event

ChBE Spring2007 Seminar Series

Primary tabs

Baron Peters, a Post-doc at CECAM ENS-Lyon, presents Obtaining Reaction Coordinates by Likelihood Maximization as part of ChBE's spring seminar series.

* Refreshments will be served at 3:30 PM in the Lower Level 1 Gossage Atrium
* Lecture commences at 4:00 PM in L1255 in the Ford ES&T Building

Seminar Abstract
When a few strong bonds are broken or formed in a reaction, unstable modes at saddle points on
the potential energy surface are good reaction coordinates. However, complex processes like
protein folding and nucleation involve the disruption and reorganization of many weak bonds
each of strength comparable to kT. Reaction coordinates for these complex processes are best
described by collective variables, but finding the correct collective variables is a major
challenge. The standard approach is to guess a reaction coordinate, and then test whether the
coordinate is correct. The trial and error approach is tedious, unreliable, and computationally
expensive. Aimless Shooting, a new approach related to transition path sampling, samples
reactive trajectories independent of the reaction coordinate. Based on data harvested during
Aimless Shooting, likelihood maximization can rapidly determine the best reaction coordinate
from a list of many candidate collective variables. The first method to quantify reaction
coordinate error for complex systems will also be presented. These new methods quantified for
the first time the trade-off between surface area and size of critical nuclei in the Ising model.
Additionally, the new methods identified the reaction coordinate and critical nuclei in the
polymorph transition of terephthalic acid.

Status

  • Workflow Status:Published
  • Created By:Josie Giles
  • Created:05/20/2010
  • Modified By:Fletcher Moore
  • Modified:10/07/2016