TABLE III.
Predictors with replicable effects and their importance ranks in data sets 1 and 2 from Random Forests arid RuleFit
| Rank in Random Forests |
Rank in RuleFit |
|||
|---|---|---|---|---|
| Predictor | Data set 1 |
Data set 2 |
Data set 1 |
Data set 2 |
| Age | 1 | 1 | 1 | 1 |
| Serum glucose | 2 | 2 | 3 | 3 |
| BMI | 3 | 4 | 2 | 2 |
| HDL-C | 4 | 3 | 14 | 4 |
| Fibrinogen | 8 | 8 | 5 | 7 |
| Homocysteine | 7 | 5 | 13 | 5 |
| log(Lp(a)) | 11 | 9 | 8 | 12 |
| Systolic blood pressure | 5 | 6 | 7 | 25 |
| log(triglyceride) | 6 | 7 | 15 | 20 |
| Cholesterol | 10 | 11 | 28 | 14 |
| Sex | 12 | 16 | 20 | 16 |
| LDL-C particle size | 13 | 14 | 9 | 41 |
| NOS3_rs1800780 | 16 | 47 | 31 | 34 |
| GPR35_rs3749172 | 19 | 41 | 38 | 45 |
The rank is based on the total of 304 variables including 287 tagSNPs and 17 risk factors. BMI, body mass index; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SNP, single nucleotide polymorphism.