Table 2.
Top COAD AUC improvements by combining three data types.
| PATHWAY | MEAN GENE | MEAN METHYLATION | MEAN SNP | MEAN 3 TYPES (97.5% LCI) |
|---|---|---|---|---|
| Differentiation in PC12 cells | 0.52 | 0.53 | 0.55 | 0.66 (0.64) |
| Leukocyte transendothelial migration | 0.60 | 0.60 | 0.53 | 0.68 (0.65) |
| Cell adhesion molecules | 0.60 | 0.61 | 0.59 | 0.68 (0.66) |
| BCR | 0.53 | 0.48 | 0.53 | 0.61 (0.59) |
| Apoptosis | 0.62 | 0.56 | 0.53 | 0.69 (0.67) |
| Biopeptides | 0.57 | 0.51 | 0.57 | 0.64 (0.62) |
| Integrin mediated cell adhesion | 0.55 | 0.55 | 0.50 | 0.62 (0.59) |
| Angiotensin II mediated activation of JNK | 0.55 | 0.51 | 0.57 | 0.63 (0.61) |
| Death | 0.68 | 0.60 | 0.63 | 0.73 (0.71) |
| Fc epsilon receptor I-mediated signaling | 0.55 | 0.49 | 0.57 | 0.62 (0.60) |
Notes: The top 10 ranked pathways using the AUC measure in the COAD dataset. “Mean” denotes the mean accuracy of the pathway’s classification of early versus advanced stage over 50 iterations. LCI is calculated as defined in the Materials and Methods section.