Affiliations: [a] CAS Key Laboratory of Genome Sciences and Information, Beijing Institute of Genomics, Chinese Academy of Sciences, Beijing, China | [b] University of Chinese Academy of Sciences, Beijing, China | [c] Center of Alzheimer's disease, Beijing Institute for Brain Disorders, Beijing, China | [d] Department of Neurology, Xuanwu hospital, Beijing, China | [e] UC Davis Genome Center and Department of Biomedical Engineering, Davis, CA, USA
Correspondence:
[*]
Correspondence to: Hongxing Lei, CAS Key Laboratory of Genome Sciences and Information, Beijing Institute of Genomics, Chinese Academy of Sciences, Beijing 100101, China. Tel./Fax: +86 10 84097276; E-mail: [email protected]; Yong Duan, UC Davis Genome Center and Department of Biomedical Engineering, One Shields Avenue, Davis, CA 95616, USA. Tel.: +1 530 754 7632; E-mail: [email protected].
Abstract: Meta-analysis of data from genome-wide association studies (GWAS) of Alzheimer's disease (AD) has confirmed the high risk of APOE and identified twenty other risk genes/loci with moderate effect size. However, many more risk genes/loci remain to be discovered to account for the missing heritability. The contributions from individual singe-nucleotide polymorphisms (SNPs) have been thoroughly examined in traditional GWAS data analysis, while SNP-SNP interactions can be explored by a variety of alternative approaches. Here we applied generalized multifactor dimensionality reduction to the re-analysis of four publicly available GWAS datasets for AD. When considering 4-order intragenic SNP interactions, we observed high consistency of discovered potential risk genes among the four independent GWAS datasets. Ten potential risk genes were observed across all four datasets, including PDE1A, RYR3, TEK, SLC25A21, LOC729852, KIRREL3, PTPN5, FSHR, PARK2, and NR3C2. These potential risk genes discovered by generalized multifactor dimensionality reduction are highly relevant to AD pathogenesis based on multiple layers of evidence. The genetic contributions of these genes warrant further confirmation in other independent GWAS datasets for AD.
