Table 2. The χ2 goodness-of-fit test p-values of total tumors with position-based filtering.
| Combination Case | MFaT | Driver Definition | GEV p-value | Pareto p-value |
|---|---|---|---|---|
| ICGC-IntOgen | ICGC | IntOgen | 0.405 | 0.336 |
| ICGC-DoCM | ICGC | DoCM | 0.529 | 0.316 |
| ICGC-Tokheim | ICGC | Bailey | 0.244 | 0.203 |
| COSMIC-IntOgen | COSMIC | IntOgen | 1.000 | 1.000 |
| COSMIC-DoCM | COSMIC | DoCM | 1.000 | 1.000 |
| COSMIC-Bailey | COSMIC | Bailey | 1.000 | 0.672 |
| CHANG-IntOgen | CHANG | IntOgen | 1.000 | 1.000 |
| CHANG-DoCM | CHANG | DoCM | 1.000 | 1.000 |
| CHANG-Bailey | CHANG | Bailey | 0.632 | 0.625 |
Note that in this setting, a lower significance level with higher p-value such as p = 0.50 weakly supports a notion that the observed and theoretical distributions are alike. In this table, p-values are equal or higher in the GEV than the Pareto distribution.