Parsing autism heterogeneity: phenotypic differences between individuals across genes of different effect sizes
Clinical Genetics and Therapeutics
-
Primary Categories:
- Clinical Genetics
-
Secondary Categories:
- Clinical Genetics
Introduction:
SPARK (Simons Foundation Powering Autism Research for Knowledge) is the largest recontactable research cohort of individuals with autism and their family members, with more than 50,000 individuals with autism with genomic data and ~3,000 individuals with likely pathogenic/pathogenic (LP/P) autism-related variants. Although it is known that cognitive impairment is more frequent in individuals with rare loss-of-function (LoF) variants in genes associated with autism, it is not understood how the effect size of autism risk genes correlates with quantitative measures of core autism symptoms and co-occurring conditions. The aim of the current study is to investigate differences in these variables between individuals with predicted deleterious variants in genes that have large effect sizes for autism (odds ratio (OR) > 10) compared to genes with smaller effect sizes (OR < 10).
Methods:
We performed exome sequencing on 55,478 individuals with autism and genome sequencing on 2,769 individuals with autism, as well as 94,392 first degree family members. We identified 1,202 individuals with autism who had predicted loss-of-function (LoFs) variants in 76 genes (in which LoF is the known mechanism of action) with a published OR for autism. We filtered out individuals with mutations in more than one autism-related gene. We defined “large effect size genes” as genes with an OR >10 and “smaller effect size genes” as genes with an OR < 10. Sixty-eight percent (816) had LoFs variants in genes with OR > 10 (202 de novo in 240 trios with OR>10 and 57 de novo in 93 trios with OR<10). When available, we analyzed the Vineland-3 Adaptive Behavior Scales (VABS) scores for socialization (SOC), daily living skills (DLS), communication (COM) , and overall score (Adaptive Behavior Composite, or ABC), the Social Communication Questionnaire (SCQ) summary score, the Repetitive Behavior Scale – Revised (RBS-R), and other phenotypic variables, such as age at autism diagnosis, language abilities, developmental milestones, support requirements, and medical and mental health diagnoses (ADHD, schizophrenia, bipolar disorder, congenital anomalies, seizures) using statistical tests in R.
Results:
Individuals with LoFs in large effect genes had significantly lower scores on all VABS measures except for socialization (DLS: mean difference=6.8, p=0.0005; COM: mean dif.=6.0, p=0.007; SOC: mean dif.=3.8 p=0.1; ABC: mean dif.=4.7, p=0.008) than individuals with LoFs in lower effect size genes. Among the VABS domains, individuals with LoFs in large effect size genes scored lowest on the DLS domain. Individuals with LoFs in large effect size genes did not have significantly different scores on the SCQ (mean dif.=0.5, p=0.4). However, people with LoFs in large effect size genes were significantly more likely to have intellectual disability (OR=1.8, CI:1.4-2.4 p=8.5e-07), require more substantial supports (OR=1.6, CI:1.1-2.3, p=0.02), and have lower language abilities (OR=1.6, CI:1.1-2.4, p=0.01). The age at autism diagnosis was significantly lower (mean dif.=8.8 months, p=0.05) for individuals with LoFs in large effect size genes.
In contrast, individuals with LoFs in smaller effect size genes were significantly more likely to have ADHD (OR=1.3, CI:1.0-1.7, p=0.05) and bipolar disorder (OR=2.2, CI:1.2-4.2, p=0.01) than individuals with LoFs in large effect genes. Individuals with LoFs in smaller effect size genes also displayed more repetitive behaviors on the RBS-R (mean dif.=4.5, p=0.007).
Conclusion:
We found that overall, individuals with LoFs in large effect size genes had lower scores on adaptive behavior measures and cognitive impairment, but neither the large effect nor smaller effect genes had a significant effect on social behaviors. However, individuals with LoFs in smaller effect size genes were more likely to have co-occurring ADHD and bipolar disorder as well as more repetitive behaviors. Additional data are needed to better understand the phenotypic correlations with effect size among people with autism with genetic diagnoses.
