composed of multiple crosstalk pathways that even contain over-
lapping regulatory loops, is highly challenging. We proposed a
modeling strategy that combines ODE-based and logic-based
models to accommodate a large-scale, nonlinear system of interac-
tions that we here refer to as “hybrid models” [72].2.4.3 Hybrid Model Hybrid models combine different modeling formalisms to handle
systems that contain multiple aspects; for example discrete and
continuous, linear and nonlinear dynamics. The biological system
of cell cycle is an appealing example that contains discrete and
continuous aspects [73, 74]. Alfieri et al. modeled the cell cycle as
a hybrid system using hybrid automata where the R-point transition
was modeled as a discrete event while the mitogenic stimulation of
the system was realized as a continuous state by ODEs [74]. In
Khan et al. we proposed hybrid modeling formalism that combines
the feature of ODEs and logic-based frameworks provide an effi-
cient solution to model large-scale, non-linear biochemical net-
works [72]. The network is organized and divided into different
parts with distinctive regulatory features (Fig.7) and each part is
modeled with suitable modeling formalism. For instance,
sub-networks that enriched with feedback and feed-forward loops,
and which are therefore expected to display a highly nonlinear
behavior are modeled using ODEs, whereas the target gene module
that contains activation or inactivation regulation of dozens to
hundreds of genes is modeled using logic-based formalism. Hybrid
model provides good compromise between quantitative/qualitative
accuracy and scalability when considering large networks.
2.5 Integration
of Omics Data
Recent advances in the high-throughput techniques made it possi-
ble to measure spatio-temporal genomics, transcriptomics, proteo-
mics, and metabolomics data in the context of complex diseases.
Most omics technologies are already at impressive level regarding
data quality, robustness, time, and cost efficiency. Integration of
Omics data with biochemical disease networks has been shown to
acquire better insights of system-wide impact of perturbation and
therapy in the progression and management of the diseases
[75]. Among various Omics datasets, analysis of gene expression
data and its integration with biochemical disease networks is the
most fundamental process to answer questions like to which degree
a gene is active in the process under investigation, what environ-
mental changes alter its expression, which cellular processes are
associated with it, and in case of deregulation, which diseases can
be caused or mediated by this particular gene. Thus, integration of
expression data on the network lets us develop a clearer picture of
the role of the genes in the disease progression. In case of cancer,
genes that function as tumor suppressors can cause tumorigenesis if
their production (expression) is reduced; on the other hand, the
increase of production of oncogenes may have similar effects.Integrative Workflow for Predicting Disease Signatures 263