Investigating Complex Molecular Processes Via Likelihood Maximization and the Finite Temperature String Method

Bernhardt Trout
Massachusetts Institute of Technology

The tools of chemical reactivity have for the most part been worked out for reactions involving simple systems. These reactions are associated with smooth reaction surfaces like the one on the left below and include reactions of small molecules in the gas phase, on surfaces, and in solutions in which the solvent does not play a direct role in the reaction. Chemical engineers, however, are interested in a wide range of complex processes, which have rugged reaction surfaces like the one on the right below. Our research group focuses on the development and application of molecular computational tools to study complex processes of industrial relevance. In this talk, I present some of the new methodologies that we have developed and show how we have applied these methods to nucleation and crystallization of small organic molecules and to the stabilization of biopharmaceuticals. Both of these application areas are of importance to pharmaceutical companies. Our methods are currently, however, not applicable to highly diffusive problems, for which we have teamed up with Giovanni Ciccotti and Eric van den Eijnden to apply and further develop the finite temperature string method. We discuss application of this method to such complex problems as protein conformational transformations and crystal nucleation.

Back to Rare Events in High-Dimensional Systems