We report a genetic association between social communication spectrum phenotypes and a high-risk autism spectrum disorder locus (8
) in members of a large U.K. birth cohort representative of the general population. We found robust evidence for single-trait associations involving children's communication and joint association effects that were related to the phenotypic profile of studied traits. These findings were consistent with the autism spectrum disorder risk profile of the SNP identified in previous association studies (8
) and are unlikely to be affected by bias, multiple testing, phenotype interrelatedness, or random genotype dropout.
The strongest single-trait associations were observed for stereotyped conversation and pragmatic communication skills, as measured by two related CCC scales (see ), where a higher load of the risk allele was related to more communication deficits. Severe problems with pragmatic aspects of communication appear to be universal in autism spectrum disorder and are usually ascribed to an impaired theory of mind (37
)—that is, an impaired understanding of other people's minds and mental states. Assuming the dimensionality of communication problems, which is consistent with the previously observed continuum of traits contributing to the autistic spectrum in ALSPAC (18
), our findings provide evidence that variation at a high-risk locus for autism spectrum disorder is also related to a broader autism phenotype in the general population that is likely to manifest as inappropriate use of language in social contexts.
It is also possible that variation at rs4307059 is linked to some aspects of behavioral adjustment as reflected by special educational needs, although the identification of special educational needs is based on a large collection of criteria, including cognitive skills and the presence or absence of attention deficit hyperactivity disorder, speech problems, and physical difficulties (21
In addition, this study produced support for joint association effects that were not solely explained by single-trait associations, as the combined effect was stronger than for any single association alone and was still present when highly associated measures were excluded. Joint effects were related to a highly consistent pattern of single allelic associations involving language, communication, cognition, social interaction, and behavioral adjustment such that less functionality and more problems were associated with a higher frequency of the common rs4307059 risk allele, although the confidence intervals for single effects were often wide. Deviation from this pattern was detected for only one of 29 selected phenotypes. As our findings indicate, the likelihood that the same risk-associated allele would predict variation in these phenotypes within members of the general population by chance is very remote.
The exact risk structure of the joint association profile, however, is still unknown and likely to be complex. Given that even for highly associated single traits, the SNP explains only a small proportion of the phenotypic variance (<0.3%), as expected for genetic variation in general, the phenotypic relatedness among the measures is difficult to disentangle, especially as the joint effect captures both pleiotropic effects and shared sources of phenotypic variation. However, exclusion of measures of overall behavioral difficulties and special educational needs (in addition to highly associated single-trait associations) attenuated the combined effect only marginally. This lends support to the hypothesis that total behavioral difficulties might be mediated through social ability since adjustment problems, particularly in school, might be the consequence of inadequate social communicative skills (18
). More importantly, our findings provide evidence that the combined signal involves multiple social, communicative, and cognitive impairments, including lower verbal IQ (WISC-III) and impaired social communication skills (SCDC), none of which showed association with rs4307059 when adjusted for multiple testing. Common variation at rs4307059 may therefore be related to small impairments in overall social and communication skills, which might not be detectable as single signals (subthreshold impairments) but might increase the susceptibility for autism spectrum disorder. As such, variation at rs4307059 may provide the genetic background against which a more severe disruption of an autism spectrum disorder risk locus is more likely to result in a diagnosis of autism (38
). Although the molecular mechanisms are still unidentified, it has been hypothesized that rs4307059 tags causal variants in CDH9/CDH10
, a theory that is supported by pathway-based association analyses implicating neuronal cell-adhesion molecules in the etiology of autism spectrum disorder (8
Taken together, our results support the hypothesis that variability at rs4307059 is related to an underlying QTL for autism spectrum disorder in the general population. Moreover, they suggest that multiple facets of a broader autism phenotype share underlying genetic risk factors, which is consistent with previous reports on shared genetic influences for social, communication, and language skills (39
These findings must be interpreted within the context of potential limitations. First, although measures were derived from educational databases and mother and child reports, mother-reported data were overrepresented in the study and may have contributed to greater variance sharing. Second, the analysis included only white European children, and the reported genetic association therefore needs to be replicated in populations of different ethnic backgrounds. Third, instruments were chosen that focus on social communication spectrum phenotypes assessed between ages 3 and 12 for the majority of children in ALSPAC. However, they nonetheless represent a cohort-specific selection of measures, which may not be exhaustive. This is unlikely to affect the reported single-trait associations, but it may impinge on the total of the investigated phenotypic variance, which is reflected by the joint signal.