[gmx-users] EM problem for homology model !!

[Kamil Tamiola]

narender wrote:

>Dear Users,
>Hi, i am getting problems in doing energy minimizing of my system of 3400 
>atoms (371 residues). This system is a homology build protein and i want to 
>do atleast 1000 steps of steepest descent. But its converginh to Fmax after 
>just 4 steps . I tried changing emtol and emstep but of no use.  
>Please help !!. Also it would be great if someone could tell me how can i fix 
>my protein's backbone and minimize only the sidechains.
>
>thank in advance
>NSM
>
>
> (My EM input file (.mdp is) 
>
>;
>cpp                 =  /lib/cpp
>define              =  -DPOSRE
>constraints         =  all-bonds
>integrator          =  steep
>nsteps              =  1000
>;
>;       Energy minimizing stuff
>;
>emtol               =  2300
>emstep              =  0.01
>
>nstcomm             =  1
>ns_type             =  grid
>rlist               =  1
>rcoulomb            =  1.0
>rvdw                =  1.0
>Tcoupl              =  no
>Pcoupl              =  no
>gen_vel             =  no
>~
>~
>~
>~
>_______________________________________________
>gmx-users mailing list
>gmx-users at gromacs.org
>http://www.gromacs.org/mailman/listinfo/gmx-users
>Please don't post (un)subscribe requests to the list. Use the 
>www interface or send it to gmx-users-request at gromacs.org.
>
>
>  
>
Try to increase the stepsize... for example 0.5 (I had got the same 
problem and it was the solution in my case)
As for fixing backbone, there is an option "freezegrps" in 
non-equilibirum dynamics, please refer to the manual - version 3.2.1 
page 143.
 Try to minimize the whole structure
first and than freeze the backbone an run some short MD with velocity 
generation at specified temprature. After this
extract ener.edr file and investigate the potential and total energy to 
se if there is an equilibrium and when it has equilibrated.
You can average the structure over given time an be sure that it is in 
local energy equilibirum.


Good luck,

Kamil