Figure 1

Illustration of the difference between all-vs-all and one-vs-all discriminative subnetworks. Illustration of the difference between all-vs-all and one-vs-all discriminative subnetworks. The aggregate expression profiles of five hypothetical subnetworks are shown. Red and green respectively represent positive and negative expression, with intensity representing magnitude. The subnetworks ${\mathcal{S}}_1,{\mathcal{S}}_2,{\mathcal{S}}_3$ and ${\mathcal{S}}_4$ are one-vs-all discriminative, respectively indicating classes (CRC stages) A, B, C and D, since the expression profile of each subnetwork in samples that belong to the respective class can discriminate these samples from other classes. On the other hand, ${\mathcal{S}}_5$ is an all-vs-all discriminative subnetwork, since it discriminates all classes from each other.