Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2014 Oct 6;111(42):15161–15165. doi: 10.1073/pnas.1409204111

Autism spectrum disorder severity reflects the average contribution of de novo and familial influences

Elise B Robinson a,b,1, Kaitlin E Samocha a,b,c, Jack A Kosmicki a,b,d, Lauren McGrath e,f, Benjamin M Neale a,b, Roy H Perlis b,f, Mark J Daly a,b
PMCID: PMC4210299  PMID: 25288738

Significance

Autism spectrum disorder (ASD) research is complicated by heterogeneity. There are several types of genetic risk factors for ASDs, and that diversity may be reflected in case presentation. This study presents evidence for systematic variation in the genetic architecture of ASDs in which higher functioning cases, defined through cognitive and behavioral assessments, are more likely to manifest familial influences. This finding suggests that genetic and neurobiological research into ASDs and other neuropsychiatric disorders may be pursued more efficiently through greater phenotypic characterization.

Keywords: neuropsychiatric genetics, epidemiology, heterogeneity, phenotype

Abstract

Autism spectrum disorders (ASDs) are a highly heterogeneous group of conditions—phenotypically and genetically—although the link between phenotypic variation and differences in genetic architecture is unclear. This study aimed to determine whether differences in cognitive impairment and symptom severity reflect variation in the degree to which ASD cases reflect de novo or familial influences. Using data from more than 2,000 simplex cases of ASD, we examined the relationship between intelligence quotient (IQ), behavior and language assessments, and rate of de novo loss of function (LOF) mutations and family history of broadly defined psychiatric disease (depressive disorders, bipolar disorder, and schizophrenia; history of psychiatric hospitalization). Proband IQ was negatively associated with de novo LOF rate (P = 0.03) and positively associated with family history of psychiatric disease (P = 0.003). Female cases had a higher frequency of sporadic genetic events across the severity distribution (P = 0.01). High rates of LOF mutation and low frequencies of family history of psychiatric illness were seen in individuals who were unable to complete a traditional IQ test, a group with the greatest degree of language and behavioral impairment. These analyses provide strong evidence that familial risk for neuropsychiatric disease becomes more relevant to ASD etiology as cases become higher functioning. The findings of this study reinforce that there are many routes to the diagnostic category of autism and could lead to genetic studies with more specific insights into individual cases.

The set of conditions diagnosed as autism spectrum disorders (ASDs) vary enormously in their presentation (1). The most severely impaired individuals—often those with intellectual disabilities, limited speech, and severe behavioral problems—can require lifelong care. At the other end of the functional spectrum, people diagnosed with ASDs can be verbally fluent and academically gifted and can achieve independence in adulthood (2, 3). The broad range of cognitive and behavioral profiles seen in diagnosed ASDs has been long viewed as a challenge by the research community (4). Although it is well established that (i) the cognitive/behavioral profile of people diagnosed with ASDs varies widely and (ii) the set of genetic factors related to ASDs varies widely (5, 6), the degree to which phenotype can be used to predict patterns in disease architecture remains unclear.

Recent insights into the genetic influences on ASDs offer an opportunity to investigate this question through the lens of de novo vs. familial effects. On average, ASDs run in families. The siblings of children with ASDs are 10–20 times more likely to receive a diagnosis of ASD themselves (7, 8); the parents of children with ASDs are more likely to manifest autistic features, as well as a variety of other neuropsychiatric conditions, such as schizophrenia and bipolar disorder (9, 10). These epidemiologic observations are consistent with analyses suggesting that ASDs are influenced by thousands of common genetic variants transmitted between generations. It has been estimated that common, genotyped SNPs account for 20–60% of variation in ASD risk, although the effect of any individual SNP is likely very small (11–13). Many of these influences are shared with other psychiatric disorders (12, 14), which at least in part explains the familial clustering of different types of behavior problems.

