Impact of SNP Space on Polygenic Risk Modeling for Alzheimer’s Disease

Our findings underscore the potential of using biologically relevant SNP subsets to simplify models while maintaining performance. They also indicate that future PRS models for AD can be effectively developed using whole exome SNPs, providing a practical balance between predictive power and model simplicity. Taken together, these results support the use of more focused variant sets in cases where genome-wide approaches may be computationally intensive or unnecessarily complex.