Table 3.
Overall model performance of APOE-e4 count, polygenic risk score, and Elastic Net SNP models in dementia genetic prediction in validation of All of Us sample, stratified by genetic inferred ancestry
| HLA (N = 526) | AA (N = 2,644) | ||||
|---|---|---|---|---|---|
| N case | Cases | Controls | Cases | Controls | |
|
| |||||
| N | 81 | 445 | 181 | 2,463 | |
| Model | AUPRC | AUROC | AUPRC | AUROC | |
|
| |||||
| APOE | e4 count | 0.425 (0.39, 0.468) | 0.64 (0.62, 0.67) | 0.352 (0.317, 0.39) | 0.603 (0.573, 0.632) |
| Best single AD PRS | AFR gene-annotated | 0.395 (0.34, 0.484) | 0.62 (0.58, 0.68) | 0.347 (0.299, 0.404) | 0.599 (0.549, 0.646) |
| Best SNPs | Gene-annotated Neuro SNPs | 0.475 (0.384, 0.533) | 0.69 (0.61, 0.73) | 0.371 (0.328, 0.414) | 0.628 (0.591, 0.66) |
Abbreviations: AA, African Americans; AD, Alzheimer’s Disease; AFR, African American; APOE, apolipoprotein E; AUROC, Area Under the ROC Curve; AUPRC, Area Under the Precision-Recall Curve; HLA: Hispanic Latino Americans; PRS, Polygenic Risk Score; SNP, Single-Nucleotide Polymorphism.