Re: accelMD parameter choice

From: Chris Harrison (charris5_at_gmail.com)
Date: Sat Mar 24 2012 - 17:23:50 CDT

Dear Grant,

On Sat, Mar 24, 2012 at 5:08 PM, Grant Schauer <grantschauer_at_gmail.com>wrote:

>
> For choosing E, Hamelberg, Mongan, and McCammon suggest in J.Chem.Phys
> 2004 to use an E greater than <V>, the average potential from the starting
> structure. So when I'm only modifying the dihedral potential (accelMDdihe
> on), I take this to mean the average dihedral energy from the output file
> as evaluated during energy minimization- is this correct?
>

This is correct, but consider instead a short MD run instead of an energy
minimization.

>
> For choosing alpha, they also suggest to use alpha=E-<V>. So if the
> average dihedral energy from the first 100 minimization steps is ~1500, and
> I want an an alpha for fairly aggressive sampling, does E= 1600 and
> alpha=100 seem like a reasonable starting point? Or is it important to
> scale alpha to the size of my system as suggested here:
> http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2011-2012/0629.html
> ?
>

In lieu of comparing against differing alphas for your given system, Jeff's
suggestion is the wiser approach than selecting one unscaled alpha and
assuming it's ok. For best results, attempt and compare the results of a
few "trial" accelerated MDs with different alphas.

Best,
Chris

-- 
Chris Harrison, Ph.D.
NIH Center for Macromolecular Modeling and Bioinformatics
Theoretical and Computational Biophysics Group
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave., Urbana, IL 61801
http://www.ks.uiuc.edu/Research/namd       Voice: 773-570-6078
http://www.ks.uiuc.edu/~char                          Fax: 217-244-6078

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:21 CST