Table 1 Main results of the model comparisons
From: Using LDpred2 to adapt polygenic risk score techniques for methylation score creation
Model | p-value | Nagelkerke R2 | AIC | Probes clustered* |
|---|---|---|---|---|
P + T (Tesfaye et al. 2024) | 3.49 × 10− 8 | 0.0347 | 1597.7 | 12% |
LDpred2, CMR blocks | 0.432 | 0.000687 | 1628.8 | 18% |
LDpred2, CMR blocks, no singletons | 8.39 × 10− 7 | 0.0275 | 1604.3 | 18% |
LDpred2, CMR sliding window blocks** | [1.83 to 1.88] × 10− 8 | 0.0361 to 0.0362 | 1596.3 to 1596.5 | 74–86% |
LDpred2, TAD blocks | 1.68 × 10− 8 | 0.0363 | 1596.2 | 99% |
LDpred2, random cluster blocks | [2.02 to 2.79] × 10− 8 | 0.0352 to 0.0359 | 1596.6 to 1597.2 | 99% |