Genome-Wide Association Study of the Frailty Index - Atkins et al. 2019

Genome-wide summary statistics from the GWAS analysis of the Frailty Index in UK Biobank participants of European descent aged 60 to 70 years. Preprint available https://doi.org/10.1101/19007559 We used a Frailty Index (FI) based on the accumulation of deficits model (Searle et al. 2008), as validated in UK Biobank previously for its ability to predict all-cause mortality (Williams et al. 2018). The FI was derived using 49 self-reported baseline data variables in UK Biobank. Variables were based on a variety of physiological and mental health domains, and included symptoms, disabilities and diagnosed diseases, which were self-reported by participants at baseline. The FI was generated using a complete-case sample with information on all 49 individual components and presented as a proportion of the sum of all deficits. The FI was quantile normalised (i.e. transformed into a normal distribution) prior to the genome-wide association study (due to the skew of the untransformed trait). The analysis included 164,610 participants of European descent aged 60 to 70 with complete Frailty Index data. Results are from UK Biobank genetic data release version 3 (including imputation from HRC and the combined UK10K and 1000 Genomes panels) including 16.4million genetic variants that met the following criteria: minor allele frequency (MAF) >0.1%, Hardy-Weinberg p-value >1x10-9, and imputation quality >0.3. We used the BOLT-LMM (v2.3.2) software used for the GWAS itself (Loh et al. 2015), which uses linear mixed-effects modelling to account for genetic relatedness and confounding by ancestry. Models included age, sex, assessment centre (22 categories), and genotyping array (two categories: Axiom or BiLEVE) as covariates. If used please cite paper Atkins et al. 2019 "A Genome-Wide Association Study of the Frailty Index Highlights Synaptic Pathways in Healthy Aging" Fields are as follows: SNP: dbSNP name of genetic marker, if available CHR: chromosome BP: base-pair position on CHR (hg19 / b37) ALLELE1: effect allele ALLELE0: non-effect allele A1FREQ: frequency of ALLELE1INFO: Imputation quality BETA: effect size from BOLT-LMM approximation to infinitesimal mixed model with respect to ALLELE1 SE: standard error of effect size P_BOLT_LMM: non-infinitesimal mixed model association test p-value

本数据集包含针对欧洲血统、年龄60至70岁的英国生物银行（UK Biobank）参与者的虚弱指数（Frailty Index, FI）开展全基因组关联分析（Genome-Wide Association Study, GWAS）得到的汇总统计数据。预印本可访问：https://doi.org/10.1101/19007559。

我们采用基于缺陷累积模型的虚弱指数（Searle et al. 2008），该指数此前已在英国生物银行中得到验证，可用于预测全因死亡率（Williams et al. 2018）。本研究的虚弱指数源自英国生物银行中的49项自我报告的基线数据变量，这些变量涵盖多种生理与心理健康维度，包含受试者在基线时自我报告的症状、残疾与确诊疾病。

虚弱指数通过包含全部49项组成成分信息的完整病例样本计算得到，以所有缺陷总和的占比形式呈现。在开展全基因组关联分析前，我们对虚弱指数进行了分位数正态化处理（即转换为正态分布），以修正未转换性状的偏态分布问题。本分析共纳入164610名欧洲血统、年龄60至70岁且具备完整虚弱指数数据的参与者。

研究结果源自英国生物银行遗传数据第3版（包含基于人类参考组合HRC、UK10K与1000 Genomes联合面板的基因型填充数据），共纳入1640万个符合以下标准的遗传变异：次要等位基因频率（minor allele frequency, MAF）>0.1%、哈迪-温伯格平衡p值>1×10^-9、基因型填充质量>0.3。

我们采用BOLT-LMM（v2.3.2）软件开展全基因组关联分析（Loh et al. 2015），该软件通过线性混合效应模型校正遗传相关性与祖先混杂因素。分析模型纳入年龄、性别、评估中心（共22个类别）以及基因分型芯片（共2个类别：Axiom或BiLEVE）作为协变量。若使用本数据集，请引用论文Atkins et al. 2019 "A Genome-Wide Association Study of the Frailty Index Highlights Synaptic Pathways in Healthy Aging"。

各字段含义如下：
SNP：遗传标记的dbSNP名称（若可获取）
CHR：染色体编号
BP：对应染色体上的碱基对位置（hg19 / b37版本）
ALLELE1：效应等位基因
ALLELE0：非效应等位基因
A1FREQ：ALLELE1的等位基因频率
INFO：基因型填充质量
BETA：基于BOLT-LMM近似无限混合模型得到的ALLELE1效应量
SE：效应量的标准误
P_BOLT_LMM：非无限混合模型关联检验的p值