Autism spectrum disorders (ASD) include everything from the relatively mild Asperger syndrome (characterized by more mild social and language impairments) to full-fledged autism (characterized by severe social and communicative handicaps, limited interests, and repetitive behaviors). ASD is relatively common, and in the United States it is estimated that one out of every 54 boys is affected (the frequency in girls is considerably lower, averaging one affected girl out of every 252). Although genetics certainly plays a role in these conditions, exactly how is not fully understood, and this newest study, led by Richard Anney (Trinity College of Dublin) and Bernie Devlin (University of Pittsburgh School of Medicine), is the next a long line of scientific inquiries to this point. Their study, known as the Autism Genome Project (AGP), was conducted in two stages. The first stage of consisted of a genome-wide association study using genetic data from 1400 families affected by autism; the second stage checked the associations discovered in the first stage using the genetic data from an additional 1301 ASD-affected families and included another new genome-wide association study which combined the study subjects from both stages. When all the analyses were said and done, no SNPs (common genetic variations) were significantly associated with ASD. Furthermore, when some of the SNPs that had been identified in the first study as possibly associated with ASD were tested in the second-stage families, the associations failed to hold up. This lack of common SNPs associated with ASD is both disappointing and enlightening. Knowledge of what is not true, paradoxically, is knowledge of what is true. For instance, if I tell you that my pet Tyger is not a dog, you are one step closer to knowing Tyger’s a cat. Most of science progresses through “not trues” — the failed hypotheses that bring us closer to real understanding. A perfect example of this mode of scientific progress is this recent genetic study. Their lack of findings was quite a finding. The fact that no SNPs were associated with ASD in such a large study suggests that the effect of any one common SNP is quite small. Even when all the SNPs were used together to predict whether a study subject would have ASD, the model explained less than 1% of differences in risk for the condition. The fact that these common SNPs have little power to predict ASD does not mean that genetics is unimportant. Twin studies suggest that ASD is at least moderately heritable, suggesting that genetics does play a role. This current study’s lack of findings supports the idea that common genetic variations may play a smaller role than rare mutations or copy number variation (the number of copies of a given gene) in ASD risk. Thus, the lack of ASD associations today has, hopefully, brought us closer to tomorrow’s discoveries.
Editor’s note: Pending an FDA decision, 23andMe no longer offers new customers access to health reports referred to in this post. Customers who purchased prior to November 22, 2013 will still be able to see their health reports, but those who purchased after that time will not. Those customers will have access to ancestry information as well as access to their uninterpreted raw data.