Section: New Results

Kinematics-inspired algorithms for macromolecular modeling

At a geometric level, RNA is much more flexible than protein, and undergoes smooth transitions between its various conformations. Such transitions are difficult to observe, but can be predicted using algorithms inspired by kinematics and motion-planning. With our partners at Stanford, we designed and implemented such an algorithm within the KGS library [8] to morph between two RNA conformations while keeping distance constraints induced by base pairs and, more importantly, avoiding clashes. In a more preliminary work, we also used similar approaches to automagically fit multi-conformer ligand models into electron density maps [16].