Analyses of plasma multi-omic data across ancestries identify novel pathways implicated in Alzheimer's disease

INTRODUCTION: Few genetic studies on Alzheimer's disease (AD) have incorporated multiple ancestries and omic datasets to pinpoint actionable AD risk effectors for each ancestry. METHODS: Here, we first performed genetic colocalization between molecular phenotypes (proteomics and metabolomics) from two ancestral groups (European [EUR] and African [AFR]) and the two largest EUR AD genome-wide association studies. We next performed pathway enrichment analyses to identify biological mechanisms. RESULTS: We found 21 proteins and one metabolite colocalized with AD risk that were shared between the EUR and AFR ancestry groups. We also identified 25% AFR and 60% EUR proteins; 50% AFR and 10% EUR metabolites were unique. The pathway enrichment analyses nominated interleukin-1 production and lipid pathway were shared underlying proteomic and metabolomic findings, respectively. DISCUSSION: Our findings indicate that these four plasma datasets may pinpoint different effectors of AD risk in diverse populations; findings from AFR participants require validation with AFR-based genome-wide association study data. Highlights: For proteomics, 61% of findings for European (EUR) ancestry and 72% for African (AFR) ancestry were not previously reported. For metabolomics, 83% of findings for EUR ancestry and 50% AFR ancestry were not previously reported. Both convergent and divergent pathways were identified in EUR- and AFR-ancestry stratified analyses in either proteomics or metabolomics findings.