During plant development the regulatory networks controlling growth and patterning must somehow interfere with physical processes to generate specific shapes. How this is achieved, i.e. how molecules assemble into complex systems with a particular form is not known in any organism. We are addressing this central issue in developmental biology using the shoot apical meristem of the higher plant Arabidopsis, which initiates all the aerial organs of the plant.

As a first step, we are currently trying to link the activity of regulatory genes to specific morphogenetic events at the inflorescence meristem. To this end, we have developed a computational pipeline, aimed at expressing gene function in terms of quantified, geometrical changes in tissue shape. Using this tool, we are monitoring anisotropy and growth rates in gene expression domains and in mutant backgrounds.

In parallel we have started to analyse the molecular basis of shape control. Using a combination of physical, mathematical and biological approaches we have provided evidence for a model where molecular networks would impact on two separable processes. First, we have identified a microtubule control of cell wall anisotropy, which feeds back on local stress and strain patterns. This process seems to define particular morphogenetic events and is coupled to the hormone-based control of overall growth patterns, associated with the rapid outgrowth of organs at particular locations.

Models in the form of virtual flowers are being developed to interpret the data and to propose precise hypotheses regarding gene function. These models, using finite element approaches, are able to express both mechanical and biochemical properties. Using 3D reconstructions of real floral meristems as a template we are currently trying to reproduce the growth patterns observed during flower development, taking into account the mechanistic insight we have obtained.