Systems Biology (Methods in Molecular Biology)

2.3 Network
Analysis

After gathering and managing information regarding biological

systems in the form of interactions networks, structural and dyna-

mical analysis provides useful information about the network archi-

tecture and the dynamics of regulatory pathways. The structural

analysis of networks allows the identification of functional modules,

regulatory motifs (including feedback and feed-forward loops), and

node properties (including node degree (ND) and betweenness

centrality (BC)). In biochemical networks, nodes can be protein,

gene, miRNA, etc. Modules are the aggregations of the densely

interconnected neighboring nodes. It is observed that the func-

tionally related nodes are located in close proximity and thus form

functional modules [30, 31]. These functionally related genes

could be associated with the same biological pathway and have

similar effects on certain disease phenotype (which fits to the prov-

erb “guilt by association”) and may be targeted by structurally

similar drugs [32, 33].

Biological networks are enriched in recurring structural pat-

terns called motifs. Network motifs are the interacting patterns that

recur significantly more often than in random networks [4]. These

motifs are sort of small molecular circuitry that the cell uses to

process information and governing dynamic response to external or

internal fluctuations [34, 35]. Feedback and feedforward loops

(FBL and FFL) are the important regulatory network motifs. Feed-

back loops are characterized by direct/indirect inhibition/activa-

tion of a node by its own target, e.g., Fig.4a, bshow the indirect

activation/inhibition of node “X” by its own target. FBLs can be

either positive or negative depending on the parity of negative links

in the loop. If parity is odd the FBL is negative (Fig.4a) and if it is

even then the FBL is positive (Fig.4b). Feedforward loops are

characterized by interactions in which a node is a mutual target of

a node and its target, e.g., in Fig.5a, bwhere “X” regulates “Y,”

and then “X” and “Y” mutually target “Z.” FFLs can either be

X

Y Z

Z

X

Y Z

X

Y Z

X

Y

Z

X

Y Z

X

Y Z

X

Y

Z

X

Y

Fig. 4Representation of all possible feedback loops (either positive or negative)

in a three-node network

254 Faiz M. Khan et al.