Massively expedited genome-wide heritability analysis (MEGHA)

摘要: The discovery and prioritization of heritable phenotypes is a computational challenge in variety settings, including neuroimaging genetics analyses the vast phenotypic repositories electronic health record systems population-based biobanks. Classical estimates heritability require twin or pedigree data, which can be costly difficult to acquire. Genome-wide complex trait analysis an alternative tool compute from unrelated individuals, using genome-wide data that are increasingly ubiquitous, but computationally demanding becomes apply evaluating very large numbers phenotypes. Here we present fast accurate statistical method for high-dimensional SNP termed massively expedited (MEGHA) accompanying nonparametric sampling techniques enable flexible inferences arbitrary statistics interest. MEGHA produces significance measures with several orders magnitude less time than existing methods, making heritability-based millions based on individuals tractable first our knowledge. As demonstration application, conducted global local morphometric measurements derived brain structural MRI scans, 1,320 young healthy adults non-Hispanic European ancestry. We also computed surface maps cortical thickness empirically localized regions where were significantly heritable. Our demonstrate unique capability large-scale screening profile construction.