Search | Site Map | Contact

Research: Project

Quantitative modelling of the regulatory network of flowering time genes

Felipe Leal Valentim, Plant Research International, Applied Bioinformatics group

Biochemical and molecular biology approaches have revealed which genes are involved in the control of flowering-time in the model plant Arabidopsis thaliana. As outcome, a comprehensive list of the key genes together with their targets and regulators is now available. However, some knowledge gaps still exist between the upstream regulators and their ultimate effect, e.g. Are the known components and interactions of the network sufficient to predict the dynamics of expression of regulatory genes? How to predict the flowering-time based on molecular expression levels? Previous studies have proposed models of part of the complex transcriptional network involved in flower transition and flower formation but these were based on Boolean network analysis and do not capture quantitative aspects of time, concentration dynamics, affinities, and the role of protein-protein interactions. We propose to study the gene regulatory network that controls flowering time by making use of an ordinary differential equation (ODE) model that allows for accurate prediction of when the plant starts flowering as a function of transcription factor concentrations.

The key genes and their regulatory relationships are identified by literature survey from which it is possible to define the topology of the gene regulatory network. To model the transcription regulation of genes we use Michaelis-Menten functions in the differential equations where the presumed role of protein-protein interactions is explicitly incorporated. For quantitative reliability of ODE models using Michaelis-Menten functions, detailed parameter information is essential. The parameters in the equations address biological features (e.g. the transcript concentration needed to achieve a half-maximum transcription rate, the maximum transcription rate, dimer decay rate into non-functional components and association and protein-protein interaction affinity) which are not fully experimentally elucidated yet. Therefore, the parameters are estimated indirectly by a parameter estimation procedure. The expression of the key flowering-time genes have been measured by RT-qPCR (by Angenent/Immink group - partner 2) to build the dataset which is used to estimate the unknown parameters in the equations. The parameter estimation procedure makes use of an optimization approach to fit the model equations to the data in order to find the free parameters.

Systematic analysis on the measured time-course of mRNA levels of the flowering-time genes are performed in order to obtain useful insight into the dynamics of the genes during floral transition. Then, we assess our model by comparing the predicted flowering time with available plant mutant data (ectopic and knocked-down expression of key regulators). After the modelling framework has been tested and validated, the effect of mutations in the dynamic of the genes are analyzed. The effect of mutations that have so far not been made experimentally are simulated in silico to be possibly experimentally validated. As results of this project, we aim to: (1) propose the ODE model that describes the gene regulatory network of the key flowering-time genes; (2) estimate the kinetic parameters for the flowering-time regulators; (3) make use of the model for simulating unknown mutant phenotypes, thus generating testable biological hypotheses.

Figure 1