A1947
Title: fSKAT: A set-based association test for functional responses in genetic studies
Authors: Jingyeong Jeong - Chungnam National University (Korea, South) [presenting]
Eunjee Lee - Chungnam National University (Korea, South)
Abstract: Set-based association testing is widely used in genetic studies, but most existing methods are designed for scalar or low-dimensional outcomes. In biomedical and Neuroimaging studies, phenotypes are often functional data, such as trajectories, signals, imaging profiles, or connectivity curves indexed by time, space, or anatomical distance. Functional SKAT (fSKAT), a score-type variance-component test for associations between genetic variants and a functional response, is proposed. fSKAT captures genetic effects that are weak at individual points but accumulate across the functional domain. The method is based on function-on-scalar regression and constructs a kernel-based statistic from residual functions under the null model, with inference based on a mixture-of-chi-square approximation. Simulations show controlled type I error and improved power over scalar-summary approaches when genetic effects are distributed across the functional domain. fSKAT is applied to resting-state fMRI and genetic data from the Alzheimer's Disease Neuroimaging Initiative (ADNI), using seed-based functional connectivity curves as a motivating phenotype. For each subject and seed region, the curve is indexed by Euclidean distance to other brain regions, and curve values represent functional connectivity. The ADNI analysis examines genetic associations with early connectivity changes before clinical diagnosis and demonstrates the utility of fSKAT for complex functional phenotypes.
