Systems Biology (Methods in Molecular Biology)

proteins in each stable state were then highlighted to reveal

sub-networks that are expected to play key roles in the establish-

ment and maintenance of the stable state. Since the activity of each

protein was normalized to range from 0 (minimal activation) to

1 (full activation), the threshold can be selected within a reasonable

range (from 0.4 to 0.6). We found that these thresholds will not

affect the main conclusions. Figure2 showed the sub-networks of

the normal hepatocyte and cancerous hepatocyte when we set the

threshold as 0.4.

These sub-networks then provided means to directly identify a

few key features of the genetic mutation patterns in HCC. Biologi-

cally, genetic mutations can have varying effects on the function of

protein. Mutations that confer enhanced activity were defined as

gain-of-function mutations, while those that reduce or abolish

protein function were defined as loss-of-function mutations. It

was known that some random mutations in cancers were selected

and accumulated in response to phenotypic consequences

[57–59]. We reasoned that proteins that were inactive in the nor-

mal hepatocyte stable state and activate in the cancerous hepatocyte

Table 2
(continued)

Model

results Experimental results

NF-κB (Kupffer) (RELA) Up Down Up Up

IL-1 (IL1B) Up Up Up Up

IL-6 (IL6) Up Up Up Up

EGF (EGFR) Up Up Up Up

VEGF (KDR) Up Un-change Up Down

Ras (KRAS) Up Up Up Down

ERK (MAPK1) Up Up Up Down

GSK-3β(GSK3B) Down Down Down Down

β-catenin (CTNNB1) Up Up Un-change Up

Myc (MYC) Up Up Up Up

P53 (TP53) Up Down Up Down

TGF-β(TGFB2) Up Up Up Up

P21 (CDKN1A) Un-

change

Up Up Up

Relative changes of protein’s activity from stable state A to stable state B, including up, down or unchanged, obtained
from model results were listed in the second column. Relative changes of protein’s activity from normal liver to HCC
obtained from three independent gene expression data were listed in the last three columns]. Model result has agreement
ratios of 64.9%, 78.4% and 67.6% with three independent experimental data respectively)

Endogenous Molecular-Cellular Network Cancer Theory 227