Table 3. Comparison of kinship estimation with and without off-target data in WES of 762 Chinese and Malays.
Dataset | Method | Unrelated (289,205 pairs) | 3rd degree (148 pairs) | 2nd degree (147pairs) | PO/FS (437 pairs) | Self-kinship (762 individuals) | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
RMSE | BIAS | RMSE | BIAS | RMSE | BIAS | RMSE | BIAS | RMSE | BIAS | ||
Target | SEEKIN | 0.006 | -0.001 | 0.007 | -0.001 | 0.008 | -0.003 | 0.010 | -0.004 | 0.018 | -0.005 |
PC-Relate | 0.005 | 0.000 | 0.007 | -0.001 | 0.008 | -0.004 | 0.012 | -0.008 | 0.022 | -0.017 | |
Target + off-target | SEEKIN | 0.003 | -0.001 | 0.003 | 0.001 | 0.004 | 0.001 | 0.007 | 0.003 | 0.017 | 0.002 |
PC-Relate | 0.002 | 0.000 | 0.004 | -0.004 | 0.009 | -0.009 | 0.019 | -0.019 | 0.026 | -0.021 |
Evaluation was based on SNPs overlapped with the SGVP dataset in the BEAGLE+1KG3 call set of 762 individuals for both SEEKIN and PC-Relate. 40,824 SNPs within target regions and 1,054,229 SNPs across target and off-target regions were included in the analyses. RMSE is the root mean squared error and BIAS is defined as the mean difference to the array-based estimates from PC-Relate for each type of relatedness. Negative values of BIAS suggest underestimation for results based on sparse sequencing data and vice versa.