Abstract
Alzheimer's disease (AD) risk differs across ancestral populations, yet most genetic studies have focused on Non-Hispanic White (NHW) cohorts. We conducted a multi-population transcriptome-wide association study (TWAS) using whole-blood RNA-seq and genotype data from reported NHW (n=235), African American (AA; n=224), and Hispanic (HISP; n=292) participants in MAGENTA. Using SuShiE for multi-population fine-mapping, we identified credible sets of eQTLs for 8,748 genes and improved fine-mapping precision relative to analyses using fewer populations. eQTL effects were largely shared across populations, with population-specific regulation for a subset of genes. Population-stratified TWAS and sample size-weighted meta-analysis (FUSION + MAFOCUS) prioritized and and fine-mapped nine genes (FDR<0.05, PIP>0.8), including established AD loci (BIN1, PTK2B, DMPK) with consistent effects across populations. Importantly, at BIN1 we fine-mapped regulatory variants associated with gene expression and AD risk beyond the GWAS index SNP-most notably rs11682128, which is only in modest LD with rs6733839 ( r∧2 ≈ 0.34 )-demonstrating that multi-population TWAS can implicate additional functional variants not captured by single-SNP GWAS signals. We also discovered a novel association between COG4 expression and AD in NHW, implicating Golgi apparatus function. Using independent SuShiE-derived models from TOPMed MESA (PBMC), several associations replicated directionally across ancestries, with statistical significance most evident in NHW. Our results show that multi-population fine-mapping improves eQTL resolution and TWAS interpretability, reveals regulatory variants beyond GWAS index SNPs, and underscores the need to expand non-European AD cohorts to resolve shared and population-specific mechanisms.