Modeling and Analysis of Gene Regulatory Networks - Modeling in Computational Biology and Biomedicine - page 46

to stabilization of this transcription factor (hypoxia signaling pathway). As a
consequence, it recognizes Hypoxia Response Elements which are DNA motifs
associated to a series of genes involved in adaptation to low pO 2 . Among these
genes, erythropoietin (or epo ) is a well known inducer of red blood cells production;
thus the low oxygen signal leads to secretion of the EPO protein and ultimately
to an increase in erythropoiesis. This whole process aims at compensating poor
oxygenation. In contrast, under normal pO 2 conditions, this stimulation of the epo

gene is absent.
Another classical example, for bacterial cells, is the lac operon. An operon is a
typical structure of bacterial genomes. It can be viewed as a “pack” of genes that are
regulated by a unique promoter sequence. For instance, the lac operon is composed
of the lacZ , lacY and lacA genes. In an environment with no glucose but with lactose
available, the lac operon genes are transcribed, leading to lactose consumption as
a source of energy. In that case, the level of transcription factors does not change
directly, but the efficiency of transcription is regulated by a lactose repressor protein
(impairing transcription in the absence of lactose) and by a Catabolite Activator
Protein (which favors transcription in the absence of glucose).
These examples illustrate the complexity of gene regulation networks (GRN)
for eukaryotic as much as prokaryotic cells. From an experimental point of view,
biologists can access different intermediaries of these networks: genomic data
(presence/absence of a DNA motif, complete sequence determination, mutations),
mRNA quantification (large scale semi-quantitative screening with DNA arrays,
or more focused and more precise analysis with Quantitative Polymerase Chain
Reaction), quantitative gene activity measurements (luciferase reporter genes),
quantitative protein detection (use of specific antibodies, fluorescent fusion pro-
teins), or even molecular interactions estimation (semi-quantitatively with F orster/
Fluorescence Resonance Energy Transfer, double hybrid, co-precipitation). Dynam-
ics can also be followed thanks to tools like time-lapse microscopy or Fluorescence
Recovery After Photo-bleaching microscopy on living cells (for an overview of
some of these techniques see [ 15 , 28 ]).

2.1.2 Mathematical Modeling

Therefore, large amounts of data, of more or less qualitative nature, are now avail-
able; one of the main challenges of molecular biology is to develop methodologies
for using these data to address biological questions. Because of the complexity of
the networks, it is necessary to design models, describing the dynamical functioning
of the GRN. Indeed, the expression of genes, the concentrations of mRNA and
proteins evolve with respect to time, and possibly converge toward some steady
state, some periodic behavior or some other complex dynamical attractor. The
emergence of these patterns from the dynamical interactions between the elements
of the network, and the comparison with experimental data, will provide new keys
to the comprehension of molecular biology, and enable scientists to solve important
problems.