Using rare cancer predisposition alleles derived from The Cancer Genome Atlas (TCGA) and high cancer prevalence (14% of participants) in All of Us (version 6), we assessed the impact of these rare alleles on cancer occurrence in six broad groups of genetic similarity provided by All of Us: African/African American (AFR), Admixed American/Latino (AMR), East Asian (EAS), European (EUR), Middle Eastern (MID), or South Asian (SAS). We observed that germline susceptibility to cancer consistently replicates in EUR-like participants but less so in other participants. We found that All of Us participants from the EUR (p = 1.8 × 10-7), AFR (p = 0.018), and MID (p = 0.0083) genetic similarity groups who carry a rare pathogenic mutation are more likely to have cancer than those without a rare pathogenic mutation. With the advent of combining medical records and genetic mutations, we also performed a phenome-wide association study (PheWAS) to assess the effect of pathogenic variants on additional phenotypes. This analysis again showed several associations between predisposition variants and cancer in EUR-like participants, but fewer in those of the other genetic similarity groups. As All of Us grows to 1 million participants, our projections suggest sufficient power (>99%) to detect cancer-associated variants that are common, but limited power (∼28%) to detect rare mutations when using the entire cohort. This study provides preliminary insights into genetic predispositions to cancer across a diverse cohort and demonstrates the value of All of Us as a resource for cancer research.
Keywords: MeSH; diverse genetics; genetic predisposition to disease; genotype-phenotype association; neoplasms; phenome-wide association study.
Copyright © 2025. Published by Elsevier Inc.