However, statistics about ASD heritability reflect an average. For example, there are likely many affected families for whom sibling recurrence risk is less than 10–20%. The strongest evidence toward this claim comes from studies of rare, severely deleterious genetic events that are associated with ASDs (15–20). Events of this type, for example copy number variants and loss of function (LOF) mutations, are often de novo (not seen in an affected individual’s parents). Although cases of ASDs involving a de novo mutation could reflect a concert of spontaneous and inherited genetic events, de novo events of large effect may reduce the likelihood of seeing psychiatric problems in an affected individual’s family members.

Identifying Variables That Index Differences in Genetic Architecture

The link between phenotypic and genotypic heterogeneity could involve many types of variables. This analysis focuses on intelligence and symptom severity because existing literature suggests they are strong candidate sources of etiologic differences. First, variation in genetic architecture has been associated with the intelligence quotient (IQ) distribution itself. Although IQ overall is highly heritable (21–24), there is evidence that severe intellectual disabilities, which exist at the low tail of the distribution, may have fewer inherited, familial influences (25). This finding is consistent with genetic reports of strong de novo influences on severe intellectual disability (26, 27) and suggests that the IQ distribution within ASDs could follow a similar trend. Second, studies have associated de novo copy number variants (CNVs) and point mutations with reductions in average IQ in ASD (20, 28, 29), although the extent to which this pattern extends to broader patterns of genetic architecture or case severity is unclear. Symptom severity, which can include impairments in social, communication, and daily living abilities, are associated with IQ in ASD but are likely, in many cases, to provide an independent route to reproductive disadvantage. To examine that possibility, these analyses test etiologic structure in a group of ASD individuals identified for extreme behavioral and language impairments. Last, cognition and case severity are associated with the ASD sex ratio, which may also be indicative of patterns in etiologic structure (30). The male to female ratio in ASD, which is estimated to be 4:1 overall, increases across the IQ range (31). To the extent that females are protected from familial influences on ASD risk (32), this may suggest that higher-functioning ASD cases reflect a greater average degree of inherited genetic influences.

This report presents evidence for systematic association between clinical phenotype and ASD etiology using multiple lines of analysis, suggesting that higher functioning male cases have a stronger familial component. Lower-functioning cases, and female cases overall, are more likely to reflect sporadic genetic influences. These findings could lead to more efficient and interpretable genetic analyses in ASD.

Methods

This study uses data from the Simons Simplex Collection (SSC), a deeply phenotyped sample of more than 2,800 individuals with ASDs and their families (33). The SSC includes only simplex cases of ASD, defined through the lack of another affected family member as close or closer than first cousins. A subset (n = 774) of the SSC families contributed to exome sequencing studies, the methodological details of which have been previously described (15–17). Each of those analyses suggested a role of de novo LOF mutations in the etiology of ASDs and, in a joint analysis of those datasets, LOF mutations—specifically any nonsense, frameshift, or splice site mutation—were significantly associated with disease status (29). In this analysis, we determined which individuals carry de novo LOF mutations using the published SSC exome sequencing studies and linked those data to the SSC phenotype registry. LOF status and detailed phenotypic and family history were available for every individual in the sequenced subset. We evaluate the previously published copy number variation data (20) in the same fashion.

Measures.

Several measures of IQ were used across SSC sites. In assessing IQ, we only used those tests designed to provide a comparable full scale IQ score (34–38) (Table S1). Full-scale IQ was the primary predictor in all analyses; among the 2,256 individuals for whom it was measured, full-scale IQ was very highly correlated with both verbal (r = 0.88, P < 0.0001) and nonverbal (r = 0.94, P < 0.0001) IQ, suggesting a cohesive construct. Individuals who did not attempt an IQ test, or were instead given a test of developmental stage [e.g., Mullen Scales of Early Learning (39)], were not included in the IQ analyses (n = 251, 8.8%). Individuals who attempted an IQ test, indicated by a partially complete test record, but who did not receive a full-scale IQ score were labeled “IQ attempted but unscored” (n = 350, 13.4% of those who attempted an IQ test). Individuals who attempted but did not receive an IQ score had, on average, substantially greater social, communication, and daily living skills impairments as measured by the Vineland Scales of Adaptive Behavior (40) than those who did receive an IQ score [measured IQ mean 76.7 (SD 10.3); attempted but unscored mean 61.1 (SD 10.0); P = 5.6e-138]. Individuals in the IQ attempted but unscored group were also more than 20 times more likely not to have developed phrase speech, controlling for age at time of Autism Diagnostic Interview (ADI) (P = 6.1e-30), suggesting that language and functional impairment interfered with valid IQ test administration. The IQ attempted but unscored group had greater average symptom severity than ASD individuals with IQ less than 55 (P = 0.002; Table S2); the severity is accordingly unlikely to be purely a function of unmeasured but very low IQ.

