During the last decade, network approaches became a powerful tool to describe protein structure and dynamics. Here we review the links between disordered proteins and the associated networks, and describe the consequences of local, mesoscopic and global network disorder on changes in protein structure and dynamics. We introduce a new classification of protein networks into 'cumulus-type', i.e., those similar to puffy (white) clouds, and 'stratus-type', i.e., those similar to flat, dense (dark) low-lying clouds, and relate these network types to protein disorder dynamics and to differences in energy transmission processes. In the first class, there is limited overlap between the modules, which implies higher rigidity of the individual units; there the conformational changes can be described by an 'energy transfer' mechanism. In the second class, the topology presents a compact structure with significant overlap between the modules; there the conformational changes can be described by 'multi-trajectories'; that is, multiple highly populated pathways. We further propose that disordered protein regions evolved to help other protein segments reach 'rarely visited' but functionally-related states. We also show the role of disorder in 'spatial games' of amino acids; highlight the effects of intrinsically disordered proteins (IDPs) on cellular networks and list some possible studies linking protein disorder and protein structure networks. # Mr. Zsolt Hoksza is a high school student of the Fazekas High School (www.fazekas.hu) in Budapest, Hungary, who started his research as a member of the EU Descartes Award winning High School Research Student organization (www.kutdiak.hu) founded by P.C. and offering research opportunities for several thousands of high school students since 1996.
STRUCTURE AND DYNAMICS OF PROTEINS AS NETWORKS
2.A. Definition of protein structure networks.Protein structure networks (also called protein contact networks) form the basal layer of the cellular network hierarchy. At the residue level of coarse-graining, the nodes are amino acids, while links connect amino acids whose inter-distance is below a cut-off (usually 0.4 to 0.85 nm) in the native fold. Protein structure networks may have weighted links instead of distance cut-offs, and may discriminate between individual atoms (like αC or βC atoms). Covalent bonds may be included or excluded in the network representation [1-9].2.B. Key topological properties of protein structure networks. In the small world of protein structure networks any two amino acids are separated by only a few links, and the network diameter grows logarithmically with increasing number of amino acids. This smallworldness promotes the fast transmission of perturbations (conformational changes). Typically, protein structure networks have a Poisson degree distribution rather than the widely characteristic scale-free degree distribution. This means that they have fewer hubs than expected, and amino acid "mega-hubs", having an extremely large number of neighbors do not exist, whic...