SPARK (Simons Foundation Powering Autism Research for Knowledge) is the largest recontactable research cohort of individuals with autism and their family members, with more than 50,000 individuals with autism with genomic data and ~3,000 individuals with likely pathogenic/pathogenic (LP/P) autism-related variants. Although it is known that cognitive impairment is more frequent in individuals with rare loss-of-function (LoF) variants in genes associated with autism, it is not understood how the effect size of autism risk genes correlates with quantitative measures of core autism symptoms and co-occurring conditions. The aim of the current study is to investigate differences in these variables between individuals with predicted deleterious variants in genes that have large effect sizes for autism (odds ratio (OR) > 10) compared to genes with smaller effect sizes (OR < 10).
Methods:
We performed exome sequencing on 55,478 individuals with autism and genome sequencing on 2,769 individuals with autism, as well as 94,392 first degree family members. We identified 1,202 individuals with autism who had predicted loss-of-function (LoFs) variants in 76 genes (in which LoF is the known mechanism of action) with a published OR for autism. We filtered out individuals with mutations in more than one autism-related gene. We defined “large effect size genes” as genes with an OR >10 and “smaller effect size genes” as genes with an OR < 10. Sixty-eight percent (816) had LoFs variants in genes with OR > 10 (202 de novo in 240 trios with OR>10 and 57 de novo in 93 trios with OR<10). When available, we analyzed the Vineland-3 Adaptive Behavior Scales (VABS) scores for socialization (SOC), daily living skills (DLS), communication (COM) , and overall score (Adaptive Behavior Composite, or ABC), the Social Communication Questionnaire (SCQ) summary score, the Repetitive Behavior Scale – Revised (RBS-R), and other phenotypic variables, such as age at autism diagnosis, language abilities, developmental milestones, support requirements, and medical and mental health diagnoses (ADHD, schizophrenia, bipolar disorder, congenital anomalies, seizures) using statistical tests in R.
Results:
Individuals with LoFs in large effect genes had significantly lower scores on all VABS measures except for socialization (DLS: mean difference=6.8, p=0.0005; COM: mean dif.=6.0, p=0.007; SOC: mean dif.=3.8 p=0.1; ABC: mean dif.=4.7, p=0.008) than individuals with LoFs in lower effect size genes. Among the VABS domains, individuals with LoFs in large effect size genes scored lowest on the DLS domain. Individuals with LoFs in large effect size genes did not have significantly different scores on the SCQ (mean dif.=0.5, p=0.4). However, people with LoFs in large effect size genes were significantly more likely to have intellectual disability (OR=1.8, CI:1.4-2.4 p=8.5e-07), require more substantial supports (OR=1.6, CI:1.1-2.3, p=0.02), and have lower language abilities (OR=1.6, CI:1.1-2.4, p=0.01). The age at autism diagnosis was significantly lower (mean dif.=8.8 months, p=0.05) for individuals with LoFs in large effect size genes.
In contrast, individuals with LoFs in smaller effect size genes were significantly more likely to have ADHD (OR=1.3, CI:1.0-1.7, p=0.05) and bipolar disorder (OR=2.2, CI:1.2-4.2, p=0.01) than individuals with LoFs in large effect genes. Individuals with LoFs in smaller effect size genes also displayed more repetitive behaviors on the RBS-R (mean dif.=4.5, p=0.007).
Conclusion:
We found that overall, individuals with LoFs in large effect size genes had lower scores on adaptive behavior measures and cognitive impairment, but neither the large effect nor smaller effect genes had a significant effect on social behaviors. However, individuals with LoFs in smaller effect size genes were more likely to have co-occurring ADHD and bipolar disorder as well as more repetitive behaviors. Additional data are needed to better understand the phenotypic correlations with effect size among people with autism with genetic diagnoses.