Systems Biology (Methods in Molecular Biology)

metastatic integrative co-expression network is more than that of

the primary integrative network which indicates that a large num-

ber of biological processes and pathways are dysregulated in the

transition from the primary to metastatic state in prostate cancer.

3.3 TF and miRNA
Co-expression
Network Analysis

In order to identify key molecular signatures responsible for the

transition from healthy to primary and from primary to metastatic

phenotypes in prostate cancer, we analyzed two integrative net-

works shown in Fig.10. In particular, we considered two important

network topological parameters: (1) degree centrality and

(2) betweenness centrality of nodes in the network. The degree of

a node is the number of edges connected to the node. Degree

centrality of a node is a local centrality index that indicates hub

nodes which are involved in a large number of interactions with

other nodes in network. Betweenness centrality differs from the

other centrality measures. A node can have quite low degree, and

still with high betweenness centrality depending on the position of

node in the flow of information. In other words, betweenness

centrality shows to what extent a node can serve as a bridge in the

network. This measurement also indicates how much a given node

in a network has control on the interactions of other nodes

[38]. Betweenness centrality for nodeυεGis the fraction of short-

est paths between pairs of nodesi,jεGthat pass through nodeυ.

The betweenness centralityBC(υ) of a nodeυis computed as

follows:

BCðÞ¼υ

X

i,jεG

σijðÞυ

σij

,where i 6 ¼j 6 ¼υ:

In this equation,σij(υ) denotes the total number of shortest

paths betweeniandjthat pass through nodeυandσijdenotes the

total number of shortest paths betweeniandj.

Z-Scores for these two parameters were computed for all the

nodes in the integrative networks of primary and metastatic pros-

tate cancer. Nodes withz-score>2.0 for degree and betweenness

centrality were considered network key elements and may have a

critical effect in prostate cancer initiation and progression. Based on

the node degree and betweenness centrality of highly differentially

expressed and significantly correlated nodes, we selected top three

candidates from each of the networks. These are BCL2, PGR,

HOXD10 in case of primary stage and STAT3, JUN and JUNB

in case of metastatic state of the prostate cancer.

HOXD10 is a nuclear sequence-specific transcription factor. Its

downregulation has been reported in prostate cancer [73]. IGFBP3

that also present in the integrated co-expression network for pri-

mary cancer is one of the HOXD10 downstream target genes.

IGFBP3 is a pro-apoptotic and anti-angiogenic protein, and

induces cell apoptosis [87, 98]. PGR (progesterone receptor) and

Integrative Workflow for Predicting Disease Signatures 267