Computational Drug Discovery and Design

(backadmin) #1
data of Entrez IDs, log FCs, and adjustedPvalues to a .txt file
named DEgenes.

3.4 Network
Analysis


Over 300,000 experimentally validated molecular interactions are
included in InnateDB [33, 34]. They are curated according to
MIMIx standards [35]. We conduct a network analysis by upload-
ing the DEgenes file to InnateDB (Fig.10).
To determine the relationship among the genes, we conduct
correlation analysis to determine the interactions among uploaded
genes. If we are strict in controlling the data quality, we can return
the curated interactions only. If we would like to find more inter-
actions among uploaded genes, we should include the interactions
predicted by orthology.

3.5 Module
Identification


The network generated by InnateDB is downloaded in XGMML
format and visualized by Cytoscape (Fig.11). The node and edge
attributes include GO terms and number of
supporting publications.
ThroughjActiveModules (seeNote 2)[36], the connected
region with significant changes in gene expression is identified as
a module (Fig.12).
As the size of the first module is larger than the others, we only
select the first module for subsequent analysis.

Fig. 9Gene annotation. (a) R code for gene annotation. (b) Results of gene annotation


190 Sze Chung Yuen et al.

Free download pdf