Table 1. Regression of phenotype on predicted value for BayesR, BSLMM, LMM and GPRS in WTCCC data.
| Disease | BayesR | BSLMM | LMM | GPRS | ||||
|---|---|---|---|---|---|---|---|---|
| BD | 1.13 | (0.280) | 1.09 | (0.236) | 1.12 | (0.246) | 2541 | (1403.5) |
| CAD | 0.96 | (0.216) | 0.92 | (0.179) | 0.99 | (0.187) | 1529 | (919.0) |
| CD | 0.98 | (0.137 | 1.01 | (0.134) | 1.05 | (0.341) | 35.4 | (15.50) |
| HT | 1.08 | (0.461) | 0.98 | (0.313) | 1.04 | (0.356) | 2124 | (1041.9) |
| RA | 0.99 | (0.080) | 0.98 | (0.096) | 1.07 | (0.330) | 33.1 | (20.2) |
| T1D | 1.00 | (0.037) | 0.99 | (0.096) | 1.03 | (0.169) | 57.3 | (14.8) |
| T2D | 0.92 | (0.265) | 0.94 | (0.188) | 0.99 | (0.322) | 1894 | (789.9) |
For prediction assessment we performed 20 random 80/20 splits for each trait. The values parentheses are standard deviations over 20 replicates.