[gmx-users] mdp params and forcefields

[Justin A. Lemkul]

Алексей Раевский wrote:
> Hi. I need your help dealing with parametrization of mdp file. Look, I 
> want to understand how can we change such params like rlist, rvdw, 
> cut-off, rcoulomb. As I know each forcefield has its own properties and 
> calculated values of this params. But I've found that in different 
> articles this values were the same, though the forcefields were not... 

Some settings are transferable between force fields, others are not.

> Also I've read that each meaning of this params I can change depeding on 
> my task... Changing rlist we will get more accurate results, as there 
> are more atoms around the one, which motion is calculated, will be 

This is not necessarily true.  There is a delicate balance within all force 
fields between the contributions of the nonbonded terms and the distance at 
which they are evaluated.  Considering more interactions may actually break the 
force field and make your results less accurate.  Unless you can demonstrate 
that your results with an altered value of rlist (or any other, really), then 
stick with what the force field prescribes.

> considered. And what about other params? What for, as example, we can 
> cut-off an influence of coulumb forces with reducing the distance of 
> them? In what situation I can change them and how it will affect the 
> result?

See above.  Don't tamper unless you know you can prove what you're doing is 
better, using a very reliable test system.

> And also were can I find some defined parameters of mdp for charmm27 or 
> amber99, I mean not from articles, but any supplements which come with 
> forcefield files...

As this information is already in the literature, no such supplements are provided.

-Justin

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================