Re: Dihedral angle energy in silica block

From: Eduardo (chucruz_at_ks.uiuc.edu)
Date: Mon Mar 01 2010 - 15:46:54 CST

Dear Jindal;

Without the structures (PSF/PDB) or the configuration file (NAMD), it
is going to be very difficult for anybody to figure out what is the
problem.

Since you are using MacKerrell force field, you should also use his
structures, that can be downloaded from:

http://pubs.acs.org/doi/full/10.1021/jp055341j

and also from the CHARMM website. If you still plan to use a
different structure, check carefully the topology of your structure
before performing a simulation. Atom names and bonding terms should
comply with MacKerrell force field; otherwise your simulation will
not run.

On Mar 1, 2010, at 9:11 AM, Jindal Shah wrote:

> Thanks, Eduardo.
>
> I am actually simulating crystobalite and adopting the FF of
> MacKerell as was done in Nangia et al. JPhysChem B, 2007.
> Irrespective, I cannot figure out why the dihedral energy blows up
> in the very first step when all the dihedral K_chi terms
> are set to zero.
>
> Thanks,
> Jindal
>
>
> On Sun, Feb 28, 2010 at 11:41 PM, Eduardo <chucruz_at_ks.uiuc.edu> wrote:
> The high energy can be due to constraints in your structure, but it
> can also be due to a mistake in your PSF assignment.
>
> As far as I remember, they provided all silica structures in the
> supplementary materials of their paper. Just need to use those ones.
>
>
>
> On Feb 28, 2010, at 9:51 PM, Jindal Shah wrote:
>> Hello NAMD users,
>>
>> I am trying to minimize a silica block generated from crystobalite
>> crystal structure.
>> The force field I am using is that from Lopez and MacKerell. The
>> starting
>> structure is in highly strained conformation for this FF and the
>> minimization
>> fails in the first step step as it cannot generate search
>> direction. The initial energies
>> for dihedral angles is 999999999.99. Even if setting all the
>> dihedral constants K_chi
>> to zero, I have not been able to get this energy to go to zero. It
>> is calculated to be
>> the same as before.
>>
>> Any suggestions?
>> Thanks,
>> Jindal
>
> =====================
> Eduardo Roberto Cruz Chu
> =====================
>
>
>
>
>

=====================
Eduardo Roberto Cruz Chu
=====================

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:53:50 CST