[gmx-users] eNERGY mINIMIZATION

[Mark Abraham]

On 18/01/2011 2:50 PM, bharat gupta wrote:
> Hi,
> I did the energy minimization of my protein (230 amino acids) after 
> adding ions .. and after checking the potential energy of the system I 
> was getting thee graph which is different from the one given  in the 
> lysozyme tutorial i.e. the the potential line in graph is not parallel 
> to the x axix..

You're comparing apples and oranges... maybe lysozyme is already pretty 
close to the minimum. Consider the axis scales, also. Note that the 
energy change per EM step doesn't have any physical meaning.

> Here is are conditions that I used for minimization :
>
> define = -DFLEX_SPC
>
> constraints = none
>
> integrator = steep
>
> nsteps = 50000
>
> ;
>
> ; Energy minimizing stuff
>
> ;
>
> emtol = 2000
>
> emstep = 0.01
>
> nstcomm = 1
>
> ns_type = grid
>
> rlist = 1
>
> coulombtype = PME
>
> rcoulomb = 1.0
>
> rvdw = 1.0
>
> Tcoupl = no
>
> Pcoupl = no
>
> gen_vel = no
>
> ------------------------
> I got the following result :-
> Steepest Descents converged to Fmax < 2000 in 296 steps
>
> Potential Energy = -6.9540075e+05
>
> Maximum force = 1.9863752e+03 on atom 1669
>
> Norm of force = 3.7864201e+01
>
> ------------------------------
>
> I want to know which parameters are important  to get a correct 
> minimized structure and what value do I have to take for emtol and 
> nsteps to minimize the structure properly.. Also how would I know that 
> at what value of emtol my protein will be minimized..
>

How long should a piece of string be? Your PE is OK, so if you reduce 
emtol then you can reduce the maximum force. This will prolong the EM if 
you think you need to do that.

> I checked the manual to get some details but there only the basic 
> theory of minimzation is written.. A detailed explanation will be more 
> helpful...
>

The purpose of EM is normally to relieve bad contacts, fill solvent 
voids, etc. Do something plausible (like you've done) and then try the 
next step. If you get problems there, go back and do EM harder (among 
other strategies)

Mark