Network models can be used in combination with experimental data to dissect drug mode of action and for drug repositioning. (A) In transcriptional networks nodes are individual genes and edges represent pair-wise functional or regulatory interactions. These networks can be “reverse-engineered” from gene expression profiles (GEPs) with different computational methods or derived from literature. Transcription network models can be used to filter for GEPs following drug treatment in order to infer the primary targets causing the observed ranscriptional changes. (B) Protein interaction networks can be used to model signaling pathways, where edges imply phosphsorylation/de-phosphorelation events. Signaling network models can be inferred from phosphoproteomic data. These models can be used to simulate in-silico the drug effects on signal transduction. (C) Drug similarity networks describe similarities between drugs, such as similar transcriptional responses or similar adverse-reaction. Drug networks can be easily inferred from gene expression profiles following multiple drug treatments.