Identifying the variations could help doctors intervene earlier.

Scientists plan to classify the symptoms of autism, which can vary dramatically. Credit: Keith Myers/Kansas City Star

Scientists are set to begin the autism 'phenome' project, which aims to classify the differences in autistic children. Identifying the molecular and anatomical variations of the disease might allow researchers to tailor treatments.

Autism is a developmental disorder marked by learning deficits and poor social skills. But autistic children also often show a host of other problems, including seizures and gastrointestinal disorders. Scientists have been unable to pinpoint the genetic or environmental causes of the disorder, in part because both appear to differ in subgroups of affected children.

The phenome project aims to classify the biomedical and behavioral aspects of the disorder. “The goal is to take autism as we know it and fractionate into more homogenous subtypes,” says David Amaral, research director of the MIND Institute at the University of California in Davis.

The goal is to take autism as we know it and fractionate into more homogenous subtypes. , . David Amaral, MIND Institute

A pilot version of the project—given high priority in the US National Institutes of Health's 2003 research roadmap—has already begun at the MIND Institute. Scientists plan to ultimately run rigorous genetic, proteomic, anatomic and immune tests on 1,800 children—600 with autism, 600 with other developmental disabilities and 600 normally developing children—at four US sites.

Autistic children have striking variations in proteins and immune cells in their blood, Amaral reported at a conference in Boston in May. He and his colleagues analyzed 4,000 proteins in 105 children —70 with autism and 35 without—and found significant differences in the levels of 100 proteins. They also found that autistic children had higher levels of B cells and natural killer cells. In a second study, immunologist Judy Van de Water found that immune cells collected from autistic children respond differently to bacteria and viruses than those from normally developing children.

Scientists plan to investigate these differences in the phenome project. Amaral also hopes to analyze the blood of young infants to find a predictive biomarker for autism.

The MIND Institute has raised more than $1 million to support the first 18 months of the project. Funds for the remaining costs—approximately $30 million—have not yet been lined up.

In the project's first phase, the researchers plan to follow 55 children for five to eight years after diagnosis. The children will be given a thorough medical checkup, including assessment of brain size, genetic studies to look for candidate genes and blood tests to find differences in protein composition.

The researchers plan to pool their data on a publicly accessible database. The project will help scientists interpret the genetics data collected by the autism community and study gene-environment interactions, says Dan Geschwind, a neuroscientist at the University of California at Los Angeles.

The researchers also plan to study which treatments work best for specific subtypes of the disorder, which could help doctors intervene earlier. “We have a window of opportunity for treatment,” says Amaral. “If you get it wrong, you can't go back.”