Integration of polygenic risk with single cell methylation profiles for depression

Abstract

Large scale genome-wide association studies (GWAS) have identified hundreds of risk loci for major depression disorder (MDD) with their functional understanding being largely unknown. We integrate MDD polygenic risk from GWAS with methylation at a single cell level resolution to gain insights into the role of methylation in driving MDD risk. We introduce a new approach that leverages the polygenic risk of disease with single-cell methylation data to provide a methylation single cell disease relevance score (met-scDRS) for every cell in a single-cell methylation-seq experiment. We analyzed human atlas single cell methylation data to find 54.0% of layer 2/3 intratelencephalic (L2/3-IT) neurons and 46.5% of layer 5 extratelencelphalic (L5-ET) neurons in the dataset showing significant met-scDRS enrichment. We identified gradient of met-scDRS from inferior temporal gyrus to middle temporal gyrus and variations in posterior to anterior brain axis within L2/3-IT neurons. Met-scDRS identifies functional pathways such as synaptic cellular component, somato-dendritic compartment, post-synapse, cell junction organization that are implicated in diseases and identifies genes that are more disease associated. We contrasted met-scDRS for MDD across 75 other traits including brain, immune/blood, metabolism, and other trait categories to identify diverging and converging cell types and prioritized pathways across different traits. Finally, we demonstrated that met-scDRS is portable across non-CpG and CpG methylation data in providing robust signal.