Premium
O2‐06‐04: Neuron‐Specific Methylome Analysis Reveals New Pathomechanism in Alzheimer’s Disease Brains
Author(s) -
Iwata Atsushi,
Mano Tatsuo,
Ohtomo Ryo,
Nagata Kenichi,
Saido Takaomi C.,
Yamashita Satoshi,
Ushijima Toshikazu,
Hashimoto Tadafumi,
Iwatsubo Takeshi,
Tamaoka Akira,
Tsuji Shoji
Publication year - 2016
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2016.06.427
Subject(s) - dna methylation , biology , neun , microbiology and biotechnology , cpg site , neuron , neuroscience , genetics , immunohistochemistry , gene , gene expression , immunology
Background: Using genome-wide common variant genotype data from the Alzheimer’s Disease Sequencing Project (ADSP), a Presidential Initiative on the study of Alzheimer’s Disease (AD), we estimated the genetically regulated portion of brain gene expression level using the program PrediXcan. This approach uses models that combine the predictive value of eQTLs established by the Genotype-Tissue Expression Project (GTEx) to impute gene-based expression levels from common genetic variants genome-wide. Methods:We estimated the genetically regulated expression levels ofw13,000 genes in a total of 41 tissues. We focused our preliminary assessments on findings in 8 brain tissues in 171 Caucasian family-based study participants with genotypes derived from the whole genome sequencing data. In this initial cohort, we required prediction models to have greater than 0.10 correlation with gene expression and include multiple polymorphisms. Results:We identified 31 genes whose predicted expression level is suggestively or significantly associated with AD after genome-wide multiple test correction. These include genes involved in angiogenesis (SNX11, p value 3.8 x 10), neuronal development (SGOL2, p value 3.1 x 10), and retromer dysfunction (CCDC53, p value 1.9 x 10). Conclusions: To build on these preliminary analyses several expansions of the approaches are underway. We are analyzing cosegregation of differentially predicted expression levels across AD cases and controls to increase power by leveraging pedigree structures. In addition we have detected distant cryptic relatedness across pedigrees with PADRE and are utilizing distant relationships to inform these analyses. Finally, we are currently predicting expression genome-wide in 10,341 unrelated AD cases and controls in all 41 GTEx tissues to validate these findings and identify new genes whose genetically regulated expression is significantly associated with AD risk.