Table 2.
Association of polygenic risk scores and VTE risk in the test set individuals of European and African ancestry.
| PRS testing population | |||||||
|---|---|---|---|---|---|---|---|
| European | African | ||||||
| Method | PRS tuning population | PRS | Number of SNPs | AUC (95% CI) | Odds ratio per SD (95% CI) | AUC (95% CI) | Odds ratio (95% CI) |
| (1) LDpred2 trained in EUR | - | LDpred2EUR | 604,741 | 0.6 (0.59, 0.6) | 1.42 (1.39, 1.45) | 0.54 (0.53, 0.56) | 1.17 (1.11, 1.23) |
| (2) LDpred2 trained in AFR | - | LDpred2AFR | 1,184,805 | 0.52 (0.52, 0.53) | 1.09 (1.06, 1.11) | 0.56 (0.54, 0.57) | 1.26 (1.2, 1.33) |
| Combine (1) + (2) | European | aLDpred2_combined EUR | 1,212,566 | 0.6 (0.59, 0.61) | 1.42 (1.39, 1.46) | 0.55 (0.53, 0.56) | 1.19 (1.13, 1.25) |
| Combine (1) + (2) | African | aLDpred2_combined AFR | 1,212,566 | 0.6 (0.59, 0.6) | 1.42 (1.39, 1.45) | 0.57 (0.55, 0.58) | 1.26 (1.2, 1.33) |
| (3) PRSCS trained in EUR | - | PRSCSXEUR | 591,788 | 0.61 (0.6, 0.61) | 1.47 (1.44, 1.51) | 0.56 (0.55, 0.57) | 1.26 (1.2, 1.33) |
| (4) PRSCS trained in AFR | - | PRSCSXAFR | 586,660 | 0.58 (0.58, 0.59) | 1.35 (1.32, 1.39) | 0.58 (0.57, 0.6) | 1.35 (1.28, 1.42) |
| Combine (3) + (4) | European | aPRSCSX_combinedEUR | 598,977 | 0.61 (0.6, 0.62) | 1.48 (1.45, 1.52) | 0.57 (0.56, 0.59) | 1.3 (1.24, 1.37) |
| Combine (3) + (4) | African | aPRSCSX_combinedAFR | 598,977 | 0.6 (0.6, 0.61) | 1.47 (1.44, 1.51) | 0.59 (0.57, 0.6) | 1.38 (1.3, 1.45) |
a
Combined PRSs were generated using the formula α0 + α1PRS1 + α2PRS2 where α0, α1 and α2 are the weights obtained by fitting a logistic regression model with VTE as outcome, PRS1 and PRS2 as explanatory variables using the validation data set. The weights for the considered combination of PRSs can be found at https://github.com/yonhojee/VTE_PRS.