The family history analysis focused on major adult psychiatric disorders with previous genetic and familial association to ASDs: depressive disorders (DD), bipolar disorder (BPD), and schizophrenia (SCZ) (10, 14, 41). Each of those disorders, as well as family history of psychiatric hospitalization, was assessed through parent report questionnaire. Parents indicated whether any blood relatives had, for example, SCZ, and if yes noted which relative. Age at diagnosis was not available in the majority of cases. The SSC measured family history to the level of first cousins; relatives by marriage were not assessed. We used three autoimmune disorders— type 2 diabetes, rheumatoid arthritis, and hyperthyroidism— with similar collective family history prevalence (54.99%) to DD, BPD, and SCZ (50.25%) to examine the specificity of any family history effect. Core ASD behavioral and language variation was measured using the Vineland Scales of Adaptive Behavior and the ADI–Revised (40, 42). History of seizures in the proband was assessed with the SSC medical history questionnaire. Each of the measures is described in detail in Table S1.

Analyses.

We examined the proband full-scale IQ distribution for evidence of a mixture of causal influences by comparing the empirical distribution to one simulated to have a similar mean shift but an SD consistent with that expected in the general population. In the population, the IQ distribution is normed to an SD of 15. Quantitative genetic theory predicts that a consistent set of etiologic influences on IQ in ASD would shift the mean but maintain the distribution’s shape. Mixture effects are therefore suggested by changes in the shape or SD of a distribution.

We then investigated variation in the genetic architecture of ASDs using two approaches. First, in the sequenced subset of the SSC (n = 774), we used Poisson regression to associate the rate of de novo LOF mutation with proband full-scale IQ (n = 614 individuals with measured IQ values). The rate of LOF mutation in the IQ attempted but unscored group (n = 102) was analyzed against expectation using a Poisson probability distribution. The expected rate of LOF mutations (0.085 per exome) was estimated using a statistical model of the rate of LOF mutation across the genome (29). Because ∼8.5% of the general population carries a de novo LOF mutation, no individual de novo LOF mutation can be considered causal. The purpose of this analysis was to examine how the rate within ASD varies against that background rate.

Second, in the full SSC sample, we examined whether the LOF association was more broadly indicative of variation in the genetic architecture of ASDs. We categorized family history of major adult psychiatric disorders into (i) history of DD, BPD, or SCZ, and (ii) history of psychiatric hospitalization. We regressed the log odds of both history categories against IQ and examined the history probability separately in the IQ attempted but unscored group. In each analysis we examined all probands together, followed by an analysis stratified by proband sex. The family history analyses controlled for proband race/ethnicity (white non-Hispanic, black Non-Hispanic, Asian, Hispanic, other) which has been linked to variation in psychiatric disease prevalence and reporting (43). Although the broad autism phenotype (BAP) could also in principle be used to examine variation in ASD familiality (44, 45), we did not include BAP or autistic trait measures in the analysis because suspected ASD in family members (which could be indexed by a high concentration of autistic traits) was used as an SSC exclusion criterion (33). Last, we also examined the relationship between LOF rate and other variables in the sequenced subset—including language development, behavioral severity, and proband seizure history—controlling for proband full-scale IQ in each analysis. Controlling for IQ test used did not impact any of the above associations with IQ (de novo LOF rate, family history of psychiatric disease), and test was accordingly not included in the final analyses.

