Mapping the Genotype–Phenotype Relationship in Cellular Signaling Networks: Building Bridges Over the Unknown - Systems Biology: Networks, Models, and Applications - page 140

The expression of each of these catalogued genes is regulated by a
complex interplay of transcription factors and other cellular proteins.
Chromatin immunoprecipitation is used to identify sites on a genome
to which regulatory proteins bind. The proteins are crosslinked to DNA
with formaldehyde and small fragments are precipitated, amplified,
and identified. Recently, a high-throughput implementation of this
technique (called genome-wide location analysis [19]) has enabled
extensive characterizations of the regulatory patterns in yeast [20],
and it is being developed for application to mammalian systems [21].
For example, Zeitlinger et al. found that different conditions result in
the binding of yeast transcription factor Ste12 to distinct promoters in
vivo, which, in turn, cause different sets of genes to be expressed, and
different developmental programs to be specified [22].
In addition, cDNA microarrays can indicate which genes are
expressed under specified conditions. Microarray technology has
progressed to the point where community standards regarding the
depositing and sharing of data have emerged [23]. Although initially
there was criticism regarding the low signal to noise ratio, technical
standards have significantly improved. Microarray data coupled with
other data sources can generate remarkably accurate pictures of net-
work function [24]. Microarrays complete the picture regarding the
genotype of a given signaling network. Genome sequencing, coupled
with expression arrays, delineates the protein participants. Genome-
wide location analysis indicates which regulatory proteins control the
expression of the given genes.
The first stage in mapping the genotype-phenotype relationship is
a precise characterization of the genotype of interest. The methods
highlighted above delineate the genotype, or “available parts list,” for
a given signaling network. Rapid and cost-effective sequencing efforts
may lead to genotyping of individual cells in a population of highly
mutable cells. This precise characterization may generate further
understanding of a disease etiology as complex and diverse as cancer.
Although many efforts to reconstruct signaling networks piece together
global signaling reactions, only a subset of these reactions is ever
expressed in any given system. Genome-wide location analysis may
lead to signaling network reconstructions that correspond to the regu-
latory protein-gene interactions that occur in vivo in a given system.
Advances in sequencing technologies and methods for identifying reg-
ulatory protein-gene interactions connect the signaling network to the
regulatory program.
What questions do these technologies fail to answer? They do not
indicate whether any of these proteins actually interact with each other
in vivo . The approaches described above cannot generate hypotheses
regarding the physical mechanism of their interaction. Furthermore,
they do not account for nonprotein components of signaling networks,