“Autism is not really one thing,” explained geneticist Brian O’Roak, in describing why, despite evidence of a genetic basis for autism, there’s yet no “autism gene” fingered as the culprit. O’Roak is a postdoctoral fellow with the Eichler and Shendure labs at the University of Washington, and his research involves a good deal of high-tech detective work.

His latest paper, reporting new findings from exome sequencing, was published in the May 2011 issue of Nature Genetics. They were testing the hypothesis that sporadic, protein-altering mutations were linked to sporadic cases of autism, and the initial results (from a small sample) are promising–they may have found a good way to begin to map out a spectrum of genes involved in producing the autism spectrum.

O’Roak and I were sitting on the patio of the Vista Café, on UW campus, at the UW’s $150 million genomic sciences headquarters, the William H. Foege Building. Everything is new and gleaming, and even the café furniture boasts shiny metal finishes that give you the impression they’re suitable for al fresco lab work.

O’Roak is soft-spoken and mutton-chop-sideburned, and unfailingly courteous. Before we begin the interview, he cautions me to remember the deep, personal impacts of autism for many, and not mislead with a headline that promises a cure within the year. (As it happens, I’m aware of the first part, having an older brother who is autistic.)

The results of his latest work, while giving his team the impression they’re on the right track, are still the results of a pilot study. They’re in the midst of a larger effort to confirm their initial findings.

Technological advances in gene sequencing are both heartening–“For about a year now, this technology has been available where you can actually look at the entire genome through the exome, and do it in an unbiased way”–and humbling: The joint forces of genetic heterogeneity and phenotypic heterogeneity create universes of possible outcomes in any given individual. That is, there are genes, and then there’s what the genes do, and their interrelations are very, very complicated. At the moment, it’s estimated that, of the genetic factors that cause autism, some 70 percent of those remain unknown.

To tease apart some of those interactions, O’Roak, et al, turned to the Simons Simplex Collection for help. Sequencing just the exome–the protein-coding part of the genome–made that part of the task easier. “Jay Shendure, who’s the co-senior author, his lab developed one of the approaches to doing exome sequencing, basically about a year before I got here,” said O’Roak. But the researchers still needed a way to better sift mutations that might be causative for autism from mutations that aren’t. The hunch was that sporadic cases of autism could be the key, and, thanks to the Simons Simplex Collection, that key was now available:

For about three years now they have been collecting [genetic samples from] families throughout the U.S. They have twelve different sites. Their whole mission has been to develop this public repository where any researcher can apply and get samples; but what they’re doing that’s different is getting these simplex families (or single affected families, or sporadic families, as we call them in the paper).

O’Roak’s study took “20 trios–so that’s father, mother, child–where the affected individual looked like the only person affected in their family. We only really expected, based on what we knew about mutation rates, about one new mutation per individual. So we could quickly get down to only a handful of variants to sift through, as opposed to the hundreds of variants you might find in any family.”

The de novo mutations they uncovered begin to give a better picture of what’s involved in autism. Mutation at GRIN2B suggested impairment of glutamatergic neurotransmission (glutamate is extremely important for learning); a mutation at SCN1A brought in a gene previously associated with epilepsy; disturbance to LAMC3 would affect limbic and cortical development; mutations at FOXP1 often results in language difficulties.

Looking at the data, O’Roak was struck by the rareness of the mutation sites (of the 21 mutations found, eleven were protein-altering, and four of the mutations, at FOXP1, GRIN2B, SCN1A and LAMC3 were found in severely affected individuals).

It meant they could be on to something–that they had found a way to work backwards from autism’s expression to specific genetic points of interest. If the next, larger study bears this first look out, it’s a step toward a future in which gene therapy could target these sites. But even earlier, this kind of information could impact intervention techniques, as we learn precisely what areas have been affected, and in what way.