Results

Evidence of Mixture Effects.

The empirical distribution of IQ in ASD suggests causal heterogeneity. The SD (20.28) of IQ is increased relative to that expected in the general population, resulting in a distribution that is heavy at the tails. In SSC data, both high (IQ >130; P = 2.90e-64) and low (IQ <70; P = 5.23e-32) IQ are significantly overrepresented compared with expectation (Fig. S1).

Case Severity, LOF Mutations, and Sex.

Of the 774 individuals with trio sequencing data, 106 had a de novo LOF mutation. This corresponds to a de novo LOF rate per exome that significantly exceeds that expected in the population (P = 1.38e-07), as previously reported (29). As shown in Fig. 1, increases in IQ were associated with lower LOF rate (P = 0.03), but that relationship was only independently significant in males (males, P = 0.04; females, P = 0.72). The overall association was consistent across the IQ distribution, as shown in Fig. S2. The sex difference was not statistically significant and, because of a limited number of female probands, the extent to which the LOF–IQ association extends to female probands is accordingly unclear. The relationship between IQ and de novo rate was similar when LOF mutations and CNVs were analyzed jointly (P = 0.01), although the CNV-only association was not independently significant (P = 0.30; Fig. S3). LOF rate was elevated against expectation in both males (n = 72; rate = 0.181; P = 0.01) and females (n = 30; rate = 0.266; P = 0.02) with IQ attempted but unscored. Together with the regression of IQ against LOF rate, these findings suggest that the clinical category of ASDs includes both (i) individuals with a de novo LOF rate equivalent to that seen in severe intellectual disability (26, 27), a phenotype associated with extreme reproductive disadvantage, and (ii) individuals with no increased burden of de novo mutations relative to the rate expected in the general population. The severity effects were consistent across the continuous and categorical analyses—the highest rates of LOF mutation were noted in the low and unscored IQ groups. Because the IQ unscored group is most notable for severe behavioral and language impairment, this suggests that LOF mutations in ASD predispose individuals to both low IQ and severe autism-like behavioral impairments that may have equal or greater effect on functional ability. As shown in Table 1, females had nearly twice the LOF rate compared with males (females: 0.204 per exome; males: 0.129 per exome; P = 0.01). There was no difference in the de novo LOF rate between male and female unaffected siblings (ratemales = 0.064, nmales = 297; ratefemales = 0.061, nfemales= 311; P = 0.52).

Fig. 1.

Fig. 1.

Open in a new tab

De novo LOF mutations and IQ in ASDs. Error bars on IQ attempted but not scored estimates indicate ±1 SE.

Table 1.

De novo LOF per exome in sequenced SSC subset (n = 770)

Sex N Observed P (vs. expected rate in the population) P (vs. male rate in SSC)
Males 618 0.129 0.002 —
Females 152 0.204 3.96E-05 0.01
Open in a new tab

Case Severity and Family History of Psychiatric Disease.

Family history analyses suggested that the de novo pattern is more broadly indicative of differences in the genetic architecture of ASDs. Fig. 2 illustrates that increases in measured IQ were associated with increases in family history of major adult psychiatric diseases (DD, BPD, or SCZ; P = 0.003) and psychiatric hospitalizations (P = 0.04). The consistency of the psychiatric disease association across deciles of the IQ distribution is shown in Fig. S4. In contrast, there was no overall association between measured IQ and family history of autoimmune disease (P = 0.92). The IQ attempted but unscored group had low family history rates, similar to individuals with low measured IQ. Males with IQ attempted but unscored (n = 275) did not have significantly different family history of DD, BPD, or SCZ (46.2%) or psychiatric hospitalization (14.2%) than females with IQ attempted but unscored [n = 69; DD, BPD, SCZ frequency 32.3% (P = 0.09); hospitalization history 6.2% (P = 0.10)]. The major adult psychiatric disease association was also statistically significant in the sequenced subset alone (n = 774; P = 0.01), and the association did not change when controlling for proband de novo LOF mutations.

