Department of Biostatistics, Columbia University, New York City, NY, USA.
Division of Nephrology Department of Medicine College of Physicians and Surgeons, Columbia University, New York City, NY, USA.
Nat Genet. 2023 Jun;55(6):1057-1065. doi: 10.1038/s41588-023-01392-0. Epub 2023 May 11.
Fine-mapping is commonly used to identify putative causal variants at genome-wide significant loci. Here we propose a Bayesian model for fine-mapping that has several advantages over existing methods, including flexible specification of the prior distribution of effect sizes, joint modeling of summary statistics and functional annotations and accounting for discrepancies between summary statistics and external linkage disequilibrium in meta-analyses. Using simulations, we compare performance with commonly used fine-mapping methods and show that the proposed model has higher power and lower false discovery rate (FDR) when including functional annotations, and higher power, lower FDR and higher coverage for credible sets in meta-analyses. We further illustrate our approach by applying it to a meta-analysis of Alzheimer's disease genome-wide association studies where we prioritize putatively causal variants and genes.
精细映射通常用于识别全基因组显著位点的假定因果变异。在这里，我们提出了一种贝叶斯精细映射模型，该模型相对于现有方法具有多个优势，包括灵活指定效应大小的先验分布、联合建模汇总统计数据和功能注释以及考虑元分析中汇总统计数据与外部连锁不平衡之间的差异。通过模拟，我们将性能与常用的精细映射方法进行了比较，结果表明，在包含功能注释时，所提出的模型具有更高的功效和更低的假发现率（FDR），并且在元分析中具有更高的功效、更低的 FDR 和更高的置信区间覆盖率。我们通过将其应用于阿尔茨海默病全基因组关联研究的荟萃分析进一步说明了我们的方法，在该分析中我们对假定的因果变异和基因进行了优先级排序。
