statistical genetic "fishing net" that they have cast into a sea of complex genetic data on autistic children to harvest an elusive autism gene.
Moreover, the researchers said that the success of the approach will be broadly applicable to studying genetic risk factors for other complex genetic diseases, such as hypertension, diabetes and multiple sclerosis.
In this case, the gene, which encodes part of a brain neurotransmitter docking station called the gamma-Aminobutyric Acid Receptor beta3-subunit (GABRB3), has been implicated in autism previously, but never positively linked to the disease. Their findings will be published in the March 2003 issue of the American Journal of Human Genetics.
"Many research groups have been actively looking for genetic risk factors that can lead to autism, but without much success," said Margaret Pericak-Vance, Ph.D., director of the Duke Center for Human Genetics and lead investigator of the study.
Autism is the common term that encompasses an overlapping group of complex developmental disorders that are diagnosed in about one in 1,000 children under the age of 3. Each autistic child has a unique set of characteristics that affect his or her behavior, communication skills and ability to interact with others. It is the very diverse, complex nature of autism that has made it so difficult to locate distinct genetic risk factors, said Pericak-Vance.
After several genetic studies turned up only a few vague genetic clues, the research team decided a new approach was needed. Pericak-Vance hypothesized that grouping patients with similar traits together statistically might enhance the scientists' ability to distinguish relevant genetic risk factors. To provide guidance, the scientists turned to Michael Cuccaro, Ph.D., a clinical child psychologist at Duke with extensive experience diagnosing and treating autism. Cuccaro noticed that some but not all autistic children exhibit repetitive compulsions and extreme difficulty with changes to their daily routine. This character trait -- defined by Cuccaro as "insistence on sameness" or "IS" -- helped the research team identify a subset of autism family data to study in more detail.
Researchers, led by Yujun Shao, Ph.D., a genetic epidemiologist at Duke, reorganized data collected from families in which more than one child is affected by autism and grouped together all the families that reported their autistic child had difficulty with change.
Cuccaro's theory that autistic children could be subdivided into at least two groups gave the team of scientists from Duke and the University of South Carolina an opportunity to test a new statistical method, called "ordered subset analysis," developed by Elizabeth Hauser, Ph.D., assistant research professor of medicine at Duke. This new genetic fishing net allows scientists to sift through complex genetic data and extract genetic risk factors that affect only some of the total group.
In this case, when the researchers applied the new test only to those families whose children scored high in the IS category, they discovered a strong link to the GABRB3 gene on chromosome 15q, where no such link had appeared before.
"This is the first successful application of ordered subset analysis to help us pinpoint a genetic risk factor that would be missed by looking at the larger group." said Pericak-Vance.
The researchers emphasize that this discovery is only the first step in understanding how the GABRB3 gene, or others genes in the same region of chromosome 15 might be involved in autism. Another clue may be gained from previous research that has shown the same area on chromosome 15 is just as responsible for Angelman Syndrome and Prader-Willi Syndrome -- two genetic disorders in which a subset of affected children also exhibit repetitive behavior. Additional research will be necessary to understand how defects in the GABRB3 gene might contribute to autistic disorder, and how other genes or environmental factors also play a role.
"In the short term, however, I think what this will allow us to do is encourage clinicians and researchers working with autistic children to think about autism as consisting of different types or subgroups and not a one-dimensional disorder," said Cuccaro. "I think that subgrouping, over time, will allow us to develop a better understanding of how to treat each individual with autism."
This is a case, said Cuccaro, where identifying subsets of patients based on clinical observations has resulted in a significant neurobiological finding, and it perhaps is pointing a way to bring clinical observations to bear on complex genetic problems.
"The genomic revolution has given us a tremendous wealth of information in terms of a road map and markers for finding disease genes," said Pericak-Vance. "Now, we need to be able to look at complex clinical information and come up with methods that can help us dissect diseases that have multiple risk factors. This new statistical test will allow us to find meaningful genetic risk factors that are diluted out when tested as part of a larger heterogeneous group."
Members of the research team also included Marissa Menold, Chantelle Wolpert, Leigh Elston, Karen Decena, Shannon Donnelly, Robert DeLong, M.D., and John Gilbert, Ph.D., of Duke; and Sarah Ravan, Ruth Abramson and Harry Wright, M.D., of the W.S. Hall Psychiatric Institute at the University of South Carolina. The research was supported by grants from the National Institutes of Health and the National Alliance of Autism Research.