Evolution of biomolecular networks — lessons from metabolic and protein interactions

Abstract

Biomolecular networks, such as protein–protein interaction (PPI) or metabolic networks, organize the 'parts lists' generated by various large-scale approaches and are therefore frameworks that facilitate many discoveries in molecular biology. Nodes represent proteins (specifically enzymes in metabolic networks), whereas PPIs in PPI networks and enzyme–enzyme interactions through shared metabolites in metabolic networks are considered as links. Using this framework, general knowledge on the topology of networks can be applied. However, specification of biomolecular networks, such as the impact of the environment and experimental conditions, also have to be taken into account. The experimental conditions raise important issues regarding the accuracy and coverage of these networks, which also have an impact on the conclusions about the evolution of the networks. The evolutionary dynamics of PPI and metabolic networks is mostly based on two classes of genetic events. The first is duplication and loss of regions encompassing complete genes, which implies the addition and loss of nodes and links. The second is more fine-tuned and includes point mutations, small insertions or deletions, and mutations that affect the regulation of genes, which implies the addition and loss of links. Owing to different biological functions and distinct topological features of PPI and metabolic networks, changes of nodes and links in each are subject to different selection. So far, most of the research on networks is devoted to in vitro and static networks, and these are usually considered in two dimensions (2D networks) — that is, without spatial (3D) or temporal (4D) resolution. Many network features and their evolution can be understood only when taking spatiotemporal resolution into account.