TABLE 3.
Simulation results for p = 600,q= 600, and n = 300
| Method | numSNP | Estimation loss | Model error | True positive | MCC |
|---|---|---|---|---|---|
| IVGC | βk = 0.5, k = 1,…,10 | ||||
| 3 | 1.65 (1.08) | 0.73 (0.27) | 10.00 (0.00) | 0.74 (0.14) | |
| 4 | 1.68 (0.91) | 0.84 (0.25) | 9.98 (0.14) | 0.78 (0.15) | |
| 5 | 2.93 (1.76) | 1.25 (0.51) | 9.13 (1.38) | 0.75 (0.15) | |
| βk ~ U(0.5, 1), k = 1,…,10 | |||||
| 3 | 2.59 (0.94) | 1.22 (0.27) | 9.99 (0.10) | 0.74 (0.12) | |
| 4 | 2.62 (1.04) | 1.37 (0.36) | 9.91 (0.45) | 0.78 (0.13) | |
| 5 | 4.36 (2.64) | 1.93 (0.70) | 9.15 (1.42) | 0.75 (0.18) | |
| IV | βk = 0.5, k = 1,…,10 | ||||
| 3 | 2.95 (1.21) | 1.13 (0.25) | 9.84 (0.39) | 0.70 (0.14) | |
| 4 | 3.28 (1.21) | 1.31 (0.28) | 9.63 (0.56) | 0.73 (0.15) | |
| 5 | 4.86 (1.23) | 1.77 (0.35) | 7.71 (1.03) | 0.64 (0.15) | |
| βk ~ U(0.5, 1), k = 1,…,10 | |||||
| 3 | 3.86 (1.48) | 1.61 (0.35) | 9.90 (0.30) | 0.71 (0.13) | |
| 4 | 4.45 (1.10) | 1.86 (0.35) | 9.66 (0.59) | 0.72 (0.12) | |
| 5 | 6.91 (1.69) | 2.64 (0.44) | 7.87 (1.08) | 0.65 (0.16) | |
Abbreviation: IV, instrumental variable; IVGC, IV regression with graph-constrained regularization.
The numbers in the parentheses are the empirical SE.