Cells achieve reproducible outputs while relying on intrinsically stochastic molecular processes. In plants, cell division orientation is accurately predicted before mitosis by a microtubular ring, the preprophase band (PPB), through an unknown mechanism. Using high-throughput microscopy, quantitative image analysis, biomechanics and stochastic modeling, we will investigate how the stochasticity of microtubule (MT) self-organization is used, or filtered out, during PPB formation to sense temporal, geometric and mechanical cues in order to generate a robust placement of cell division planes. In short, we will test whether the PPB acts as a macromolecular mechanosensor. Using statistical mechanics such as Monte Carlo, event-based modeling and mean-field theory, we will assess how stochasticity leads to distinct dynamical states for MT arrays. We will develop biophysical models of dynamic MTs in 3D cells to explore how stochastic MTs self-organize into PPBs and how spatiotemporal cues modulate that process. Using the Arabidopsis shoot apical meristem, we will identify correlations between MT array dynamics, PPB behavior, cell shape, growth and mechanics, cell cycle progression, and cell division plane from statistically representative sample involving hundreds of dividing cells. To challenge the robustness of PPB and cell division in vivo and in silico, we will globally and locally increase variability in these cell parameters with mutants (MT dynamics, wall mechanics, cell cycle, mechanotransduction), inducible lines, mosaics, and micromechanical perturbations. Last, we will unravel the molecular mechanism processing molecular stochasticity to channel cell divisions: we will investigate how the TTP (TON1-TRM-PP2A) complex, a key regulator of PPB formation in connection with the cell cycle, contributes to generate reproducible divisions by monitoring MT self-organization, integrating geometric, mechanical and temporal cues. Implications of this project go beyond cell division robustness: while many cellular pathways are adapted to respond to rapid and discontinuous changes, we expect here to unravel mechanisms managing slow and continuous signals, like shape, growth or tissue-stress. We will also gain insight into the mathematical properties of stochastic extended objects like microtubules, and we will propose a mechanism for cells to perceive directional cues.