The clinical heterogeneity of autism spectrum disorders majorly challenges their genetic study. Autism spectrum disorders symptoms occur in milder forms in the general population, as autistic-like traits, and share genetic factors with autism spectrum disorders. Here, we investigate the genetics of individual autistic-like traits to improve our understanding of autism spectrum disorders. We meta-analysed four population-based genome-wide association studies investigating four autistic-like traits – ‘attention-to-detail’, ‘imagination’, ‘rigidity’ and ‘social-skills’ ( n = 4600). Using autism spectrum disorder summary statistics from the Psychiatric Genomic Consortium ( N = 46,350), we applied polygenic risk score analyses to understand the genetic relationship between autism spectrum disorders and autistic-like traits. Using MAGMA, we performed gene-based and gene co-expression network analyses to delineate involved genes and pathways. We identified two novel genome-wide significant loci – rs6125844 and rs3731197 – associated with ‘attention-to-detail’. We demonstrated shared genetic aetiology between autism spectrum disorders and ‘rigidity’. Analysing top variants and genes, we demonstrated a role of the immune-related genes RNF114, CDKN2A, KAZN, SPATA2 and ZNF816A in autistic-like traits. Brain-based genetic expression analyses further linked autistic-like traits to genes involved in immune functioning, and neuronal and synaptic signalling. Overall, our findings highlight the potential of the autistic-like trait–based approach to address the challenges of genetic research in autism spectrum disorders. We provide novel insights showing a potential role of the immune system in specific autism spectrum disorder dimensions. Lay abstract Autism spectrum disorders are complex, with a strong genetic basis. Genetic research in autism spectrum disorders is limited by the fact that these disorders are largely heterogeneous so that patients are variable in their clinical presentations. To address this limitation, we investigated the genetics of individual dimensions of the autism spectrum disorder phenotypes, or autistic-like traits. These autistic-like traits are continuous variations in autistic behaviours that occur in the general population. Therefore, we meta-analysed data from four different population cohorts in which autistic-like traits were measured. We performed a set of genetic analyses to identify common variants for autistic-like traits, understand how these variants related to autism spectrum disorders, and how they contribute to neurobiological processes. Our results showed genetic associations with specific autistic-like traits and a link to the immune system. We offer an example of the potential to use a dimensional approach when dealing with heterogeneous, complex disorder like autism spectrum disorder. Decomposing the complex autism spectrum disorder phenotype in its core features can inform on the specific biology of these features which is likely to account to clinical variability in patients.

BackgroundLarger than average head and brain sizes are often observed in individuals with Autism Spectrum Disorders (ASDs). ASDs and brain volume are both highly heritable, with multiple genetic variants contributing. However, it is unclear whether ASDs and brain volume share any genetic mechanisms. Genes from the mammalian target of rapamycin (mTOR) pathway influence brain volume, and variants are found in rare genetic syndromes that include autism spectrum disorder features. Here we investigated whether variants in mTOR-related genes are also associated with ASDs and if they constitute a genetic link between large brains and ASDs.MethodsWe extended our analyses between large heads (macrocephaly) and rare de novo mTOR-related variants in an intellectual disability cohort (N=2,258). Subsequently using Fisher’s exact tests we investigated the co-occurrence of mTOR-related de novo variants and ASDs in the denovo-db database (N=23,098). We next selected common genetic variants within a set of 96 mTOR-related genes in genome-wide genetic association data of ASDs (N=46,350) to test gene-set association using MAGMA. Lastly, we tested genetic correlation between genome-wide genetic association data of ASDs (N=46,350) and intracranial volume (N=25,974) globally using LD-score regression as well as mTOR-specific by restricting the genetic correlation to the mTOR-related genes using GNOVA.ResultsOur results show that both macrocephaly and ASDs occur above chance-level in individuals carrying rare de novo variants in mTOR-related genes. We found a significant mTOR gene-set association with ASDs (p=0.0029) and an mTOR-stratified positive genetic correlation between ASDs and intracranial volume (p=0.027), despite the absence of a significant genome-wide correlation (p=0.81).ConclusionsThis work indicates that both rare and common variants in mTOR-related genes are associated with brain volume and ASDs and genetically correlate them in the expected direction. We demonstrate that genes involved in mTOR signalling are potential mediators of the relationship between having a large brain and having ASDs.

Background Larger than average head and brain sizes are often observed in individuals with autism spectrum disorders (ASDs). ASD and brain volume are both highly heritable, with multiple genetic variants contributing. However, it is unclear whether ASD and brain volume share any genetic mechanisms. Genes from the mammalian target of rapamycin (mTOR) pathway influence brain volume, and variants are found in rare genetic syndromes that include ASD features. Here we investigated whether variants in mTOR‐related genes are also associated with ASD and if they constitute a genetic link between large brains and ASD. Methods We extended our analyses between large heads (macrocephaly) and rare de novo mTOR‐related variants in an intellectual disability cohort (N = 2,258). Subsequently using Fisher's exact tests we investigated the co‐occurrence of mTOR‐related de novo variants and ASD in the de‐novo‐db database (N = 23,098). We next selected common genetic variants within a set of 96 mTOR‐related genes in genome‐wide genetic association data of ASD (N = 46,350) to test gene‐set association using MAGMA. Lastly, we tested genetic correlation between genome‐wide genetic association data of ASD (N = 46,350) and intracranial volume (N = 25,974) globally using linkage disequilibrium score regression as well as mTOR specific by restricting the genetic correlation to the mTOR‐related genes using GNOVA. Results Our results show that both macrocephaly and ASD occur above chance level in individuals carrying rare de novo variants in mTOR‐related genes. We found a significant mTOR gene‐set association with ASD (p = .0029) and an mTOR‐stratified positive genetic correlation between ASD and intracranial volume (p = .027), despite the absence of a significant genome‐wide correlation (p = .81). Conclusions This work indicates that both rare and common variants in mTOR‐related genes are associated with brain volume and ASD and genetically correlate them in the expected direction. We demonstrate that genes involved in mTOR signalling are potential mediators of the relationship between having a large brain and having ASD.