Graduate Student Seminar Set for April 15

The next Beckman Institute Graduate Student is set for Wednesday, April 15 at Noon in Room 1005 of the Beckman Institute. The talks feature: "Membrane Sculpting by BAR Domain Protiens" by Ying Yin, "Stochastic Multiphysics Modeling of Micro-ElectroMechanical Systems (MEMS)" by Nitin Agarwal, and "Molecular Dynmaics Simulations of Villin Headpiece Folding" by Peter Freddolino. A pizza lunch will be served.

The next Beckman Institute Graduate Student is set for Wednesday, April 15 at Noon in Room 1005 of the Beckman Institute. The talks feature: "Membrane Sculpting by BAR Domain Protiens" by Ying Yin, "Stochastic Multiphysics Modeling of Micro-ElectroMechanical Systems (MEMS)" by Nitin Agarwal, and "Molecular Dynmaics Simulations of Villin Headpiece Folding" by Peter Freddolino. A pizza lunch will be served. Abstracts for the talks are listed below.

Membrane Sculpting by BAR Domain Proteins

Ying Yin

Membrane compartments of manifold shapes are found in cells, often sculpted by cellular proteins. In particular, proteins of the BAR domain superfamily participate in membrane sculpting processes in vivo and reshape also in vitro low-curvature membrane liposomes into high-curvature tubes and vesicles, achieving their role by binding with their curved, positively charged surfaces to negatively charged membranes. Recent observations revealed that membranes are shaped actually through the concerted action of multiple BAR domains arranged in a lattice. However, information on the dynamics of membrane bending and an explanation of the lattice's role are still lacking. Here we show by means of coarse-grained molecular dynamics simulations totaling over 1 millisecond, how lattices involving parallel rows of amphiphysin BAR domains sculpt flat membranes into tubes. A highly detailed, dynamic picture of the formation of membrane tubes by lattices of BAR domains over time scales of 100 microseconds is obtained. Lattice types inducing a wide range of membrane curvatures are explored. The results suggest that multiple lattice types are viable for efficient membrane bending.

 *****

Stochastic Multiphysics Modeling of Micro-ElectroMechanical Systems (MEMS)

Nitin Agarwal

Micro-ElectroMechanical Systems (MEMS) have been used in widespread sensing and actuation applications such as micro-switches, gyroscopes, accelerometers, position controllers etc. For the purpose of functionality prediction and reliability assessment of these devices, it is required to accurately model the interaction of various physical fields, such as – mechanical, electrical and fluidic. State-of-the-art design methodologies for MEMS assume that the geometrical and physical properties of these devices are known in a deterministic sense. However, in reality, significant uncertainties in these properties are inevitable, and must be considered during the development of computational models.

To this end, we present an advanced computational framework based on high-dimensional sparse grid interpolation, which allows quantifying the effect of stochastic variations in various design parameters on the performance of electrostatic MEMS devices. In addition to accurately computing important information such as – moments (mean and variance) and failure probabilities, regarding relevant quantities of interest, it also estimates the sensitivities with respect to design variables. Moreover, the uncertainty quantification data can be effectively used to identify critical design parameters, which can then be controlled during fabrication, in order to achieve desired performance. This approach is straightforward to implement, and offers orders of magnitude improvement over traditionally used sampling based Monte Carlo methods.

 *****

Molecular Dynamics Simulations of Villin Headpiece Folding

Peter Freddolino

Molecular dynamics simulations of protein folding can provide very high resolution data on the folding process; however, due to computational challenges most studies of protein folding have been limited to small peptides, or made use of approximations such as G\={o} potentials or implicit solvent models. We have performed a set of molecular dynamics simulations totaling over 50 microseconds on the villin headpiece, one of the most stable and fastest-folding naturally occurring proteins, in explicit solvent. We find that the wild type villin headpiece reliably folds to a native conformation on timescales similar to experimentally observed folding, but that a fast folding double-norleucine mutant shows significantly more heterogeneous behavior. Along with other recent simulation studies, we note the occurrence of non-native structures which may yield a native-like signal in the fluorescence measurements typically used to study villin folding. Based on the wild type simulations we propose alternative approaches to measure the formation of the native state.