Biology Reference
In-Depth Information
have described a detailed formal definition and the properties of HFPNe [Nagasaki
et al.
, 2004; Nagasaki
et al.
, 2005].
In HFPNe, to bridge the gap between computer science and biology, the Petri net terms of place,
transition, arc, and token are renamed to the more intuitive terms entity, process, connector, and content,
respectively.
Cell Illustrator is a software implementation that includes HFPNe as well as an extended graphical
user interface for building and simulating biological networks. By using Cell Illustrator a researcher
can directly draw a network map using icons to represent Petri net elements (entities, processes and
connectors), assign speed rules to the processes and directly simulate the dynamics of the network. A
plot with the concentration change of the different entities is displayed during simulation time.
Models of the flowering network have been described and simulated in the past, originally as a Boolean
gene network [Mendoza and Alvarez-Buylla, 1998] which included only 10 genes, but was later refined
to a logical network including 15 genes [Espinosa-Soto
et al.
, 2004]. These network analyses could
correctly identify the steady state gene activation patterns for each of the floral organs, and demonstrated
the importance of the network architecture above of that of the initial parameters assumed.
In the model presented here, we include both direct regulation between genes (mediated by a protein)
and the formation and regulatory effect of heterodimeric transcription factor complexes, thus creating a
larger and more complex network. Explicit translation reactions in the model also allow for the inclusion
of post-transcriptional regulation, such as translational inhibition by miRNAs. Individual transcription
reactions dependent upon different regulatory elements can be associated with each gene, enabling us to
distinguish distinct transcription factor binding events, allowing promoter elements bound by different
transcription factors to be separated and providing the ability to model binding-site competition. Different
floral homeotic protein complexes may compete for binding sites, as is suggested by overlapping DNA-
binding preferences of different floral homeotic factors (reviewed in Melzer
et al.
, 2006). Our model
assumes that organ-specific developmental programmes are stabilized by autoregulatory loops involving
all members of a floral homeotic protein complex, which is further supported by our perturbation analysis
results.
METHODS
Modelling and simulation software: Cell Illustrator
The Cell Illustrator software implements the HFPNe architecture with highly tuned modelling and sim-
ulation graphical user interfaces [Nagasaki
et al.
, 2003; 2009b]. Publicly available models created on Cell
Illustrator are maintained in two websites (
http://www.csml.org/
and
http://genome.ib.sci.yamaguchi-
u.ac.jp/˜gon
)
. The genetic network controlling flower development was implemented using the latest
Cell Illustrator Online 4.0 version (
http://www.cellillustrator.com/
[Nagasaki
et al.
, 2009a]).
Construction of the network
The blueprint of the regulatory network was built by compiling information from current literature,
and ensuring that each proposed entity and connector is qualitatively supported by genetic data and,
whenever possible, by molecular data that confirms a direct physical interaction. The basic module used
in this network was composed of a transcription reaction producing an mRNA, connected to a translation
reaction producing a protein and degradation reactions for both of the mRNA and protein products
