Table 1.
SNV | Data set | Method | BIAS | SEE | SSE | RMSE |
---|---|---|---|---|---|---|
47956424 (MAF = 0.3435) | Simulated phenotype | “True model” | 0.166 | 1.206 | 1.121 | 1.131 |
“Full model” | 0.663 | 4.215 | 4.280 | 4.320 | ||
BAC | 0.440 | 1.587 | 1.277 | 1.347 | ||
Q1 | “True model” | 0.006 | 0.996 | 1.105 | 1.102 | |
“Full model” | 0.025 | 3.483 | 3.591 | 3.582 | ||
BAC | 0.089 | 1.313 | 1.203 | 1.203 | ||
47908815 (MAF = 0.0026) | Simulated phenotype | “True model” | 1.617 | 3.771 | 3.844 | 4.161 |
“Full model” | 5.115 | 6.89 | 6.739 | 8.447 | ||
BAC | 3.438 | 5.322 | 5.329 | 6.331 | ||
Q1 | “True model” | 0.129 | 3.115 | 3.113 | 3.108 | |
“Full model” | 0.369 | 5.694 | 5.621 | 5.619 | ||
BAC | 0.179 | 4.419 | 3.964 | 3.958 |
BIAS is the difference between the mean of estimates of ACE and the true value; RMSE is the root mean square error; SEE is the mean of standard error estimates; SSE is the standard error of the estimates of ACE
Results are based on 200 simulated phenotypic or Q1 data sets. In simulated phenotypic data, the true ACE of SNV at position 47956424 (47908815) is −6.094 (−7.732). In Q1 data, the true ACE of the two SNVs is zero