15693
Common Polygenic Variations in Autism
Autism Spectrum Disorder (ASD) has a strong genetic component. The inheritance pattern of ASD in most families is complex and caused by multiple genetic and environmental factors. While rare and de novo genetic variation are known to affect liability, a recent study suggests that common genetic polymorphisms exert substantial additive genetic effect on ASD liability and could explain up to 40% of genetic variance of autism in simplex families and 60% in multiplex families.
Objectives:
To better characterize the common genetic background of multiplex and simplex autism and identify an optimal set of single nucleotide polymorphisms (SNPs) that could discriminates affected from non-affected individuals when combined in a polygenic model.
Methods:
A genome-wide association study based on 1,490 children with autism from the Autism Genetic Resource Exchange (AGRE) and the Simons Simplex Collection (SSC) and 1,335 independent controls (Discovery sample) was performed. Genetic risk scores (GSs) were computed from SNPs selected according to varying liberal P-values criteria and evaluated in an independent Validation sample of 452 affected children from the Autism Genome Project (AGP), including multiplex and simplex families and 1,144 independent controls. Using a principal component analysis, we limited the two samples to individuals with a homogeneous European genetic background to avoid any confounding effects. For statistical power consideration, all cases were required to be diagnosed with an autistic disorder. In the Validation sample we explored which SNPs jointly explained most of the genetic variance in multiplex and simplex autism based on P-value criteria and consistent direction of SNP’s allele effect. New GSs were evaluated in an independent Replication sample combining 554 affected individuals without sibling history of ASD from the AGRE, SSC and AGP collections, excluded from previous analysis because of a heterogeneous European genetic background and 1,986 controls.
Results:
The GSs computed in the Discovery sample were significantly associated with autistic disorder and explained up to 6% of the genetic variance in the Validation sample. Exploratory analyses in the Validation sample determined that a subset of 5,284 SNPs could explain up to 24% and 28% of genetic variance in simplex and multiplex autism. Consistency of risk allele between the Discovery and the Validation sample stratified according to multiplex/simplex information was observed for 2,061 SNPs. Because of different genotyping technology, only 1,706 SNPs were available for analysis in the Replication sample. The corresponding GS was significantly associated to autistic disorder (P=1.4x10-169) and explained up to 24% of genetic variance with an estimated 0.82 area under the curve (AUC). Assuming a pre-test risk of 24.8% for children with developmental delays, a GS based on 1,706 SNPs identified children with up to 50% and 72% risk of autism.
Conclusions:
This study provides further evidence for a common polygenic component in autism and a common genetic background between simplex and multiplex autism. Based on these results, we show that a combination of multiple risk-associated common variants in a GS may have the ability to identify children with a higher risk of autism.