[gmx-users] Use different fudgeLJ and fudgeQQ values in simulations with the AMBER force field

[Mark Abraham]

On 12/08/2011 12:16 AM, intra\sa175950 wrote:
>
> Dear All,
>
> Few years ago, C. Neale (thanks to him!) posted in the GROMACS mailing 
> list a very useful tutorial [1] to scale the Coulombic 1-4 
> interactions when we combine forcefields with different fudgeLJ and 
> fudgeQQ values. I am trying to apply this trick to a system containing 
> a peptide and a micelle simulated with AMBER99SB-ILDN and GLYCAM force 
> fields, respectively (see my previous message [2]).
>
> I have followed all the steps described in [2] and computed the 
> sigma_ij and epsilon_ij values for the peptide and the micelle with 
> comb_rules 2 (i.e. sigma_ij=1/2(sigma_i+sigma_j) and 
> epsilon_ij=sqrt(epsilon_i*epsilon_j). The epsilon_i and sigma_i are 
> taken from the [ atomtypes ] section in the ffnonbonded.itp file
>
> However, in the step 3 in [2], it is said that the peptide the pair 
> interaction values (i.e. epsilon) should be divided by 10 and these 
> "values/12" include in the pairtypes section. I suspect a typo here. 
> Indeed, if I set the general fudgeLJ and fudgeQQ to 1.0 and 0.1666666, 
> I should have a ratio of 10 (for (0.83333/(0.166666)))*1/2 for 
> epsilon_ij and not "12". Am I right?
>
> To test the changes, I have performed three nsteps=0 runs in NVT 
> ensemble for a simple system (1 peptide and 1 glycolipid solvated in 
> water) with the initial and modified and non-modified nonbonded 
> parameters:
>
> 1- With ffnonbbonded.itp file non-modified with fudgeLJ=0.5 and 
> fudgeQQ=0.83333   --- > peptide
>
> 2- With ffnonbbonded.itp file non-modified with fudgeLJ=1.0 and 
> fudgeQQ=1.0  --- >  micelle
>
> 3- With ffnonbbonded.itp file modified with fudgeLJ=1.0 and 
> fudgeQQ=0.1666666  --- >  all
>

Using mdrun -rerun allows you to compare more than one configuration in 
a run, and avoid any issues associated with bond constraints and 
neighbour lists.

> and compared the *.trr files obtained from these the three runs.
>
> The only differences I found are, indeed, in the force section of the 
> *.trr files for atoms involved in the modified pair section. To be 
> sure that my modifications are corrects, I would like to obtain the 
> individual pair energy values of the micelle and the peptide. How to 
> obtain this values?
>

Use grompp -pp, and selectively zero out charges and/or VDW parameters 
for appropriate [moleculetypes] in the reruns.

Mark

> [1] 
> http://lists.gromacs.org/pipermail/gmx-users/2006-September/023761.html
>
> [2] http://lists.gromacs.org/pipermail/gmx-users/2011-April/060839.html
>
> Thank you again for your help and advices.
>
> Stephane
>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110812/d61dcf4a/attachment.html>