Section: New Results

Motion Planning for Quasi-Static Simulation

Participants : Leonard Jaillet, Stephane Redon.

Recently, motion planning methods inspired from Robotics have been applied to the study of biological molecular systems [8] . These approaches rely on compact graph representations that aim to capture large amplitude motions more efficiently than classic simulation techniques, despite their lower resolution.

We developed within the SAMSON's architecture a new motion planning strategy to perform quasi-static simulation at the nano-scale.

The user provides as inputs the initial and final state of the system he or she wants to simulate. Then, the method searches a transition path that follows the low-energy valleys of the conformational landscape (see figure 5 ).

The adaptation of motion planning approaches to quasi-static simulation at the nano-scale comes with several challenges. First, these approaches must be adapted to tackle the high dimensionality involved in the case of nanosystems, dimensionality that is directly related to the number of atoms considered. Second, these approaches must be extended to face the complexity of the underlying physics that comes from the various types of interactions between atoms.

The method we propose is able to perform simulations involving bonds breaking. This is, up to our knowledge, the first motion planning approach able to simulate chemical reactions.