Fig. 2.

Fig. 2.

Open in a new tab

Family history of psychiatric disease and IQ in ASDs. P values are adjusted for race/ethnicity; purple circles indicate fraction of participants with a family history of psychiatric hospitalization, for males and females respectively; blue circles indicate fraction of participants with a family history of DD, BPD, or SCZ, for males and females respectively; error bars on all IQ attempted but not scored estimates indicate ±1 SE. For family history of hospitalization, the male-specific line is obscured by that for all individuals.

Language, ASD-Related Behaviors, and Seizure History.

Because cases with low IQ are likely to be lower functioning in behavioral and language domains (e.g., Vineland social impairment, r = 0.27, P = 2.74e-40; Vineland daily living skills impairment, r = 0.42, P = 1.64e-96; ADI age at language development, r = −0.40, P = 3.82e-87), the measured IQ analyses above capture aspects of severity in addition to intellectual impairment. However, none of the behavioral and language variables were independently associated with LOF rate once IQ variance was regressed out (P > 0.05 for all comparisons; Table S3). A history of seizures in the proband was also associated with lower IQ (P = 8.9e-8), but seizure history was not independently associated with LOF rate after controlling for intelligence (P = 0.50).

Comment

Genetic studies of ASDs have yielded consistent, positive associations with de novo as well as inherited variants (12–18, 46). The analyses above suggest that both of these types of genetic variation can be used to distinguish groups of individuals with different probabilities of case severity. This bears not only great relevance toward our understanding of the biological predictors of ASDs, but may also be useful in predicting clinical trajectories. For example, ASD cases with de novo LOF mutations in target genes of the fragile X mental retardation protein have a mean IQ an SD below that of cases without an LOF mutation (P = 0.0005), and higher IQ has been linked to improved adult outcomes in ASD (2, 3).

These findings also suggest that many ASD statistics reflect an average estimated across different types of cases. For example, the heritability of ASDs is often presented as a summary of previously estimated values (e.g., 60%). These results, however, predict that the heritability of ASDs will vary between cases, specifically increasing with IQ and the proband’s level of functioning. Any single estimated value—of heritability, sibling recurrence risk, the role of common variants, and many others—is likely to only apply to a subset of cases.

The analyses in this report may have benefited from the homogeneity of a simplex sample. A group of cases selected specifically for the absence of ASD in their family may yield a more homogenous and interpretable association between phenotype and etiologic structure. Although the degree to which these patterns will extend to multiplex families cannot be predicted here, the family history analyses presented above do call into question the definition of simplex as it is currently used in ASD research. Phenotypic severity may predict de novo rates in ASD better than family history, particularly if simplex status is assessed through family history of autism alone (18). This is consistent with emerging data from other neuropsychiatric disorders. The rate of de novo mutations in a neuropsychiatric disease group seems to correlate with the disorders’ associated degree of intellectual impairment [P(de novo event): Intellectual Disability (26, 27) > ASD (29) > SCZ (47)].

Interpretation of these data is limited by the lack of unrelated controls, as well as the absence of IQ data for unaffected siblings. To our knowledge, there are no large, existing trio sequenced data sets with IQ data on individuals without neuropsychiatric disorders. As such, it is difficult to determine whether there is an ASD-specific shift in the relationship between IQ and LOF rate that exists over and above any association that may exist in the general population. In these analyses, the overall IQ–LOF regression line crosses the general population LOF rate at approximately IQ = 125 (Fig. 1). Were the overall LOF burden in ASD reflecting only the increase in low IQ, one would expect those lines to cross around an IQ of 100, the general population average. The shift suggests that there may be de novo LOF relevance specific to autism outside of increased risk for intellectual disability, but it is difficult to assess in the absence of IQ matched and trio sequenced controls. Particularly in ASD cases with severe intellectual disability, it is possible that the ID mediates any LOF impacts, and the ASD symptoms are a nonspecific, but clinically significant, secondary outcome.

