Table 6.

Replication of 422-gene prediction models with different methods

The table presents the results taking in consideration: 1) whether they were part of the training or replication set; 2) the number of genes available for analysis: on top analyses including all 422 genes; on the bottom analyses where all genes were not available (and the number included)

CI of replication datasets with WIDE overlap with TCGA testing set CI: in red. CI of replication datasets with NO overlap with TCGA testing set CI: in green. AUC: area under the ROC curve. CI: confidence intervals. SE: Standard Error. PAM: Prediction analysis for microarrays; PLS: Partial least squares; Lasso: least absolute shrinkage and selection operator