16171
Maternally Acting Gene Alleles (MAGAs) in Autism: A Meta-Analysis of Two GWAS Study Results

Saturday, May 17, 2014
Atrium Ballroom (Marriott Marquis Atlanta)
W. G. Johnson1, E. Stenroos2 and S. Buyske3, (1)Neurology, Rutgers University - Robert Wood Johnson Medical School, Piscataway, NJ, (2)Neurology, Rutgers University - Robert Wood Johnson Medial School, Piscataway, NJ, (3)Statistics and Biostatistics, Rutgers University, Piscataway, NJ
Background: Maternally Acting Gene Alleles (MAGAs) act in maternal tissues prenatally to alter fetal environment and affect offspring phenotype, independently of whether or not they are inherited by the fetus. At least 70 MAGAs are known to date, mostly in neurodevelopmental disorders. From the mother's perspective, MAGAs are genetic factors, producing proteins and perhaps microRNAs or circular RNAs. From the fetus' perspective, MAGAs are environmental factors. We previously carried out the first two maternal analyses of GWAS data for MAGAs using two autism datasets, the AGRE dataset and later the AGP dataset. We found genome-wide significant peaks in regions of chromosomes 1 and 3.

Objectives: Here, we carried out a meta-analysis of these two results to identify and characterize additional regions of interest.

Methods:  We used the Weinberg log-linear method through a convenient implementation in EMIM. Study-specific results for the two datasets were then meta-analyzed.

Results: In the AGRE dataset, we identified 9 regions of interest; in the AGP dataset we identified 5 regions of interest. In the meta-analysis, we identified 5 new regions of interest. Thus, we have found 19 regions of interest so far, all of maternal origin. Three of these regions contained numerous SNPs in high LD with low p-values. Three of the SNPs with probes with high degrees of Y-homology need to be confirmed with more stringent methods. Two of the regions had been previously implicated in autism, i.e. SHANK2 and CNTN5.

Conclusions:  We have now identified 19 regions of interest for MAGAs in autism. We will now study them further by analysis of other datasets and by re-genotyping with more stringent methods and by and imputation. These studies could lead to identification of DNA variation in these regions whose action may contribute to autism. We hope these studies will lead to better understanding of the pathogenesis of autism, to approaches to identifying risk of autism prenatally or even before the onset of pregnancy, and perhaps to methods of preventing or treating autism at a very early stage.

See more of: Genetics
See more of: Genetics