Table 2.
Top predictors for lung adenocarcinoma
| Consensus cancer genes | All TCGA genes | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| C2.CP | C6 | C2.CP | C6 | ||||||||
| # | Predictor | W | # | Predictor | W | # | Predictor | W | # | Predictor | W |
| 1 | TP53 | 0.784 | 1 | TP53 | 0.672 | 1 | KEAP1 | 0.766 | 1 | KEAP1 | 0.942 |
| 2 | SMARCA4 | 0.559 | 3 | MET (CNV) | 0.590 | 2 | TP53 | 0.484 | 2 | TP53 | 0.423 |
| 3 | MET (CNV) | 0.445 | 2 | SMARCA4 | 0.541 | 3 | SMARCA4 | 0.339 | 4 | KRAS | 0.397 |
| 4 | KRAS | 0.374 | 4 | KRAS | 0.507 | 4 | KRAS | 0.249 | 3 | SMARCA4 | 0.298 |
| 5 | SETD2 | 0.268 | 8 | EGFR (CNV) | 0.319 | 5 | STK11 | 0.171 | 8 | *BAGE2 | 0.209 |
| 6 | RBM10 | 0.263 | 25 | EGFR | 0.279 | 6 | *DST | 0.166 | 7 | *FRG1B | 0.148 |
| 7 | STK11 | 0.239 | 9 | FOXA1 (CNV) | 0.262 | 7 | *FRG1B | 0.145 | 13 | MET (CNV) | 0.146 |
| 8 | EGFR (CNV) | 0.220 | 10 | MYC (CNV) | 0.241 | 8 | *BAGE2 | 0.131 | 43 | *PKHD1 | 0.137 |
| 9 | FOXA1 (CNV) | 0.213 | 44 | EP300 (CNV) | 0.229 | 9 | *SPTA1 | 0.127 | 36 | EGFR | 0.130 |
| 10 | MYC (CNV) | 0.209 | 11 | SMARCA4 (CNV) | 0.227 | 10 | *ANK2 | 0.115 | 6 | *DST | 0.127 |
Top ten gene-level somatic alteration predictors from models estimated using the TCGA lung adenocarcinoma cohort. The predictors are ranked according to a weight, W, computed as the average across all pathway models in the MSigDB collection of the product the -log(p-value) for the predictor in the unpenalized model and model . Separate rankings are shown for the C2.CP and C6 collections using both potential sets of predictors. The "#" columns contain the rank of the predictor in the list computed using C2.CP collection and the target predictor set. Predictors marked in bold represent known driver genes for lung adenocarcinoma. Predictors prefixed with a "*" are not in the COSMIC cancer gene census