Xue, Blue, Conomos, Fohner (2024) The power of representation: Statistical analysis of diversity in US Alzheimer’s disease genetics data
We determined the available array and sequencing data for US studies of Alzheimer’s disease (AD) by reported race and ethnicity, then simulated the statistical power of these datasets to identify risk variants comparable to those found in typical AD genome-wide association studies focused on White or European Ancestry populations. We found there is insufficient statistical power (probability <80%) to detect variants with low to moderate effect sizes using array data from Asian, Black, or Hispanic participants, while available and projected sequencing data from Black and Hispanic participants have statistical power to detect common variants with moderate effect sizes. This study demonstrates a wide discrepancy between proportional representation and statistically meaningful representation. We urge AD genetics researchers and institutions to continue investing in recruiting diverse participants and use community-based participatory research practices.