- Network biology makes use of the tools provided by graph theory to represent and analyse complex biological systems.
- There are several types of biological networks: genetic, metabolic, cell signalling etc. This course focuses on protein-protein interaction networks (PPINs), but the concepts and tools presented can be used in other networks too.
- Networks are represented by nodes and edges. Nodes represent different entities (e.g. genes or proteins) and edges convey information about how the nodes are linked.
Protein-protein interaction networks
- PPINs have a number of characteristics, mainly:
- Small-world effect: Network diameter is usually small (~ 6 steps), no matter how big the network is.
- Scale-free: A small number of nodes (hubs) are lot more connected than the average node.
- Transitivity: The networks contain communities of nodes that are more connected internally than they are to the rest of the network.
- Several tools are available for PPIN analysis. For example, Cytoscape, igraph, Gephi, NetworkX.
- When building a PPIN it is important to be aware of the type and quality of the data used. Confidence scoring tools such as MIscore can help select the best characterised interactions.
- Two of the most used topological methods to analyse PPINs are:
- Centrality analysis: Which identifies the most important nodes in a network, using different ways to calculate centrality.
- Community detection: Which aims to find heavily inter-connected components that may represent protein complexes and machineries
- Annotation enrichment analysis is a complementary tool often used when analysing PPINs. It uses resources such as the Gene Ontology (GO) or Reactome to infer which annotations are over-represented in a list of genes or proteins. It can produce complex results that can be further simplified using network analysis tools.