Figure 2

Schematic representation of three scenarios of contextual association between pathways. Each illustration (A-C) shows two pathways (sets of both connected and isolated genes inside ellipsoids) with varying degrees of overlapping genes. The size of the circles (genes) represents the level of differential gene expression between control and disease patients (the higher the significance associated with the expression change, the larger the circle). The lines and arrows represent KEGG-derived interactions between genes as annotated in the pooled pathway, and the thick lines represent edges connecting genes from Pathway 1 to Pathway 2. (A) Overlapping pathways with high contextual association. Whereas Pathway 1 and Pathway 2 can be associated because genes A and B are shared between the two pathways, the fact that overlapping genes are directly connected to other differentially expressed genes (thick connections to large circles) enhances this association. (B) Overlapping pathways with low contextual association. The pathway maps are exactly the same as in case (A). However, Pathway 1 is “less” associated with Pathway 2 in the context of gene expression data, as the genes connected by thick edges are modestly differentially expressed (thick connections to small circles). (C) Non-overlapping pathways with high contextual association. Although, Pathway 1 and Pathway 2 do not share any genes, genes from these two pathways are connected based on inter-pathway connectivity inferred from the pooled pathway. Pathway 1 is contextually associated with Pathway 2 because the genes connecting these two pathways are differentially expressed.