Friday, May 16, 2008
Champagne Terrace/Bordeaux (Novotel London West)
Background: Autism is a severe neurodevelopmental disorder with a strong genetic component. Despite numerous genome screens and individual candidate gene studies, the underlying genetic etiology remains largely unknown. Increasing evidence suggests that autism is more genetically complex than previously thought, and that single gene approaches toward dissecting autism genetics may not be informative. We are taking the alternative approach of testing for interactive effects of multiple genes within the serotonin pathway.
Objectives: To test serotonin related genes for association with autism.
Methods: We tested 75 SNPs within 13 different genes related to serotonin including TPH1, TPH2, HTR1A, HTR2A, HTR3A, SLC6A4, SLC7A5, YWHAZ, and DDC. SNPs were chosen to represent the linkage disequilibrium patterns across each gene, and included when possible common coding variants. The dataset consists of 151 multiplex families and 252 parent-child trios collected at two centers in the southeast United States. Initial analyses included single locus family-based association tests, considering both parental and proband gender. Subsequent analyses examined explicitly for gene-gene interactions using multifactor dimensionality reduction (MDR). Results: Single locus analyses generated marginally significant results for YWHAZ and HTR3A, however, only the HTR3A result survives correction for multiple comparisons. Preliminary two-way interaction analysis with MDR did not identify any significant interactive effects; higher-order interaction analyses are ongoing.
The HTR3A result is interesting and suggests at least a modest effect of this gene on autism. The lack of a strong two-locus interactive effect suggests that either interactions among these genes do not exert a strong effect on autism, or the effect requires a higher order interaction.