Re: using CHARMM22STAR forcefield in NAMD

From: Victor Ovchinnikov (ovchinnv_at_georgetown.edu)
Date: Fri Jun 08 2012 - 09:27:46 CDT

Aaron,

There might be a more elegant solution that does not require modifying
code, but a draconian thing you could do
is to try commenting out the part of NAMD source that disallows
dihedrals with duplicate periodicities, and recompiling.

The relevant snippet is in Parameters.C lines 1660--1678 (an older CVS
version I have):

+++++++++++++++++++

                          for (i=0; i<ptr->multiplicity; i++)
                           {
                             if (!same_as_last) {
                               iout << " k=" << ptr->values[i].k
                                    << " n=" << ptr->values[i].n
                                    << " delta=" <<
ptr->values[i].delta << "\n";
                             }
                             if (ptr->values[i].n == new_node->values[0].n)
                             {
                               iout << iWARN << "IDENTICAL PERIODICITY!
REPLACING OLD VALUES BY: \n";
                               ptr->values[i].k = new_node->values[0].k;
                               ptr->values[i].delta =
new_node->values[0].delta;
                               iout << " k=" << ptr->values[i].k
                                    << " n=" << ptr->values[i].n
                                    << " delta=" <<
ptr->values[i].delta<< "\n";
                               replace=1;
                               break;
                             }
                           }

+++++++++++++++

--Victor

On 06/08/2012 04:23 AM, Aaron Oakley wrote:
> Hi All,
>
> In the following publication, a modified CHARMM force-field called charmm22* is described:
>
> Piana S, Lindorff-Larsen K, Shaw DE. How robust are protein-folding simulations with respect to force field parameterization?
> Biophys J. 2011 May 4;100(9):L47-9.
>
> They claim that their new forcefield, CHARMM22*, has a much better balance of secondary structure stabiliy with respect to the existing CHARMM forcefields.
>
> The only "edition" of CHARMM22* I have been able to find comes with the DESMOND distribution and is in a DESMOND-specific format.
>
> The changes between the latest CHARMM and CHARMM22* are:
> -modifications to partial charges on ASP, GLU and ARG.
> -modification of dihedrals on ASP
> -removal of CMAP correction from all residues except GLY and PRO
> -modification of backbone phi and psi dihedral potentials.
>
> Its fairly straightforward to introduce these into par_all22star_prot.inp file.
>
> The problem occurs when I run NAMD. I get the following warnings:
> Warning: IDENTICAL PERIODICITY! REPLACING OLD VALUES BY:
> k=-1.825 n=1 delta=90
>
> ..and so on.
>
> This is due to the way the phi and psi dihedrals were re-parameterised by the authors.
> In charmm parameter file format they look like this:
>
> C CT1 NH1 C 0.2000 1 0.00 ! C22STAR
> C CT1 NH1 C -0.2000 2 0.00 ! C22STAR
> C CT1 NH1 C -1.8250 1 90.00 ! C22STAR
> C CT1 NH1 C 0.0540 2 90.00 ! C22STAR
> C CT1 NH1 C -0.1090 1 150.00 ! C22STAR
> C CT1 NH1 C 0.0770 2 150.00 ! C22STAR
> C CT1 NH1 C -0.0960 3 150.00 ! C22STAR
> C CT1 NH1 C -0.1060 4 150.00 ! C22STAR
>
> and
>
> NH1 C CT1 NH1 0.6000 1 0.00 ! C22STAR
> NH1 C CT1 NH1 -2.2860 1 60.00 ! C22STAR
> NH1 C CT1 NH1 -0.3620 2 60.00 ! C22STAR
> NH1 C CT1 NH1 0.0130 1 150.00 ! C22STAR
> NH1 C CT1 NH1 0.9420 2 150.00 ! C22STAR
> NH1 C CT1 NH1 0.2670 3 150.00 ! C22STAR
> NH1 C CT1 NH1 -0.1270 4 150.00 ! C22STAR
>
> is there a workaround that can be used in NAMD?
>
> Thanks,
>
> A/Prof Aaron Oakley
> School of Chemistry
> University of Wollongong
> Northfields Avenue
> Wollongong, NSW, 2522
> Australia
> Phone: (02) 4221 4347
> Fax: (02) 4221 4287
> Email: aarono_at_uow.edu.au
>
>
>

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:38 CST