Biology Reference
In-Depth Information
Fig. 2. Basic network structure implemented in Cell Illustrator. The basic regulatory module used to build this model consists
of a transcription reaction producing an mRNA, followed by a translation reaction producing a protein. A protein can form
complexes through a binding reaction, and either single proteins or protein complexes can act as activators of transcription. All
entities involved in the simulation have a degradation reaction associated.
(Fig. 2). The network was then made using the underlying HFPNe architecture in Cell Illustrator by first
drawing the entities involved on a canvas, and then connecting them by intermediary processes such as
transcription, translation, activation, inhibition and degradation. Finally speed rules were set to reproduce
biologically meaningful simulations. All entities can receive any floating-point value, and all reactions
can receive any given speed rule [Doi
et al.
, 2004; Nagasaki
et al.
, 2004]. As additional elements,
the network included binding reactions between proteins and the degradation of the resulting protein
complexes. Regulatory connections (activation, inhibition) were made between transcription factors
and the transcription of mRNAs (treated as reactions), or, in the case of the miRNA miR172, between
the miRNA and the regulated translation reaction. When evidence of different regulatory elements
was available, independent transcription reactions were added corresponding to each of the regulatory
elements in play; as an example,
AG
is known to be regulated by LFY and MADS transcription factor
complexes, so in the final model
AG
had independent transcription reactions for each regulating factor
or complex. Competition between different MADS complexes for the same sites in positive regulatory
interactions was implemented by assuming that the sites can become saturated, and the production of
mRNA would reach a certain maximal transcription speed, beyond which different levels of competing
transcription factors would make no further contribution to the process. In order to reflect this, a threshold
value for transcription speed was set on transcription reactions regulated by common factors in order to
make the different contributions additive only before the saturation threshold was reached.
The network includes static entities, whose concentration does not change during the simulation,
and dynamic entities, which are actually playing an active role in the simulation. The flowering time
complexes FD/FT and SOC1/AGL24 are static entities which are set as the starting point of the network
(Fig. 3). Their concentration is set to a constant value of 1, and their activity is set to act as a pulse,
activating the downstream entities only during a short time interval (see methods for details). All other
entities in the network with the exception of the spatial regulators SUP, UFO and miR172 are dynamic:
the meristem identity genes
AP1
/
CAL
,
LFY
and
AP2
are the first entities activated in the simulation,
followed downstream by the organ identity genes
SEP
,
AP3
,
PI
and
AG
. The production of these organ
