[gmx-users] position restraints

[Justin Lemkul]

On 6/13/15 4:47 AM, Kevin C Chan wrote:
> Sorry for digging the old thread, but I really have question concerning Justin’s reply.
>
>> On 17 May, 2015, at 05:05, Justin Lemkul <jalemkul at vt.edu> wrote:
>>
>>
>>
>> On 5/16/15 8:13 AM, Ming Tang wrote:
>>> Dear all,
>>>
>>> I want to minimize my system to relax the solvent with the positions of heavy atoms within the protein restrained. I searched the archives and got the idea that I need to put something like this in topology:
>>>
>>> #ifdef POSRES_WATER
>>> ; Position restraint for each water oxygen
>>> [ position_restraints ]
>>> ;  i funct       fcx        fcy        fcz
>>>     1    1       1000       1000       1000
>>> #endif
>>>
>>> But I still have no idea how to get the heavy atoms restrained. Could anybody tell me how to do this?
>>>
>>
>> pdb2gmx does this automatically for you.  It writes posre.itp, which specifies heavy atom restraints.  It's controlled by a similar #ifdef block.
>>
>>> I want to use gromos54a7 ff to simulate my collagen. As Justin suggested, I need to study the original paper "Definition and testing of the GROMOS force-field versions 54A7 and 54B7" to write my .mdp file. In the paper it says,
>>> 1.      Each system was energy minimised to relax the solvent with the positions of the heavy atoms within the protein and RNA restrained.
>>> 2.      All solute atom positions were restrained to their positions in the initial structure through a harmonic potential energy term with force constant of 2.5 9 104 kJ mol-1 nm-2.
>>> Could anybody tell me how to restrain the heavy atoms of protein, and how to get the solute atom positions restrained with a harmonic potential energy term with a certain constant?
>>
>> This is what posre.itp does.
>>
>> Also note that you don't necessarily have to follow that protocol exactly; my general advice for proper use of force fields is about proper nonbonded setup. Preparation of a system often depends on what the system is and if there are any special considerations.  For a simple protein system, it's pretty hard to break things.  I don't see much benefit in restraining atoms during minimization, though.
>
> I believe a significant number of the users learn MD protocols from reading papers and often it is difficult to trace back the reasons behind certain steps. Restraining atoms during minimization or even heating-up, pre-equilibrium is a very good example. While Justin mentioned that the step is kind of “useless”, I read quite a amount of similar approaches in papers that people “relax” their system step by step before a production.
>
> I will greatly appreciate if anyone can provide some original papers on discussing rigorous MD protocols (especially for nowadays complex biomolecular systems).
>

Every paper reads a bit different.  The simple answer is this: there is no 
inherently "right" or "wrong" way to prepare a system.  I don't generally think 
restraints during EM do much good, but there are instances where this is not 
true (e.g. rebuilding a loop region in a protein - restrain the known part of 
the structure and relax the loop to avoid perturbation).  The generic "I want to 
simulate a protein in water" follows a pretty easy recipe that is covered by 
most tutorials.

Restraints during initial solvent equilibration are de facto standard.  Without 
restraints, you risk (1) a bad contact by the solvent kicking the solute in an 
unphysical way that takes a long time to recover from or (2) simply crashing.

-Justin

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

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

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