eTable 1 in Supplement 2 presents the results of each PGS analysis in the screening stage. For the high-risk analysis, the best PGS was calculated using 9154 SNVs having P < .01 and LD r2 < 0.1 within 250 kb when comparing the top 5% of samples with the remaining 95% (odds ratio [OR], 2.75 [95% CI, 2.05-3.71]; P = 2.34 × 10−11). For the low-risk analysis, the best PGS was calculated using 99 135 SNVs having P < .50 and LD r2 < 0.2 within 500 kb when comparing the bottom 25% of samples with the remaining 75% (OR, 0.53 [95% CI, 0.47-0.60]; P = 2.88 × 10−26). However, using 9154 SNVs provided the best result in the high-risk group, had a comparable result in the low-risk group (for the bottom 45%: OR, 0.61 [95% CI, 0.55-0.67]) (eTable 2 in Supplement 2), and used a smaller number of SNVs; therefore, this PGS was used in the testing stage. The list of 9154 SNVs and their weights is presented in eTable 2 in Supplement 2.
