Table 3. Gene-gene interaction models for polymorphisms in NER pathways for hepatocellar cancer risk by MDR analysis.
| Model | Training Bal. Acc. | Testing Bal. Acc. | Sign Test (p) | CV Consistency | P for permutation test |
|---|---|---|---|---|---|
| XPC rs18701343 | 0.5798 | 0.5803 | 8 (0.0547) | 10/10 | 0.0190 |
| ERCC1 rs2298881-XPC rs1870134b | 0.5974 | 0.5977 | 9 (0.0107) | 10/10 | 0.0010–0.0020 |
| ERCC1 rs2298881-ERCC1 rs3212961-XPC rs1870134 | 0.6061 | 0.5689 | 9 (0.0107) | 4/10 | 0.0640 |
| ERCC1 rs2298881-XPC rs1870134-ERCC2 rs238417-ERCC5 rs873601a | 0.6218 | 0.6001 | 10 (0.0010) | 10/10 | 0.0010–0.0020 |
Note: The best model was selected as the one with the maximum testing accuracy and maximum CV Consistency.
a
In this study, the best interaction model was the four-factor model including ERCC1 rs2298881-XPC rs1870134-ERCC2 rs238417-ERCC5 rs873601 polymorphisms.
b
, the second best interaction model was the two-factor model including ERCC1 rs2298881-XPC rs1870134 polymorphisms.