Further, although these analyses suggest that the role of inherited genetic risk for ASDs increases among high-functioning individuals, the family history associations could also reflect shared environmental influences. Additional studies using measured genetic data or twin samples will be needed to clarify the nature of the familial risk. Further analysis will also be needed to clarify the extent to which these findings apply to female probands. Neither the IQ–LOF nor IQ–family history trends were independently significant in females, which could suggest either insufficient statistical power or a true sex difference in the relationship between functioning and etiologic structure in ASDs. Sex differences in ASD ascertainment may also affect estimated sex differences in etiologic structure, because high-functioning females may be less likely to obtain a diagnosis than high-functioning males (4). Because the SSC instated minimum requirements for cognitive function (33), ascertainment may also explain the relatively inflated male/female ratio in these analyses (∼6.2:1).

To our knowledge, these analyses provide one of the first clear examples of genetic structure being able to refine nosology in neuropsychiatric illness. These findings carry substantial implications for the design of genetic studies in autism research and suggest that phenotypic subcategories may be useful in predicting etiologic heterogeneity.

Supplementary Material

Supplementary File
pnas.201409204SI.pdf (663KB, pdf)

Acknowledgments

We thank the families who took part in the Simons Simplex Collection study and the clinicians who collected data at each of the study sites. E.B.R. was funded by National Institutes of Mental Health Grant 1K01MH099286-01A1.

Footnotes

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.

