Analysis of Regulatory and Interaction Networks from Clusters of Co-expressed Genes - Clustering Challenges in Biological Network - page 59

potential for identifying experiment-specific properties of the resulting regulatory
and functional networks.
3.2. Analysis of Regulatory Networks
3.2.1. Expression Data
The microarray data was obtained from an experiment that was conducted to ex-
amine the behavior of a bolus injection of corticosteroids upon temporal gene
expression profile of liver in vivo. This dataset was specifically chosen due to the
apriori knowledge that corticosteroids have powerful transcriptionally mediated
effects upon the rat experimental model. The data collection and preliminary anal-
ysis were previously presented in [51]. Male adrenalectomized (ADX) Wistar rats
(Rattus rattus) weighing 225-250 g were obtained from Harlan Sprague-Dawley.
Rats (3) were sacrificed by exsanguination under anesthesia at 0.25, 0.5, 0.75, 1,
2, 4, 5, 5.5, 6, 7, 8, 12, 18, 30, 48, and 72 hr after dosing. The sampling time points
were selected based on previous studies describing Glucocorticoid Receptor (GR)
dynamics and enzyme induction in liver and skeletal muscle. Four cannulated ve-
hicle treated rats were sacrificed as controls (data represented as time zero). Total
RNA was isolated from liver tissue and hybridized to a U34A GeneChip array.
3.2.2. Regulatory Network Construction and Analysis
As a result of the gene selection step on the aforementioned data set describing
the administration of corticosteroid 529 probes were isolated in 12 clusters. These
529 probes corresponded to 454 individual genes of which 339 genes had reliable
sequence information about the promoter region so that transcription factor pre-
diction could be performed.
Transcriptional networks are comprised of the links between pairs of genes:
those which code for transcription factors and the genes that they regulate. These
links are abstractions of the regulatory activity of transcription factors binding on
a given promoter region and either up or down regulate the levels of gene expres-
sion. The construction of this network can be used in a variety of ways such as the
quantification of the aggregate behavior of the system after the perturbation of a
single node, or the identification of possible key intervention points with which to
mediate the response of an organism to an experimental perturbation. Transcrip-
tional networks can be represented by bi-partite networks in which sets of tran-
scription factors are shown to directly regulate a set of genes Fig. 3.3. While tran-
scriptional networks have elements such as feed forward loops, feedback loops,
and input cascades which are not explicitly visible in a bi-partite representation,