The Autism Genome Project

Autism Spectrum Disorders are severe neurodevelopmental disorders, with a prevalence of ~60 in 10000, characterised by impairments in verbal and non-verbal communication, processing social and emotional information and repetitive or stereotyped behaviours. The cause of autism is unknown but there is strong evidence for the involvement of inherited genetic factors. The Autism Genome Project (AGP) proposes to tease apart the complex genetic etiology of autism. The AGP brings together leading experts in autism research from 13 groups in Europe and North America, with expertise in the autism phenotype, statistical genetic analysis, and high-throughput methods of gene discovery. To achieve our goals, the AGP has a powerful resource of 1,496 multiplex families (two or more affected individuals in a family) and ~2,000 trios (one affected individual and both parents). The multiplex family samples have been genotyped and analysed for linkage using a set of 10000 single nucleotide polymorphisms (SNPs) spread across the genome. Analysis of the SNP data using a qualitative autism diagnosis has identified regions on chromosomes 2q, 5p, 7q, 9p, and 11p that likely harbour variants increasing susceptibility to autism.

Our study design will take four parallel, complementary approaches to identify autism susceptibility genes. We will integrate these approaches in a staged design of experiments that maximises our chance of success while being cost effective. We will (1) follow-up our top four-five regions of linkage (2) use the richness of the autism phenotype to help us understand how genetics and different aspects of the clinical picture are related (3) search for changes in copy number (deletions and duplications) throughout the genome that may also be causative for autism and (4) analyse candidate genes for mutations and association with autism. Variants in genes emerging from linkage and fine-mapping, copy number analysis, and data from outside sources will be evaluated in the second and third years of the project through our staged experimental design. The large cohort of AGP trio families will be used as an independent sample for validation of significant findings obtained in the multiplex families. Identifying susceptibility alleles and a proper understanding of the underlying genetic and molecular mechanisms will be key to identifying any interacting environmental factors, and for developing rational preventative and treatment strategies to improve quality of life.

Autism Spectrum Disorders (ASDs) are lifelong neurodevelopmental diseases affecting about 0.6% of the population with significant economic, health and well-being costs to society and families involved. Twin and family studies point to a strong genetic predisposition involving epistatic interactions between multiple loci. The behavioural phenotype is variable and extends beyond ASDs to include social, communication and repetitive/restricted behavioural abnormalities, alone or in combination, in relatives of normal intelligence. Identifying susceptibility genes will be important for understanding the neurodevelopmental pathways involved, identifying putative environmental factors and for developing specific preventative and treatment strategies. To identify susceptibility genes, the Autism Genome Project (AGP) was initiated to pool resources, and clinical and scientific expertise from four separate consortia involving 13 North American and European sites. The AGP has a large and rich phenotypic data set on multiplex (1496) and trio (2000) families so that specific statistical and genetic approaches can be brought to bear in the search for susceptibility loci and their relationship with phenotypic expression. The AGP has recently completed a 10K SNP-based genome scan for linkage and has provided support for linkage on chromosomes 2q, 5p, 7q, 9p and 11p. We hypothesize that liability to autism is due to rare and more common genetic variation, and that inheritance of multiple combinations of genetic variants result in this disorder.

Our study design will take four parallel, complementary approaches to identify autism susceptibility genes. Each approach will be integrated in an experimental staged design to maximise power and minimise costs. (1) High-density SNP genotyping of the four-five principal regions of linkage from the AGP genome scan will be performed to search for association. (2) Refinement of the autism phenotype will identify auxiliary traits for use in detailed analyses to identify loci that influence expression of autism-related traits. (3) Copy Number Variations (CNVs) identified from dense genotype data will be analysed in the autism sample and a control sample as there is now strong evidence that liability to autism in some individuals is linked to genomic instability. (4) Compelling candidate genes, especially those with prior evidence, functional rationale for involvement in autism and those located in regions of moderate linkage will be evaluated for association to autism. Subsequently, candidate genes emerging from these approaches will be evaluated in a large cohort of AGP trio families in the second and third years of the project through our staged experimental design.