Table 2. p-values from the likelihood ratio tests comparing different models.
| CD | CEL | T2D | ||
|---|---|---|---|---|
| x1 | x2 | p-values from LRTa | ||
| ldpred-inf | AnnoPred-inf | 4.4e-15 | 2.8e-6 | 0.011 |
| ldpred-inf | PleioPred-inf | 3.9e-34 | 2.3e-7 | 0.041 |
| AnnoPred-inf | PleioPred-anno-inf | 1.5e-18 | 4.9e-8 | 0.031 |
| PleioPred-inf | PleioPred-anno-inf | 1.8e-9 | 1.9e-8 | 0.017 |
| ldpred-inf | PleioPred-anno-inf | 6.4e-31 | 1.6e-12 | 1.7e-3 |
| ldpred | AnnoPred | 1.3e-5 | 1.7e-5 | 0.066 |
| ldpred | PleioPred | 9.3e-40 | 0.022 | 0.039 |
| AnnoPred | PleioPred-anno | 8.6e-13 | 5.7e-5 | 0.021 |
| PleioPred | PleioPred-anno | 7.7e-3 | 0.014 | 0.45 |
| ldpred | PleioPred-anno | 5.2e-29 | 2.8e-7 | 0.027 |
a Likelihood ratio = -2[logL(x1)—logL(x1 + x2)], where logL(x1) and logL(x1 + x2) is the log likelihood from a logistic regression with case/control status as outcome and x1 and x2 as covariates.