The University of Queensland, Australia

Shoot Branching: From Biological Discovery to Predictive NetworksShoot branching has emerged as a powerful model for understanding how genes, hormones, metabolites and environmental signals are integrated to control a plant process. This presentation begins with historical and recent discoveries that illustrate how biological networks are assembled through the combined application of molecular genetics and physiology. Over several decades, shoot branching has become a well-characterised regulatory network with around. As such, it provides an ideal case study for exploring how biological knowledge can be transformed into predictive models.

Using shoot branching as a case study, we built a work flow and R-package (PSoup) and a resulting network representing interactions among key molecular components, hormones, sugars and aspects of the environment from a wide array of diverse studies over many years and laboratories. The resulting model achieved high predictive accuracy across diverse perturbation datasets and identified areas where current biological knowledge remains incomplete. Building on this foundation, we developed FLASH-P, an AI-enabled platform that automates literature mining, network construction, validation and refinement. Processes that previously required months or years can now be completed in minutes or an hour while maintaining transparency and traceability to the underlying evidence.

Finally, we are connecting these biological networks to functional-structural and crop growth models, creating a pathway from molecular knowledge to whole-plant and agricultural prediction. This integration offers new opportunities for hypothesis generation, model-guided experimentation and crop improvement, illustrating how detailed biological understanding can be translated into predictive frameworks for both fundamental and applied plant science.

In Arabidopsis thaliana, the specific bacterial-type ferroxidase LPR1 and the single ER-resident P5A-type ATPase PDR2 form a key regulatory module linking root growth to Pi availability through Fe-dependent signaling. LPR1 enzyme activity triggers profound changes in root meristem function in response to local root tip Pi sensing, while PDR2 maintains cellular homeostasis and restricts excessive stress responses. Our comparative genetic and biochemical studies of Arabidopsis mutants defective in local Pi sensing, ER stress response, and autophagy indicate that ER stress-dependent autophagy is rapidly activated as part of the developmental response of the root apex to Pi limitation and implicates PDR2/AtP5A as a central mediator of this process. Recent biochemical and genetic data on PDR2/AtP5A orthologues in S. cere­visiae and C. elegans uncovered ER-resident P5A-type ATPases to act as trans­mem­­brane helix dislocases, which facilitate the extraction of mistargeted tail-an­chored mem­brane proteins, to enforce ER protein quality control. To dissect the link between PDR2/AtP5A and the autophagic pathway and to understand the significance of Pi limitation-induced autophagy in plants, we used an untargeted approach to monitor global changes in the proteome of root tips. In addition to core autophagy proteins, which are known to be involved in and degraded by autophagy, our results indicate that local Pi sensing via LPR1 activity acts as a potent inducer of ER stress and ER stress-mediated autophagy, and therefore specifically influences proteostasis. The comparative proteomic data in combination with the recently obtained Pi-dependent PDR2-interactome point to PDR2 as a major hub for coordinating Pi sensing, ER quality control, and ER stress-activated autophagy in Arabidopsis root tips.

Contact : François Barbier