is organized and represented in a form of machine readable net-
work, which is the formalized representation of a large number of
individual experimental results. The network is then modeled with
suitable modeling formalisms, which after calibrating with experi-
mental data should recapitulate the biological phenomena under
consideration, is used to formulate or validate hypotheses, and is
used to support the design of new experiments. In this way, the
systems biology approach cycles into data-driven modeling and
model-driven experimentation. More specifically, we can divide
this approach into four main stages:
2.1.1 Setting the Context This step starts with the formulation of a biological question that is
investigated. For example, a general question could be what the
regulatory mechanism(s) underlying prostate tumor metastasis or
drug resistance are? This defines the project boundaries and gives
directions to collect information from the literature and databases.
This information is then converted into a machine readable format
(i.e.,in the form of network) for computational analysis. With the
help of domain experts or using computational methods, project-
specific network components or modules are then chosen.
Refine/Define
a question
Choose network
visualization
Specify &
hypothesize
mechanism(s)
Collect
information
Choose
network
components
Choose
type of
model
Design
new
experiments
Validate
model and refine
hypothesis
Analyze
the model
Specify
biochemical-
/physical
interaction
Mathematical
representation
Identify
parameter values
The
Systems Biology
Approach
Fig. 2The systems biology approach: an iterative process of data-driven modeling and model-driven
experimentation
250 Faiz M. Khan et al.