References

  • 1.American Psychiatric Association . Diagnostic and Statistical Manual of Mental Disorders. 5th Ed American Psychiatric Publishing; Arlington, VA: 2013. [Google Scholar]
  • 2.Fein D, et al. Optimal outcome in individuals with a history of autism. J Child Psychol Psychiatry. 2013;54(2):195–205. doi: 10.1111/jcpp.12037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Gotham K, Pickles A, Lord C. Trajectories of autism severity in children using standardized ADOS scores. Pediatrics. 2012;130(5):e1278–e1284. doi: 10.1542/peds.2011-3668. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Skuse DH. Is autism really a coherent syndrome in boys, or girls? Br J Psychol. 2009;100(Pt 1):33–37. doi: 10.1348/000712608X369459. [DOI] [PubMed] [Google Scholar]
  • 5.Stein JL, Parikshak NN, Geschwind DH. Rare inherited variation in autism: Beginning to see the forest and a few trees. Neuron. 2013;77(2):209–211. doi: 10.1016/j.neuron.2013.01.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Berg JM, Geschwind DH. Autism genetics: Searching for specificity and convergence. Genome Biol. 2012;13(7):247. doi: 10.1186/gb-2012-13-7-247. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Constantino JN, Zhang Y, Frazier T, Abbacchi AM, Law P. Sibling recurrence and the genetic epidemiology of autism. Am J Psychiatry. 2010;167(11):1349–1356. doi: 10.1176/appi.ajp.2010.09101470. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Ozonoff S, et al. Recurrence risk for autism spectrum disorders: A Baby Siblings Research Consortium study. Pediatrics. 2011;128(3):e488–e495. doi: 10.1542/peds.2010-2825. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Wheelwright S, Auyeung B, Allison C, Baron-Cohen S. Defining the broader, medium and narrow autism phenotype among parents using the Autism Spectrum Quotient (AQ) Mol Autism. 2010;1(1):10. doi: 10.1186/2040-2392-1-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Sullivan PF, et al. Family history of schizophrenia and bipolar disorder as risk factors for autism. Arch Gen Psychiatry. 2012;69(11):1099–1103. doi: 10.1001/archgenpsychiatry.2012.730. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Anney R, et al. Individual common variants exert weak effects on the risk for autism spectrum disorderspi. Hum Mol Genet. 2012;21(21):4781–4792. doi: 10.1093/hmg/dds301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Lee SH, et al. Cross-Disorder Group of the Psychiatric Genomics Consortium International Inflammatory Bowel Disease Genetics Consortium (IIBDGC) Genetic relationship between five psychiatric disorders estimated from genome-wide SNPs. Nat Genet. 2013;45(9):984–994. doi: 10.1038/ng.2711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Gaugler T, et al. Most genetic risk for autism resides with common variation. Nat Genet. 2014;46(8):881–885. doi: 10.1038/ng.3039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Smoller JW, et al. Cross-Disorder Group of the Psychiatric Genomics Consortium Identification of risk loci with shared effects on five major psychiatric disorders: A genome-wide analysis. Lancet. 2013;381(9875):1371–1379. doi: 10.1016/S0140-6736(12)62129-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Iossifov I, et al. De novo gene disruptions in children on the autistic spectrum. Neuron. 2012;74(2):285–299. doi: 10.1016/j.neuron.2012.04.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.O’Roak BJ, et al. Sporadic autism exomes reveal a highly interconnected protein network of de novo mutations. Nature. 2012;485(7397):246–250. doi: 10.1038/nature10989. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Sanders SJ, et al. De novo mutations revealed by whole-exome sequencing are strongly associated with autism. Nature. 2012;485(7397):237–241. doi: 10.1038/nature10945. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Neale BM, et al. Patterns and rates of exonic de novo mutations in autism spectrum disorders. Nature. 2012;485(7397):242–245. doi: 10.1038/nature11011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Levy D, et al. Rare de novo and transmitted copy-number variation in autistic spectrum disorders. Neuron. 2011;70(5):886–897. doi: 10.1016/j.neuron.2011.05.015. [DOI] [PubMed] [Google Scholar]
  • 20.Sanders SJ, et al. Multiple recurrent de novo CNVs, including duplications of the 7q11.23 Williams syndrome region, are strongly associated with autism. Neuron. 2011;70(5):863–885. doi: 10.1016/j.neuron.2011.05.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Deary IJ, et al. Genetic contributions to stability and change in intelligence from childhood to old age. Nature. 2012;482(7384):212–215. doi: 10.1038/nature10781. [DOI] [PubMed] [Google Scholar]
