BACKGROUND: A hallmark of Late-onset Alzheimer's disease (LOAD) is brain tissue atrophy across multiple cortical and subcortical regions. However, studies report no significant genetic correlation between LOAD and brain morphology, indicating that the genetic drivers of neurodegeneration are complex. Resolving this shared genetic architecture is key to uncovering the biological mechanisms linking brain morphology to LOAD. METHODS: We analyzed UK Biobank genotype data (n = 272,513; 57.1 ± 8.0 years; 54.6% female) alongside genome-wide association study summary statistics for LOAD and 89 brain morphological traits. We employed linkage disequilibrium score regression (LDSC), polygenic score (PGS), and Local Analysis of coVariant Association (LAVA) to evaluate genetic relationships at the whole-genome and locus level. The extent of polygenic overlap was quantified using bivariate causal mixture modeling (MiXeR), and specific shared loci were identified via conjunctional FDR. Identified loci were mapped to genes and followed by enrichment analyses to characterize the shared biological pathways. RESULTS: LDSC and regression analyses revealed no significant global genetic correlations between LOAD and brain morphological traits and no associations between PGS and LOAD proxy scores. In contrast, LAVA identified significant local genetic correlations (both positive and negative) across all LOAD-brain trait pairs. MiXeR analysis further revealed a substantial number of shared genetic variants characterized by mixed effect directions, explaining the lack of global correlation. We identified 183 shared loci mapping to 80 distinct genes. Enrichment analyses indicated that these shared genes are involved in biological processes related to cell differentiation, development, and cellular component organization. CONCLUSIONS: Our findings provide evidence of extensive polygenic overlap between LOAD and brain morphology, revealing that the absence of global genetic correlation stems from mixed local effect directions. These shared genetic architectures implicate pathways in cellular development and differentiation, providing a molecular framework for understanding the link between brain morphology and LOAD.
Wu et al. (Thu,) studied this question.