Table 2 Commonly studied molecular interaction networks

Association networks

Association networks model any kind of relation between molecules (e.g. binding, co-expression and structural similarities). Examples of association networks are gene co-expression networks and protein similarity networks.

Functional networks

Functional networks model functional relations between pairs of molecules (usually genes or proteins). A link implies that both are involved in the same function, process or phenotype. Genetic interaction networks represent interactions where a double mutation leads to an epistatic effect, i.e., worse or better than expected based on the single mutation.

Protein-protein Interaction Networks (PPI Networks)

Protein-protein interaction networks are undirected networks that model protein binding. PPI networks are derived from high-throughput experiments using techniques such as yeast two-hybrid screening, mass spectrometry and tandem affinity purification [2]. Signaling networks are related to protein interaction networks, but their links are directed according to the flow of molecular signals.

Transcription-regulatory Networks (TR Networks)

Transcription-regulatory networks are bipartite networks with one set of nodes representing genes and the other representing transcription factors (TFs). TFs are products of genes (modeled by gene-TF links) whilst genes are regulated by TFs (modeled by TF-gene links). Data for such networks is derived through the process of chromatin immunoprecipitation (ChIP) [34]. Gene regulatory (GR) networks are related to TR networks but contain only genes. Their links represent indirect regulatory relationships.

Metabolic networks

Metabolic Networks are bipartite networks that model the relationships between the chemical reactions that occur in cells and the substrates involved in the reactions (the solid gray lines in Figure 2a). Reduced, non-bipartite metabolic networks containing only metabolites or only reactions are also often studied.