Family-based genome-wide association study in autism

Background: Autism has been regarded as the most heritable neuropsychiatric disorder. However, no single major gene has been identified conclusively through linkage or association studies during the past decade suggesting an extensive genetic heterogeneity in its etiology. Recently, several studies have implicated de novo copy number variations in autism.
Objectives: To comprehensively examine the association between genomic structure variations (SNP and copy number) and autistic risk
Methods: The Miami Institute for Human Genomics is processing a genome screen on a completely independent dataset using the Illumina 1-M beadchip. We completed a preliminary analysis of the first 189/600 trios.  All genotypes were called in beadstudio and Plink was used for the TDT analysis. A total number of 189 trios (sample call rate>98%) and 793,380 SNPs per sample (call frequency>95%, genTrain score>0.4 and not in copy number region) are included.
Results: Our results show no SNPs meet a genome-wide significance at this stage. Over 428 SNPs gave a P-value less than 0.001, and these are spread across the genome. Among the top 10 significant SNPs (based on their empirical p-values), none of them fall within previously reported candidate regions or genes.
Conclusions: These data further suggest that the genetic etiology of autism is complex and that no single polymorphism carries a large effect.