Fig. 1

Workflow of TOPDRIVER. The figure depicts genes by g _i, edge weights by w _i, disease by D _i, and vertex score by S(g _i). Gene similarity matrix were constructed by integration of four datasets that were depicted in upper left corner. The G(N) were built with application of gene similarity matrix, disease-disease similarity matrix, and gene-disease association. The G(N) depicts the discrimination among gene scores by coloring high-scored vertices with red and the other low-scored vertices by yellow. As G’(N) shows, the jagged network was amended using diffusion kernel. Afterward, we select an induce graph using gene-disease association by finding a set of genes related to an intended disease. Finally, we applied a permutation test and p-value cut-off to choose related candidate genes