[gmx-users] making topology file manually

[Justin Lemkul]

On 6/8/16 10:49 AM, Mohsen Ramezanpour wrote:
> Thanks Justin,
>
> This is why I have got confused then, as each force field treats them in a
> different way.
>
> Knowing the details of force field seems necessary then. As another
> solution to this:
>
> Can we just make the topology file using some web servers or software (ATB,
> SwissParam, Antechamber, PRODRG, ...) and trust the output and continue
> with?
>
> By "trust" I literally mean to:
> 1) not be worry about the assigned lines (atom types, as well as bonds,
> angels, dihedrals, impropers, and pairs between these atom types) in the
> topology file as recognised by the above mentioned tools
> AND
> 2) check for the correct numbers of assigned lines based on information in
> force field files (ffbonded, ffnonbonded) and of course validating the
> simulation results
>

Never trust a black box.  Antechamber is generally a pretty robust program for 
AMBER.  I can't speak to the accuracy of SwissParam, but I've heard it's pretty 
good.  I prefer CGenFF for CHARMM topology generation because it gives you an 
explicit parameter penalty score to alert you to parameters that may require 
more work.  ATB is generally good but often requires minor revision.  PRODRG 
always requires a complete overhaul of the charges (at minimum).

Always validate.  Go back to the original force field literature and carry out 
whatever types of small-molecule validation were done in the parent force field. 
  Otherwise, you're playing a pretty costly game of numerical roulette.

-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

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