Inferring the functional effect of gene expression changes in signaling pathways

Abstract

Signaling pathways constitute a valuable source of information that allows interpreting the way in which alterations in gene activities affect to particular cell functionalities. There are web tools available that allow viewing and editing pathways, as well as representing experimental data on them. However, few methods aimed to identify the signaling circuits, within a pathway, associated to the biological problem studied exist and none of them provide a convenient graphical web interface. We present PATHiWAYS, a web-based signaling pathway visualization system that infers changes in signaling that affect cell functionality from the measurements of gene expression values in typical expression microarray case-control experiments. A simple probabilistic model of the pathway is used to estimate the probabilities for signal transmission from any receptor to any final effector molecule (taking into account the pathway topology) using for this the individual probabilities of gene product presence/absence inferred from gene expression values. Significant changes in these probabilities allow linking different cell functionalities triggered by the pathway to the biological problem studied. PATHiWAYS is available at: http://pathiways.babelomics.org/.

Figure 1.

Graphical representation of Wnt pathway as described in the KEGG repository (1). This pathway has a canonical Wnt/β-catenin cascade and two non-canonical pathways, named Wnt/Planar cell polarity (Wnt/PCP) pathway and Wnt/Ca2+ pathway, which conform the complete Wnt signaling pathway (A). Here, we represent the Wnt pathway in a gene expression microarray experiment in which colorectal cancer patients are compared with control subjects (14). In the diagrams, nodes represent proteins and edges represent the type of interactions among them. (A) The results provided by PATHiWAYS in which the stimulus–response circuits with significantly decreased activity in cancer patients with respect to healthy control subjects are highlighted in blue. Solid edges indicate activations and dotted edges indicate inhibitions. (B) Representation, using the cytoscape program, of the individual gene expression activations (in red) or deactivations (in blue) when comparing the cancer samples with respect to the normal control samples. Edges ended in an arrow indicate activations and edges ended in ‘T’ indicate inhibitions. Despite many genes in many circuits increase or decrease their activity (B), only a few circuits are really affected by these changes (A).