47. STRATIFYING SCHIZOPHRENIA RISK LOCI THROUGH GENETICALLY REGULATED CEREBROSPINAL FLUID PROTEIN ABUNDANCE

Background Schizophrenia (SCZ) is a highly heritable psychiatric disorder with a complex genetic architecture and extensive pleiotropic effects. However, the biological mechanisms linking genetic risk loci to disease-relevant molecular pathways and clinical manifestations are not fully understood yet. Methods In this study, we integrated the lasted SCZ genome-wide association study summary statistics [1] with cerebrospinal fluid (CSF) proteomic data [2] to stratify SZC risk loci into protein-regulatory modules and characterize their biological functions. Results We first identified 250 CSF proteins whose genetic regulation colocalized with SCZ risk loci. We then applied a soft clustering framework developed by Smith et.al [3] to classify 316 SCZ-associated genetic variants according to their predicted regulatory effects on these 250 proteins. This analysis identified seven protein-regulatory modules: c1) a mitochondrial/stress-associated regulatory module; c2) a neurodevelopment-associated module; c3) an MHC-complement inflammatory module; c4) an extracellular matrix/barrier inflammatory microenvironment module; c5) a synaptic/AKT/NMDA-associated module; c6) a neuroendocrine-associated module; c7) a developmental/metabolic-associated inverse-regulatory module. To further characterize the functional relevance of these modules, we constructed cluster-specific partitioned polygenic scores (pPSs) using published GWAS summary statistics. We found psychiatric traits showed widespread positive associations across multiple clusters, consistent with shared genetic liability among neuropsychiatric disorders. Beyond psychiatric traits, we observed distinct module-specific cross-trait associations. The c3 module showed strong associations with complement C4 and monocyte count, consistent with an MHC-centered complement immune axis. In contrast, c7 exhibited a distributed multi-locus association pattern, including negative associations with serum lipid traits and coronary artery disease, suggesting a metabolic pleiotropic axis. Discussion Together, these preliminary findings suggest that protein-regulatory stratification of SCZ risk loci can capture heterogeneous biological projections of SCZ genetic risk beyond a single pathway. Ongoing analyses will further assess module stability and function, locus-level contributions, and robustness to major genomic regions such as the MHC.