Premium
Polygenic analysis of late‐onset Alzheimer’s disease in a Japanese population
Author(s) -
Kikuchi Masataka,
Miyashita Akinori,
Hara Norikazu,
Shigemizu Daichi,
Ozaki Kouichi,
Niida Shumpei,
Ikeuchi Takeshi,
Nakaya Akihiro
Publication year - 2020
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 - 10.1002/alz.041351
Subject(s) - genome wide association study , single nucleotide polymorphism , apolipoprotein e , genetic association , medicine , population , disease , biology , genetics , genotype , environmental health , gene
Background Late‐onset Alzheimer’s disease (LOAD) is a neurodegenerative disease characterized by cognitive impairment and has a high heritability of 60%–80%. The genetic factor, with the strongest influence identified to date, that is associated with the increased risk of LOAD is the APOE ε4 allele. However, the APOE alleles only explain approximately 6% of LOAD phenotypic variance, and other genetic factors may be responsible for the remainder. The polygenic model explains complex traits, including diseases based on multiple single nucleotide polymorphisms (SNPs) with weak effects as well as several dozen SNPs identified by genome‐wide association studies (GWASs). Recent studies have reported that the polygenic risk score (PRS), which is an index that quantifies polygenic effects, increased the ability to discriminate LOAD patients and correlated with the amount of amyloid β in cerebrospinal fluid. In this study, we calculated PRSs for approximately 2,000 Japanese people using the European LOAD GWAS statistics provided by a previous study. To examine the discrimination ability for LOAD patients, we compared ones between PRSs calculated using the LOAD GWAS statistics and PRSs computed using the GWAS statistics in other diseases. Additionally, we also compared the discrimination ability by LOAD PRS with that of APOE ε4 variation only. Method PRS was calculated for each individual as the weighted linear sum of the genotypes of SNPs and the GWAS statistics, including the p‐values or the odds ratios. In this study, we utilized the GWAS statistics from psychiatric disorders, including schizophrenia, and non‐psychiatric disorders, including type 2 diabetes. We performed bootstrapping to evaluate the discrimination ability for LOAD patients and simulation analysis to discuss the differences between populations. Result The predictive powers of LOAD were significantly higher than those from other diseases. With GWAS statistics of the same population, the accuracy prominently increased. Conclusion We demonstrated the polygenic effects of LOAD in the Japanese population. Our simulation analysis suggested that PRS constructed based on LOAD GWAS statistics in the same population increased the discrimination ability for LOAD patients.