Figure 8

Reproducibility of top-scoring pairs and triples. Panel A: Each point represents a candidate gene pair. For each pair, the probabilities of the two orderings are estimated on the combined expObreast and Miller datasets and the classifier is tested on Sotiriou. The x-coordinate of a point represents the training accuracy and the y-coordinate represents the testing accuracy. Panel B: Same as left panel, but for all triples of genes.