A0756
Title: Heritability modeling of complex functional phenotypes
Authors: Eardi Lila - University of Washington (United States) [presenting]
Keshav Motwani - University of Washington (United States)
Ali Shojaie - University of Washington (United States)
Ariel Rokem - University of Washington (United States)
Abstract: Magnetic resonance imaging has played a key role in defining structural and functional measures of brain connectivity. Univariate heritability models have been extensively used to estimate the portion of observed inter-individual differences in connectivity attributable to genetics. However, precisely characterizing how genetics and environmental factors shape these phenotypes remains a challenge due to their complex nature. A novel variance component model designed to enable computationally efficient heritability analysis of complex phenotypes, such as manifold-valued or functional data, is introduced. The proposed model allows for the estimation of primary modes of variation due to genetic and environmental factors, generalizing well-known tools such as tangent and functional principal components analysis. Its application to brain connectivity data reveals that the primary modes of variation, typically characterized by overall increases or decreases in connectivity levels, arise from more complex structural and functional connectomes influenced by genetic and environmental factors.
