Figure 1

Overview of the main logical steps of the proposed method. I) Top left: the original graph representing functional interactions between genes. White nodes correspond to labeled examples (genes belonging to a given CM), gray nodes to unlabeled examples. II) Top right: the "augmented" graph obtained by applying a 2-step random walk kernel. Red edges represent the new connections between genes added by the random walk kernel. III) Bottom left: genes scoring. The score function is applied to 4 coloured nodes: the weights of the edges (outlined in boldface) connecting the coloured nodes to the labeled "positive" nodes are added to obtain the scores associated to each coloured node. IV) Bottom right: gene ranking. The coloured nodes are ranked according to the previously computed scores.