This project develops an integrative risk model using ML classification algorithms to analyze and predict AD risk based on genetically-regulated multi-omics data imputed from multi-ancestry genotypes across ADSP. Individual-level transcriptomic and proteomic datatypes were imputed using GTEx v8 and ARIC reference data. Elastic-net regression, random forest, and neural network architectures are being constructed and their predictive performance is being compared across data modalities. Preliminary transcriptomic and proteomic analyses indicate that biologically relevant findings can be obtained from imputed data; these additional data modalities will be used to improve risk prediction for AD.