Fig. 1

Illustrated overview of PIGNON. Step 1 PIGNON builds a graph representation of the protein–protein interaction (PPI) network. Here, every node is a protein, and the edges between them a PPI. Edges can be weighted based on the fold-change (FC) of the protein expression measured between condition 1 (c₁) and condition 2 (c₂). Step 2 PIGNON annotates the proteins of the PPI network using Gene Ontology terms, and in parallel, the network is annotated with a shuffled set of these annotations. Here, annotations are represented by the various colours (blue, yellow and green). The clustering of proteins associated with a given annotation is measured using the Total pairwise distance (TPD). A Monte Carlo sampling approach followed by an approximation using normal distributions is then performed to assess the clustering statistical significance. Step 3 We assess the clustering confidence through a false discovery rate estimation. Step 4 Significantly dysregulated GO terms in the expression-weighted PPI network are reported following filtering of significant GO terms identified in the unweighted PPI network