  • 22.Davies G, et al. Genome-wide association studies establish that human intelligence is highly heritable and polygenic. Mol Psychiatry. 2011;16(10):996–1005. doi: 10.1038/mp.2011.85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Trzaskowski M, Yang J, Visscher PM, Plomin R. DNA evidence for strong genetic stability and increasing heritability of intelligence from age 7 to 12. Mol Psychiatry. 2014;19(3):380–384. doi: 10.1038/mp.2012.191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Benyamin B, et al. Wellcome Trust Case Control Consortium 2 (WTCCC2) Childhood intelligence is heritable, highly polygenic and associated with FNBP1L. Mol Psychiatry. 2014;19(2):253–258. doi: 10.1038/mp.2012.184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Nichols PL. Familial mental retardation. Behav Genet. 1984;14(3):161–170. doi: 10.1007/BF01065538. [DOI] [PubMed] [Google Scholar]
  • 26.Rauch A, et al. Range of genetic mutations associated with severe non-syndromic sporadic intellectual disability: An exome sequencing study. Lancet. 2012;380(9854):1674–1682. doi: 10.1016/S0140-6736(12)61480-9. [DOI] [PubMed] [Google Scholar]
  • 27.de Ligt J, et al. Diagnostic exome sequencing in persons with severe intellectual disability. N Engl J Med. 2012;367(20):1921–1929. doi: 10.1056/NEJMoa1206524. [DOI] [PubMed] [Google Scholar]
  • 28.Ronemus M, Iossifov I, Levy D, Wigler M. The role of de novo mutations in the genetics of autism spectrum disorders. Nat Rev Genet. 2014;15(2):133–141. doi: 10.1038/nrg3585. [DOI] [PubMed] [Google Scholar]
  • 29.Samocha KE, et al. A framework for the interpretation of de novo mutation in human disease. Nat Genet. 2014;46(9):944–950. doi: 10.1038/ng.3050. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Jacquemont S, et al. A higher mutational burden in females supports a “female protective model” in neurodevelopmental disorders. Am J Hum Genet. 2014;94(3):415–425. doi: 10.1016/j.ajhg.2014.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Fombonne E. Epidemiology of autistic disorder and other pervasive developmental disorders. J Clin Psychiatry. 2005;66(Suppl 10):3–8. [PubMed] [Google Scholar]
  • 32.Robinson EB, Lichtenstein P, Anckarsäter H, Happé F, Ronald A. Examining and interpreting the female protective effect against autistic behavior. Proc Natl Acad Sci USA. 2013;110(13):5258–5262. doi: 10.1073/pnas.1211070110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Fischbach GD, Lord C. The Simons Simplex Collection: A resource for identification of autism genetic risk factors. Neuron. 2010;68(2):192–195. doi: 10.1016/j.neuron.2010.10.006. [DOI] [PubMed] [Google Scholar]
  • 34.Wechsler D. Weschler Abbreviated Scale of Intelligence (WASI) Harcourt Assessment; San Antonio, TX: 1997. [Google Scholar]
  • 35.Wechsler D. Weschler Primary and Preschool Scale of Intelligence. 3rd Ed Harcourt Assessment; San Antonio, TX: 2002. [Google Scholar]
  • 36.Wechsler D. Weschler Intelligence Scale for Children–4th Edition (WISC-IV) Harcourt Assessment; San Antonio, TX: 2003. [Google Scholar]
  • 37.Elliott CD. DAS Administration and Scoring Manual. The Psychological Corporation; San Antonio, TX: 1990. [Google Scholar]
  • 38.Elliott CD. DAS Introductory and Technical Handbook. The Psychological Corporation; San Antonio, TX: 1990. [Google Scholar]
  • 39.Mullen EM. Mullen Scales of Early Learning. American Guidance Service Inc.; Circle Pines, MN: 1995. [Google Scholar]
  • 40.Sparrow SS, Cicchetti DV, Balla DA. Vineland Adaptive Behavior Scales. 2nd Ed Pearson; Livonia, MN: 2005. [Google Scholar]
  • 41.Vasa RA, et al. Mood disorders in mothers of children on the autism spectrum are associated with higher functioning autism. Autism Res Treat. 2012;2012:435646. doi: 10.1155/2012/435646. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Le Couteur A, Lord C, Rutter M. Autism Diagnostic Interview–Revised (ADI-R) Western Psychological Services; Portland, OR: 2003. [Google Scholar]
  • 43.Riolo SA, Nguyen TA, Greden JF, King CA. Prevalence of depression by race/ethnicity: Findings from the National Health and Nutrition Examination Survey III. Am J Public Health. 2005;95(6):998–1000. doi: 10.2105/AJPH.2004.047225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Robinson EB, et al. Evidence that autistic traits show the same etiology in the general population and at the quantitative extremes (5%, 2.5%, and 1%) Arch Gen Psychiatry. 2011;68(11):1113–1121. doi: 10.1001/archgenpsychiatry.2011.119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Lundström S, et al. Autism spectrum disorders and autistic like traits: Similar etiology in the extreme end and the normal variation. Arch Gen Psychiatry. 2012;69(1):46–52. doi: 10.1001/archgenpsychiatry.2011.144. [DOI] [PubMed] [Google Scholar]
  • 46.Sullivan PF, Daly MJ, O’Donovan M. Genetic architectures of psychiatric disorders: The emerging picture and its implications. Nat Rev Genet. 2012;13(8):537–551. doi: 10.1038/nrg3240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Fromer M, et al. De novo mutations in schizophrenia implicate synaptic networks. Nature. 2014;506(7487):179–184. doi: 10.1038/nature12929. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary File
pnas.201409204SI.pdf (663KB, pdf)

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES