HASE:Framework for efficient high-dimensional association analyses

Abstract

Large-scale data collection and processing have facilitated scientific discoveries in fields such as genomics and imaging, but cross-investigations between multiple big datasets remain impractical. Computational requirements of high-dimensional association studies are often too demanding for individual sites. Additionally, the sheer size of intermediate results is unfit for collaborative settings where summary statistics are exchanged for meta-analyses. Here we introduce the HASE framework to perform high-dimensional association studies with dramatic reduction in both computational burden and storage requirements of intermediate results. We implemented a novel meta-analytical method that yields identical power as pooled analyses without the need of sharing individual participant data. The efficiency of the framework is illustrated by associating 9 million genetic variants with 1.5 million brain imaging voxels in three cohorts (total N=4,034) followed by meta-analysis, on a standard computational infrastructure. These experiments indicate that HASE facilitates high-dimensional association studies enabling large multicenter association studies for future discoveries.