Dataa | EvaluationMetricb | Cutoffsc |
---|
100 | 200 | 300 | 400 | 500 |
---|
1 | RMSE | 90.3(27.5)d | 90.9(28.8) | 90.9 (29.2) | 90.8 (28.8) | 95.8 (23.8) |
Cor | 0.13(0.06) | 0.11(0.12) | 0.11 (0.14) | 0.11 (0.14) | 0.10 (0.13) |
2 | RMSE | 48.7(13.7) | 49.4(12.9) | 49.0 (12.9) | 48.7 (12.8) | 50.1 (14.3) |
Cor | 0.19(0.08) | 0.12(0.10) | 0.15 (0.06) | 0.17 (0.05) | 0.04 (0.20) |
3 | RMSE | 48.0(7.2) | 47.6(7.0) | 47.5 (6.9) | 46.9 (7.0) | 47.0 (6.9) |
Cor | 0.04(0.08) | 0.07(0.09) | 0.07 (0.10) | 0.13 (0.10) | 0.12 (0.12) |
- aData 1: Pretreatment DNAm data to predict the triglyceride levels measured at visit 2; Data 2: Pretreatment DNAm data to predict the triglyceride levels measured at visit 4; Data 3: Posttreatment DNAm data to predict the triglyceride levels measured at visit 4
- bRMSE root mean square error, Cor Pearson correlation between observed and predicted values
- cThe top number of CpG sites selected based on interindividual variability
- dThe averaged RMSE or Cor value and their SD from the three splits of training and test sets. The bold value indicates the model has the best performance across a several number of selected CpG sites at the given DNAm data set and performance metric