Table 2. Comparison of Cross-Validation Consistency (CVC) and Test Accuracy (TA) between GEE-GMDR and the GMDR method for two simulated continuous traits.
| GEE-GMDRd (Mean±SEM) | GMDR-T1e (Mean±SEM) | GMDR-T2f (Mean±SEM) | ||||
| Model and No. ofloci | CVC | TA | CVC | TA | CVC | TA |
| Digenica: | ||||||
| 1 | 9.010±1.446 | .573±.025 | 8.445±1.886 | .546±.035 | 8.500±1.616 | .546±.027 |
| 2 | 10.000±.000 | . 665±.019 | 9.940±.433 | . 617±.023 | 9.975±.211 | . 618±.021 |
| 3 | 6.680±2.056 | .647±.024 | 6.790±2.147 | .597±.032 | 6.460±2.025 | .594±.029 |
| 4 | 5.550±2.126 | .624±.029 | 5.485±2.141 | .575±.037 | 5.435±2.071 | .576±.031 |
| 5 | 5.230±2.131 | .603±.033 | 5.105±1.911 | .557±.035 | 4.820±2.066 | .554±.032 |
| 6 | 5.050±2.017 | .583±.033 | 4.695±1.844 | .543±.033 | 4.765±1.915 | .542±.034 |
| 7 | 4.970±1.977 | .562±.035 | 4.825±2.165 | .530±.037 | 5.195±1.999 | .535±.034 |
| 8 | 5.340±2.036 | .550±.040 | 5.465±2.140 | .528±.044 | 5.250±2.034 | .526±.038 |
| 9 | 6.705±2.110 | .545±.050 | 6.500±2.088 | .524±.050 | 6.720±2.178 | .533±.052 |
| Trigenicb: | ||||||
| 1 | 8.625±1.700 | .551±.029 | 8.095±2.017 | .533±.034 | 7.720±2.020 | .527±.031 |
| 2 | 7.805±2.017 | .568±.031 | 6.655±2.142 | .536±.034 | 6.535±2.117 | .530±.034 |
| 3 | 9.940±.409 | . 630±.023 | 8.655±2.031 | . 573±.038 | 8.585±2.031 | . 572±.035 |
| 4 | 6.765±2.069 | .607±.027 | 6.320±2.250 | .559±.037 | 6.090±2.110 | .556±.033 |
| 5 | 5.685±1.996 | .582±.030 | 5.210±1.989 | .541±.035 | 5.020±2.096 | .540±.034 |
| 6 | 5.375±2.075 | .565±.033 | 4.720±2.008 | .529±.035 | 4.805±1.922 | .527±.033 |
| 7 | 5.090±2.048 | .544±.035 | 4.585±2.016 | .520±.035 | 4.940±2.041 | .522±.034 |
| 8 | 5.395±2.020 | .534±.041 | 5.460±2.083 | .521±.041 | 5.335±2.169 | .519±.041 |
| 9 | 6.770±2.066 | .535±.052 | 6.545±2.147 | .518±.052 | 6.740±2.247 | .526±.050 |
| Tetragenicc: | ||||||
| 1 | 8.060±1.764 | .537±.029 | 7.730±2.076 | .524±.036 | 7.465±2.025 | .537±.029 |
| 2 | 7.000±2.271 | .546±.034 | 6.415±2.284 | .526±.036 | 5.870±2.251 | .518±.035 |
| 3 | 6.330±2.291 | .548±.037 | 5.285±2.302 | .522±.035 | 5.245±2.135 | .519±.033 |
| 4 | 8.580±2.132 | . 587±.037 | 5.725±2.420 | . 528±.041 | 5.715±2.213 | . 526±.034 |
| 5 | 6.105±2.125 | .564±.037 | 4.865±2.126 | .524±.036 | 4.980±1.985 | .521±.030 |
| 6 | 5.190±2.068 | .543±.038 | 4.675±1.995 | .519±.033 | 4.605±1.967 | .517±.035 |
| 7 | 4.860±1.881 | .530±.036 | 4.830±1.881 | .516±.032 | 4.840±2.087 | .515±.035 |
| 8 | 5.280±2.101 | .525±.042 | 5.350±2.109 | .514±.041 | 5.195±2.034 | .511±.038 |
| 9 | 6.845±2.030 | .527±.050 | 6.470±2.126 | .515±.051 | 6.725±2.203 | .521±.050 |
a
The genotypes with two uppercase-letter alleles (i.e., AAbb, AaBb, aaBB) are set as high-risk group and the rest as the low-risk group.
b
The genotypes with three uppercase-letter alleles are set as high-risk group and the rest as the low-risk group.
c
The genotypes with four uppercase-letter alleles are set as high-risk group and the rest as the low-risk group.
d
GEE-GMDR is the GEE-GMDR analysis for the simulated bivariate traits.
e
GMDR-T1 is the unvariate GMDR analysis for trait 1.
f
GMDR-T2 is the univariate GMDR analysis for trait 2.