A Boost for MD Sampling

Simbios researchers meld a powerful combo of tools

Reprinted with permission from Steffen Lindert, Denis Bucher, Peter Eastman, Vijay Pande, and J. Andrew McCammon, Accelerated Molecular Dynamics Simulations with the AMOEBA Polarizable Force Field on Graphics Processing Units, Journal of Chemical Theory and Computation 2013 9 (11), 4684-4691. Copyright 2013 American Chemical Society.In 2011, Simbios researchers reported achieving greater speed and accuracy in molecular dynamics simulations by tying the polarizable force fields from AMOEBA (Atomic Multipole Optimized Energetics for Biomolecular Applications) to OpenMM, the Simbios toolkit that leverages the power of GPUs to accelerate simulations. Despite those gains, additional speedups are desired to generate simulations of these polarized molecules over ever longer timescales—reaching micro- and milliseconds.


A collaboration between Steffan Lindert, PhD, a postdoc from the University of California, San Diego, his advisor, J. Andrew McCammon, and Simbios researchers, has made new inroads by implementing accelerated molecular dynamics (aMD) with the powerful AMOEBA/OpenMM combo. aMD energetically raises regions of the potential energy landscape that fall below a certain cutoff. This lowers barriers between energy wells, allowing more frequent transitions between low energy states and resulting in enhanced sampling of the conformational space. Until Lindert came along, aMD had been implemented in classical (nonpolarizable) simulations in AMBER and NAMD, but not in OpenMM. “OpenMM was the ticket to GPUs and better performance,” Lindert says.


Lindert spent a month last year as a OpenMM visiting scholar working with Peter Eastman, PhD, in the lab of Vijay Pande, PhD, at Stanford University. And the collaboration paid off: Lindert and the Simbios team found that the synergies between the aMD method and the AMOEBA force field conserve AMOEBA’s accuracy while improving sampling efficiency by two to three orders of magnitude. These results, which were published in the Journal of Chemical Theory and Computation in October 2013, suggest that the aMD sampling method as implemented in OpenMM should effectively support studies that require more extensive (micro- to milli-second) biomolecular sampling.  



The AMOEBA-aMD implementation takes full advantage of GPU computing and is publicly available in OpenMM at http://wiki.simtk.org/openmm/VirtualRepository.


To learn more about the 2014 OpenMM Visiting Scholars program, visit http://simbios.stanford.edu/OpenMMVisitingScholar.  

Post new comment

The content of this field is kept private and will not be shown publicly.
This question is for testing whether you are a human visitor and to prevent automated spam submissions.
Enter the characters shown in the image.